HELLFIRE, HOME, AND HARM: AN INVESTIGATION OF THE INTERACTION
BETWEEN RELIGIOSITY, FAMILY PROCESSES, AND
ADOLESCENT DEVIANT BEHAVIOR
Except where reference is made to the work of others, the work described in this
dissertation is my own or was done in collaboration with my advisory committee. This
dissertation does not include proprietary or classified information.
________________________________________
Lloyd Ernest Pickering
Certificate of Approval:
_________________________________
Gerald Halpin
Professor
Educational Foundations, Leadership and
Technology
_________________________________
Thomas A. Smith
Associate Professor
Human Development and Family Studies
_________________________________
Alexander T. Vazsonyi, Chair
Associate Professor
Human Development and Family Studies
_________________________________
Brian E. Vaughn
Professor
Human Development and Family Studies
___________________________
Stephen L. McFarland
Dean
Graduate School
HELLFIRE, HOME, AND HARM: AN INVESTIGATION OF THE INTERACTION
BETWEEN RELIGIOSITY, FAMILY PROCESSES, AND
ADOLESCENT DEVIANT BEHAVIOR
Lloyd Ernest Pickering
A Dissertation
Submitted to
the Graduate Faculty of
Auburn University
in Partial Fulfillment of the
Requirements for the
Degree of
Doctor of Philosophy
Auburn, Alabama
May 13, 2005
iii
VITA
Lloyd Ernest Pickering, son of Ernest Dinwoodie and Ariel Yvonne (Thomas)
Pickering, was born January 27, 1963, in Robbinsdale, Minnesota. He graduated as
Valedictorian from Emmanuel Christian School in Toledo, Ohio in June, 1980. He then
attended Baptist Bible College in Clarks Summit, Pennsylvania where, after earning
National Dean?s List and Who?s Who Among Students in American Universities and
Colleges honors and being inducted into the Alpha Gamma Epsilon honor society, he
graduated cum laude with a Bachelor of Science degree in Bible (Pastoral Studies) in
May, 1984. While serving as a church youth pastor, he entered Central Theological
Seminary in Minneapolis, Minnesota where he graduated magna cum laude with a Master
of Divinity degree in June, 1992. After completing his church ministry, he entered the
Graduate School at Auburn University in Auburn, Alabama, in September, 1995. He
graduated in March, 1999, with a Master of Science degree in Human Development and
Family Studies. On May 21, 1999, he married Kellie Rae Calhoun, daughter of David and
Darla (Haglund) Calhoun. On October 16, 2000, the same day his father passed away, his
son, Northe Crosse Ernest Pickering, was born. On August 29, 2002, his daughter,
Darielle Trystin Pickering, was born. His son, Nighte Calhoun Pickering, was born on
January 26, 2005.
iv
DISSERTATION ABSTRACT
HELLFIRE, HOME, AND HARM: AN INVESTIGATION OF THE INTERACTION
BETWEEN RELIGIOSITY, FAMILY PROCESSES, AND
ADOLESCENT DEVIANT BEHAVIOR
Lloyd Ernest Pickering
Doctor of Philosophy, May 13, 2005
(M.S., Auburn University, 1999)
(M.Div., Central Theological Seminary, 1992)
(B.S., Baptist Bible College, 1984)
Directed by Alexander T. Vazsonyi
There is an abundant literature dealing with the relationship between religiosity
and deviant behavior in adolescence (Baier & Wright, 2001; Johnson, de Li, Larson, &
McCullough, 2000; Johnson, Jang, Larson, & de Li, 2001). A considerable amount of
research has also been conducted on the nature of the family processes/deviance
connection (Baumrind, 1996; Patterson, 1982; Simons et al., 1991). There are few
studies, however, that investigate the interplay between religiosity and family processes in
deterring deviance (Benda & Corwyn, 2000; Benda & Toombs, 2000). The current study
sought to test two potentially competing models of the interrelationship between
religiosity, family processes, and deviance: 1) family processes mediating the relationship
v
between religiosity and deviance, and 2) religiosity moderating the relationship between
family processes and deviance. In addition, these models were also tested to see whether
they would be moderated by age (younger, older) or sex (male, female).
A sample of N = 865 high school students responded to survey questions dealing
with religiosity (8 single items, 3 scales), family processes (6 scales for both mothers and
fathers), and deviance (7 subscales, a total deviance scale, and 4 additional scales). Race
and SES were used as control variables. A structural equation modeling (SEM) analytic
approach (Holmbeck, 1997) was used for the mediation analyses and results indicated
that family processes never mediated the relationship between religiosity and deviance.
SEM (Mounts, 2002) was also utilized for testing moderation effects, both for religiosity
as well as sex and age. Results from these moderation analyses indicated that, not only
was it found that religiosity did not moderate the relationship between family processes
and deviance, but also neither age nor sex ever moderated any of these structural models.
Additional findings in the study included the fact that religiosity and deviance were
moderately, negatively correlated (r = .31); that the religiosity scales, particularly
relational practice, performed better than the single-item religiosity measures; and that
religiosity was related to all types of deviance, including secular and victim deviance.
Future research in this area should seek to improve methods of data collection
(e.g., multi-method) and to develop better scalar measures of religiosity.
vi
ACKNOWLEDGEMENTS
Special thanks are due to the following people:
Dr. Alexander Vazsonyi, who patiently prodded me through this arduous task.
Drs. Gerald & Glennelle Halpin, whose generosity is exceeded only by their kindness.
My wife, Kellie, who loves me when no one else does.
My son, Northe, who has no idea the inspiration he is to me.
My daughter, Darielle, whose smile brightens the darkest days.
My son, Nighte, who is, indeed, a warrior.
My mother, Yvonne, who is a living example of perseverance amid great difficulty.
My sister, Dawn, who never gave up on me
My father, Ernest, who has always been my greatest hero.
vii
Style manual or journal used: Publication Manual of the American Psychological    
Association (5th Edition)                                                                                                   
Computer software used: Word Perfect 10.0 for Windows, SPSS 12.0 for Windows,
Amos 5.0 for Windows, Microsoft PowerPoint 2003 for Windows                                      
viii
TABLE OF CONTENTS
LIST OF TABLES AND FIGURES ...................................... xvii
CHAPTER 1: INTRODUCTION ...........................................1
CHAPTER 2: LITERATURE REVIEW .....................................11
Introduction .....................................................11
Religiosity and Deviance ...........................................11
Early Research (Pre-1969) ....................................13
Studies Concluding There Is No Effect ....................13
Studies Concluding There Is An Effect ....................16
Summary of Early Research (Pre-1969) ...................19
A Landmark Study ..........................................20
Specific Responses to Hirschi & Stark (1969) .....................22
Other Important Literature ....................................28
Family and Deviance ..............................................31
Family Processes as a Dimensional Construct .....................31
Literature in the Autonomy Dimension ..........................34
Behavioral Control ....................................34
Psychological Control .................................38
Literature in the Harmony Dimension ...........................41
ix
Emotional Closeness ..................................41
Communication ......................................44
Literature in the Conflict Dimension ............................48
Summary of Family and Deviance Literature .....................52
Family and Religiosity .............................................54
Parenting .................................................55
Marriage ..................................................57
Marital Dissolution ...................................57
Marital Conflict ......................................58
Marriage and Parenting ......................................59
Summary of Family and Religiosity Literature ....................59
Theoretical Considerations .........................................60
Family, Religiosity, and Deviance ....................................64
Other Important Intervening Factors ..................................71
Type of Deviance ...........................................73
Developmental Issues .......................................77
Age ................................................77
Age and Religiosity .............................78
Age and Family ................................79
Age and Deviance ..............................80
Age, Religiosity, Family, and Deviance .............81
Summary of Age, Religiosity, Family, and Deviance ...83
x
Sex ................................................84
Sex and Religiosity .............................84
Sex and Deviance ...............................85
Sex, Religiosity, Family, and Deviance ..............85
Summary of Sex, Religiosity, Family, and Deviance ...88
Measurement ..............................................88
Summary of Other Intervening Factors ..........................91
Rationale for the Current Investigation ................................92
Hypotheses ......................................................93
Hypothesis 1: Bivariate Relationships ...........................94
Hypothesis 2: Measurement ...................................94
Hypothesis 3: Type of Deviance ...............................94
Hypothesis 4: Mediation .....................................95
Hypothesis 5: Moderation ....................................96
CHAPTER 3: METHOD .................................................98
Sample .........................................................98
Procedures ......................................................99
Measures .......................................................99
Background Variables .......................................99
Sex ................................................99
Age ...............................................100
xi
Grade .............................................100
School Age .........................................100
Race ..............................................103
SES ...............................................103
Family Income ................................103
Primary Wage Earner?s Occupation ................103
SES Scale ....................................104
Religiosity ...............................................104
Single Item Measures .................................105
Church Attendance .............................105
Church Status .................................105
Religious Involvement ..........................105
Bible Reading .................................106
Praying ......................................106
Music Listening ...............................106
Religious Conviction ...........................106
Religious Salience .............................107
Multi-Item Scales ....................................107
Ritualistic Participation .........................107
Relational Practice .............................107
Religiosity ...................................109
Family Processes ..........................................109
xii
Closeness ..........................................110
Support ............................................110
Monitoring .........................................110
Communication .....................................111
Conflict ...........................................111
Peer Approval ......................................111
Deviance ................................................112
Normative Deviance .................................112
Vandalism ...................................113
Alcohol Use ..................................113
Drug Use ....................................113
School Misconduct .............................114
General Deviance ..............................114
Theft ........................................114
Assault ......................................115
Total Deviance ................................115
Victim Deviance ....................................115
Victimless Deviance .................................116
Antiascetic Deviance .................................116
Secular Deviance ....................................116
Plan of Analysis .................................................116
Initial Analyses ............................................116
xiii
Scale Construction ...................................116
Measurement Models .................................117
Analysis Preparation .................................119
Initial Descriptive Statistics ............................121
Analyses Related to Hypotheses ..............................121
Hypothesis 1 ........................................121
Hypothesis 2 ........................................122
Hypothesis 3 ........................................123
Hypothesis 4 ........................................125
Hypothesis 5 ........................................127
CHAPTER 4: RESULTS ................................................128
Hypothesis 1: Bivariate Relationships ................................128
Part A ...................................................128
Part B ...................................................130
Part C ...................................................130
Hypothesis 2: Measurement ........................................133
Hypothesis 3: Type of Deviance ....................................134
Part A ...................................................134
Part B ...................................................135
Part C ...................................................136
Part D ...................................................137
xiv
Part E ...................................................138
Part F ...................................................138
Part G ...................................................138
Part H ...................................................139
Hypothesis 4: Mediation ..........................................139
Part A ...................................................143
Part B ...................................................148
Part C ...................................................149
Part D ...................................................153
Part E ...................................................153
Ritualistic Participation ...............................155
Relational Practice ...................................156
Religiosity .........................................156
Part F ...................................................159
Ritualistic Participation ...............................161
Religiosity .........................................162
Part G ...................................................163
Part H ...................................................163
Part I ....................................................164
Hypothesis 5: Moderation .........................................166
Part A ...................................................166
Part B ...................................................166
xv
Part C ...................................................167
CHAPTER 5: DISCUSSION .............................................170
Does religiosity matter? ...........................................170
What is the role of sex? ...........................................171
What is the role of age? ...........................................173
What is the role of measurement? ...................................174
Does religiosity deter all types of deviance or only some? ................176
Do family processes mediate the effect of religiosity on deviance? .........178
Does religiosity moderate the effect of family processes on deviance? .......183
Weaknesses of the Current Study ...................................185
Future Directions for Research .....................................187
REFERENCES .......................................................190
APPENDIX A: ADOLESCENT FAMILY PROCESSES SCALE (AFP) ..........244
APPENDIX B: RELIGIOSITY ITEMS .....................................246
APPENDIX C: NORMATIVE DEVIANCE SCALE (NDS) ....................248
APPENDIX D: RELIGIOSITY CORRELATION MATRIX ....................251
xvi
APPENDIX E: FAMILY PROCESSES CORRELATION MATRIX .............252
APPENDIX F: DEVIANCE CORRELATION MATRIX ......................253
APPENDIX G: FAMILY PROCESSES AND RELIGIOSITY CORRELATION
MATRIX ......................................................254
APPENDIX H: FAMILY PROCESSES AND DEVIANCE CORRELATION              
MATRIX ......................................................255
xvii
LIST OF TABLES AND FIGURES
Figure 1. Social control model of religiosity and deviant behavior ..................4
Figure 2. Social control model of family processes, religiosity, and deviant behavior ...5
Figure 3. Social control model hypothesizing mediating role of family processes ......5
Figure 4. Social control model hypothesizing moderating role of religiosity ..........6
Figure 5. Social control model hypothesizing two-way interaction of family processes and
religiosity ........................................................7
Figure 6. Model showing family processes mediating religiosity single-item measures
(Hypothesis 4A) and ritualistic participation (Hypothesis 4B) where there is a
significant negative pathway from family processes to deviance and no significant
direct pathway from religion to deviance ...............................95
Figure 7. Model showing the partially mediating role of family processes between
relational practice and religiosity where negative direct and indirect pathways
between religion and deviance both exist ..............................96
Figure 8. Model showing religiosity moderating between family processes and deviance
where the predictive pathways for the two groups (low and high religiosity) are
not equal ........................................................97
Table 1. Crosstabulations of age and grade variables ..........................100
xviii
Table 2. Frequencies and descriptives of dichotomous age and school age variables ..101
Table 3. Reliability statistics for study scales by sex, age, and for total sample ......108
Table 4. Fit indices showing stepwise model improvement for deviance measurement
model using correlated errors recommended by modification indices . . . . . . . . 119
Table 5. Summary of study variables .......................................120
Table 6. Correlations between demographic variables and measures of religiosity ...122
Table 7. Correlations between demographic variables and measures of
family processes .................................................123
Table 8. Correlations between demographic variables and measures of deviance ....124
Table 9. Frequency statistics for demographic variables ........................126
Table 10. Descriptive statistics for demographic variables ......................127
Table 11. Second-order correlations of religiosity measures with deviance measures for
total sample ....................................................129
Table 12. Second-order correlations of religiosity measures with deviance measures by
sex ...........................................................131
Table 13. Second-order correlations of religiosity measures with deviance measures by
age ...........................................................132
Table 14. Comparison of second-order correlations of religiosity measures (single- and
multi-item) with total deviance .....................................134
Table 15. Second-order correlations of religiosity measures with different types of
deviance for total sample ..........................................135
xix
Table 16. Second-order correlations of religiosity measures with different types of
deviance by sex .................................................136
Table 17. Second-order correlations of religiosity measures with different types of
deviance by age .................................................137
Figure 10. Direct model, testing effect of religiosity on deviance with no intervening
variable ........................................................140
Table 18. Parameter estimates and fit indices for direct models of religiosity and deviant
behavior .......................................................143
Table 19. Parameter estimates for overall models involving tests of mediating role for
family processes between individual religiosity items and deviant behavior ..144
Table 20. Model statistics for overall models involving tests of mediating role for family
processes between individual religiosity items and deviant behavior ........145
Table 21. Fit statistics in comparisons between overall models with constrained and
unconstrained pathway between single-item religiosity measures
and deviance ....................................................146
Figure 12. Actual structural model tested ...................................147
Table 22. Parameter estimates for overall models involving tests of mediating role for
family processes between religiosity scales and deviant behavior
for total sample .................................................150
Table 23. Model statistics for overall models involving tests of mediating role for family
processes between religiosity scales and deviant behavior for total sample ...151
xx
Table 24. Fit statistics in comparisons between overall models with constrained and
unconstrained pathway between religiosity scales and deviance
for total sample .................................................152
Table 25. Fit statistics in comparisons between fully variant and fully invariant models by
sex ...........................................................155
Table 26. Path coefficients and comparisons by sex: Ritualistic participation .......157
Table 27. Path coefficients and comparisons by sex: Relational practice ...........158
Table 28. Path coefficients and comparisons by sex: Religiosity .................159
Table 29. Fit statistics in comparisons between fully variant and fully invariant models by
age ...........................................................160
Table 30. Path coefficients and comparisons by age: Ritualistic participation .......162
Table 31. Path coefficients and comparisons by age: Religiosity .................162
Table 32. Direct, indirect, and total effects of religiosity scales on family processes scales
in fully variant model by sex, age, and for total sample ..................165
Table 33. Moderation of family processes by religiosity: Path coefficients and
comparisons ....................................................167
Table 34. Moderation of family processes by sex and religiosity: Path coefficients and
comparisons ....................................................168
Table 35. Moderation of family processes by age and religiosity: Path coefficients and
comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
1 Throughout this manuscript, religiosity is referred to as a ?deterrent? to
adolescent deviant behavior. The use of this word in its various forms is consistent with
previous literature. It is also consistent with a control theory perspective, since, as Hirschi
(1969) stated, bonds with conventional control agents can serve as ?a major deterrent? to
antisocial behavior (p. 83). Understandably, the use of this terminology may carry with it
some controversy, since, to some, this may imply causality. It is readily acknowledged
that the current study is unable to make inferences regarding causality due to the cross-
sectional nature of the data. Therefore, the use of words such as ?deter,? ?deterrent,? and
?deterrence? are intended only to accentuate the fact that the presence of religious
behavior is consistently accompanied by the absence of deviant behavior and that, though
the ultimate goal, on a macro-level, of any socializing agent, is to promote conformity to
social norms, the mechanism, on a micro-level, by which this is accomplished is
deterrence (e.g., teaching young people not to do certain things). As Hirschi (1969) said,
?If the bond to the [controlling agent] is weakened, the probability of delinquent behavior
increases; if this bond is strengthened, the probability of delinquent behavior declines? (p.
88).
1
CHAPTER 1: INTRODUCTION
The question of whether religion has a deterring effect1 on adolescent deviant
behavior has been studied for quite some time. Opinions among early social scientists
varied extensively, from those who simply assumed that religion led to conformity (e.g.,
Coogan, 1954; Davis & Moore, 1945) to those who felt religion was the very problem
(e.g., Barnes & Teeters, 1951; Bonger, 1916; Ellis, 1910; Durkheim, 1951; Lombroso,
1918; Lunden, 1964; Reckless & Smith, 1932; Steiner, 1924; von Hentig, 1948). An
important study by Hirschi and Stark (1969) provided finality to this debate for many
when it suggested that religiosity (i.e., religious activities and beliefs) had little or no
effect on deviance. Subsequently, and based on this evidence, a number of social
scientists assumed that religion has no deterrent effect on deviance (e.g., Cochran &
2
Akers, 1989; Schur, 1969; Sloane & Potvin, 1986). However, recent research has
provided new evidence on the importance of religiosity in adolescent deviance.
In a recent article employing latent variable structural modeling, Johnson, Jang,
Larson, and de Li (2001) modeled religiosity as a multi-item latent construct, rather than a
single item as in much of the earlier work in this area. In doing so, they found conclusive
evidence that ?the effects of religiosity on delinquency are neither spurious nor indirect?
(p. 37) referring to the fact that, in this study, religiosity had a consistent effect on
deviance even after controlling for several important background (i.e., age, race, SES)
and theoretical (i.e., delinquent peer association) predictors. Furthermore, this evidence
which was based on longitudinal data, was found at each of three time points, thus
justifying the causal inference that religiosity deters deviance. In other recent work,
Johnson, de Li, Larson, and McCullough (2000) used a systematic review approach to
examine the cumulative findings of the previous literature in this area. Forty studies were
identified and aggregated quantified findings indicated that a consistent negative
relationship existed between religiosity and deviance. The authors concluded that ?a
commitment to religious values and beliefs can have both an immediate and a long-term
impact on their [i.e., adolescents?] behavior? (p. 46). In a similar investigation, Baier and
Wright (2001) used a meta-analytic approach to survey the evidence. Analyses on data
from 60 studies confirmed that ?religion had a statistically significant, moderately sized
effect on crime? (p. 16). Again, based on the evidence, the authors suggested that there
exists ?solid evidence? of a ?deterrent effect of religion? (p. 17).
3
Clearly, recent empirical evidence suggests that religiosity may serve to deter
adolescent deviant behavior. Yet, its role as a deterrent has not been sufficiently explored.
One of the best theoretical perspectives or frameworks for such an exploration is
Hirschi?s (1969) social control theory. In it, he proposed that adolescents with strong
bonds to family and society are less likely to deviate from conventional behavior than
those with weak bonds. In other words, a strong bond to a conventional institution is the
necessary deterrent to deviant behavior. From this perspective, religiosity could be
considered a bonding mechanism for the deterrence of deviance. In fact, even Baier and
Wright (2001) recognized this when they stated that, ?Although religion as such was not
discussed in the original model [i.e., Hirschi?s], social control theory would hold that by
strengthening an individual?s bond to society, religious institutions should deter criminal
behavior? (p. 4). Thus, in a social control framework, religiosity would be considered a
conventional social structure that fortifies an individual?s bond with conventional society
and, therefore, leads to conformity or deters deviance. Indeed, social control theory offers
unique utility for the examination of this question and, in fact, religiosity has been
repeatedly conceptualized in this manner in empirical work (e.g., Cochran, Wood, &
Arneklev, 1994; Evans et al., 1996; Ross, 1994). So, a basic control model of the
relationship between religiosity and deviance may conceptualize religiosity as a social
control agent or mechanism, similar to the family or the school as originally hypothesized
by Hirschi, to which an adolescent may bond, thus deterring deviance (see Figure 1).
However, simply attributing a negative relationship to religiosity and deviance
may not be sufficient to fully understand the nature of the interaction between these two 
4
Figure 1. Social control model of religiosity and deviant behavior.
constructs. Therefore, a number of investigators have attempted to go beyond this simple
bivariate relationship by seeking to identify potentially intervening factors. The role of
family processes in preventing adolescent deviant behavior is well-established, both
theoretically and empirically. From a theoretical perspective, social control theory
establishes a place of prominence for family processes (i.e., parenting practices or family
functioning) in the deterrence of deviance when it proposes that affective ties with
conventional people (i.e., attachment), particularly parents and family members who
serve as the primary socialization agents for children, will impede deviant behavior. In
addition, previous empirical investigations have established the importance of family
processes in deviance (e.g., Baumrind, 1996; Patterson, 1982; Patterson & Dishion, 1985;
Simons, Whitbeck, Conger, & Conger, 1991).
So we know that both religiosity and family processes are important in the
explanation of adolescent deviant behavior. But an important question that remains
unanswered is how religiosity and family processes may interact in order to explain
deviance. It is possible that religiosity has an independent effect on deviance in addition
to family processes (see Figure 2). More importantly, both of these social control
elements may have a hierarchical causal order with one being more distal whose effects 
5
Figure 2. Social control model of family processes, religiosity, and deviant behavior.
are, perhaps, mediated or moderated by the other more proximal influence. For instance,
family processes might mediate the effects of religiosity on deviance (see Figure 3). In
such a mediating model, family processes might ?explain away? any relationship between
religiosity and deviance. However, if such a model were tested and religiosity maintained
a significant relationship with deviance, even after controlling for family processes, this
would be important evidence for the strength of religiosity as a separate control
mechanism. Recent work by Benda and Toombs (2000) and Benda and Corwyn (2000)
Figure 3. Social control model hypothesizing mediating role of family processes.
6
has attempted to examine the interplay between family processes and religiosity in exactly 
this manner. Results were mixed with some indication that the effect of religiosity on
deviance was mediated by family processes (Benda & Corwyn, 2000) and some
indication that it was not (Benda & Toombs, 2000). In a similar test of mediation, Foshee
and Hollinger (1996) found that maternal religiosity negatively predicted adolescent
alcohol use and that this relationship was not explained by any background (e.g., race,
gender, age, education, family structure) or control (attachment, belief in conventional
rules, commitment to conventional activities) variables.
It is also possible that the relationship between family processes and religiosity is
even more complex and can not be explained by simple mediation. For instance, it may
be that one conditions the other in a moderating role. Based on previous literature, we
know that family and religion are linked (Free, 1991; Mahoney, Pargament, Tarakeshwar,
& Swank, 2001; Snarey & Dollahite, 2001). The evidence suggests that this link is both
direct (i.e., religiosity leads to better family functioning; e.g., Brody, Stoneman, & Flor,
1996; Gunnoe, Hetherington, & Reiss, 1999; Pearce & Axinn, 1998) and indirect (i.e.,
religiosity leads to better marriage which leads to better parenting and family functioning;
e.g., Booth, Johnson, Branaman, & Sica, 1995; Brody, Stoneman, Flor, & McCrary,
Figure 4. Social control model hypothesizing moderating role of religiosity.
7
2 See later discussion (p. 34) on the use of this term.
1994; Mahoney et al., 1999) as well as bidirectional (i.e., greater attachment to parents
leads to higher levels of responsiveness to parental attempts to promote healthy
socialization practices, e.g., religious involvement; Marcos, Bahr, & Johnson, 1986).
Therefore, there is some evidence indicating a possible interaction effect. From a social
control perspective, this could be modeled by including religiosity as a moderator
between family processes and deviant outcomes (see Figure 4). In other words, is there
anything unique about the parenting practices of families that score high on measures of
religiosity in comparison to those of families that do not? Do these families perhaps
parent in systematically different ways? For instance, Bahr, Maughan, Marcos, and Li
(1998) found evidence that families that were more religious exhibited stronger bonds
between parents and children and higher levels of parental monitoring.2
In fact, based on Hirschi?s social control theory, it may even be important to test
for a two-way interaction between family processes and religiosity in their relationship to
deviance (see Figure 5). Recall that social control theory hypothesizes that the attachment
or bond that will serve to effectively deter deviant behavior will be to persons of a 
conventional nature (e.g., parents). This hypothesis makes the assumption that the parents
Figure 5. Social control model hypothesizing two-way interaction of family processes
and religiosity.
8
to whom the child is attached are conventional in their own beliefs and behaviors.
Unfortunately, of course, not all parents are conventional themselves. It is likely that
conventional parents would be more vigilant in their supervision (see footnote #2) of their
child since their own beliefs would dictate the importance of teaching conventional (i.e.,
law-abiding, non-norm-violating) behavior to their child. Furthermore, an effective bond
(i.e., one with conventional parents) might also be characterized by less parent-child
conflict and better family communication. In other words, just reporting closeness to a
parent may not be a sufficient operationalization of Hirschi?s intended meaning when he
spoke of attachment to conventional people as a positive socializing factor for the
deterrence of deviant behavior.
In conclusion, although Hirschi and Stark (1969) suggested that ?participation in
religious activities and belief in a supernatural sanctioning system have no effect on
delinquent activity? (p. 211), there seems to be sufficient recent evidence that religiosity
impacts whether an adolescent will be deviant, especially in conjunction with family
processes. Therefore, the social control perspective allows the framing of several
potentially competing and testable questions and hypotheses about these relationships: 1)
Is religiosity consistently related to deviance?; 2) Do religiosity and family processes
work together to predict deviance in an additive fashion?; 3) Does family processes
mediate the relationship between religiosity and deviance?; 4) Does religiosity moderate
the relationship between family processes and deviance?; 5) Do religiosity and family
processes interact in their relationship with deviance?
9
In addition to these questions, a further study of the relationships between family
processes, religiosity, and deviance seems very timely because of three other important
issues requiring further clarification. First, previous work has indicated that religiosity
may be differentially related to different types of deviant behavior. More specifically,
some studies have found evidence that religiosity is important with regard to ?lesser?
deviant acts (what may be termed status offenses, anti-ascetic behavior, property offenses,
or victimless crime), but not index or criminal offenses (crimes against persons; Burkett,
1993; Burkett & White, 1974; Cochran et al., 1994; Middleton & Putney, 1962).
However, findings in this area have been inconsistent and inconclusive since others have
found that religiosity impacts a wide variety of deviant behaviors, including victim or
person related crimes (Higgins & Albrecht, 1977; Jensen & Erickson, 1979). Second, key
developmental issues have been largely ignored in this area of research, including age and
gender. There has been some evidence that the relationship between family and religiosity
varies by gender (e.g., Fiese & Tomcho, 2001; Flor & Knapp, 2001; Sullivan, 2001). In
addition, we know that there are strong differences in deviant behavior associated with
age (e.g., Osgood, Johnston, O?Malley, & Bachman, 1988). At the same time, there has
been some indication that age and gender are not important in understanding the
relationship between family, religion, and deviance (e.g., Foshee & Hollinger, 1996).
Therefore, it seems that an examination of the relationships between these variables by
age and gender would be an important addition to the literature. Third, most previous
empirical research has exhibited a methodological weakness by relying on the single-item
measurement of religiosity: ?Researchers have rarely modeled religiosity as a latent
10
variable being measured by multiple indicators, which would be appropriate for an
abstract, multidimensional concept such as religiosity? (Johnson et al., 2001, p. 25). In
fact, Johnson et al. (2000) stated that the reason for inconsistent findings on the
relationship between religiosity and deviance was probably mostly methodological and
?with improvements in measurement and analytic methods, we should expect more
consistent empirical results? (p. 46).
11
CHAPTER 2: LITERATURE REVIEW
Introduction
In subsequent pages, previous empirical findings regarding religiosity, family
processes, and deviance are discussed in depth. First, an important groundwork is laid by
reviewing the literature dealing with the relationship between religiosity and deviance.
Next, the role of family processes in relation to both adolescent deviant behavior and
religiosity is examined. At this point, a theoretical framework is discussed which may
provide a helpful heuristic for the examination of the interplay between these
variables ? family processes, religiosity, and deviance. Next, the literature dealing with
such a linkage between these three concepts is detailed. Finally, three additional
potentially important intervening factors (type of deviance, developmental issues, and
measurement) are briefly reviewed. At the end of this literature review, the hypotheses for
the current study will be enumerated and a detailed rationale for the current study will
also be provided.
Religiosity and Deviance
Attempts to link religiosity to deviance are far from new in social research.
Many early sociologists believed that religion played a role as a conforming influence in
society. For example, describing the role of religion in society, Davis and Moore (1945)
wrote,
12
Through the worship of the sacred objects and the beings they symbolize, and the
acceptance of supernatural prescriptions that are at the same time codes of
behavior, a powerful control over human conduct is exercised, guiding it along
lines sustaining the institutional structure and conforming to the ultimate ends and
values. (p. 245)
Because of this firm belief that religion provided a ?powerful control,? some
scholars urged that it was not possible to find and implement an effective cure for crime
that failed to integrate the instruction of religious principles (Kalmer & Weir, 1936;
Travers & Davis, 1961). In fact, some social scientists (e.g., Coogan, 1952) felt so
strongly about the role of religion in preventing deviance that their work focused on
vehemently opposing anyone who would too quickly dismiss the importance of religion
as a deterrent to deviance. Even politicians (e.g., President Eisenhower) and influential
criminal justice officials (e.g., J. Edgar Hoover, Director of the FBI) spoke in favor of the
need for religious control over crime and delinquency (Coogan, 1954). However, on the
other side of the debate, some actually attributed to religion the blame for the delinquent
behavior of youth (see Gannon, 1967b, pp. 418-419, 430; see also Barron, 1954; Taft,
1956; Tappan, 1949). Unfortunately, the unsubstantiated claims of politicians and law
enforcement officials and the unsupported opinions of well-meaning social scientists with
no data did little to settle the debate.
13
Early Research (Pre-1969)
Studies Concluding There Is No Effect
Some early researchers did empirically examine the specific effects of religion on
the behavior of individuals, particularly deviant behavior. Several of these studies found
that religion had no significant impact on deviance. For example, in an early study on this
subject, Middleton and Fay (1941) compared eighth- and ninth-grade girls from an
institution for delinquents to those in a regular high school, both in Indiana. Interestingly,
delinquent girls not only reported being church members more than non-delinquent girls,
but also claimed to have a more favorable attitude toward church attendance. Also,
institutionalized girls indicated that they read the Bible more frequently and had a more
positive attitude toward the Bible than their high school counterparts.
Kvaraceus (1944) studied 761 delinquents (563 boys and 198 girls) to see if
church membership and/or church attendance were factors in their behavior. He found
that about 54% of these delinquents claimed to be regular attenders at church, thus
indicating that, even though half of these delinquents went to church, they were still
delinquent. The proportion of delinquents who reported regular church attendance was
found to be not significantly different from the proportion of regular church attenders
found in the general population.
In a study of 554 college students, Middleton and Putney (1962) tested the
hypothesis that, if differences existed in levels of delinquent behavior between the
religious and the non-religious, they were due to differences in standards rather than
differences in the upholding of standards. In other words, these authors felt that certain
14
standards (?common social standards,? such as shoplifting, stealing, hitting, lying,
cheating) were universally held within American society and, therefore, it was not
surprising that both religious and non-religious alike conform to them. On the other hand,
other standards (?anti-ascetic standards,? such as gambling, smoking, petting, nonmarital
sex, pornography, drinking) reflect the morality adhered to only by religious people and,
consequently, only religious people care about conforming to such standards.
Results of this study indicated that, indeed, non-religious students reported higher
levels of participation in anti-ascetic actions yet were not different from religious students
in their levels of participation in anti-social activities. Furthermore, religious students
were significantly more likely than non-religious students to believe that anti-ascetic
activities were wrong, and there was no statistically significant difference between the
groups in moral beliefs concerning anti-social actions. Finally, in a test of consistency
between belief and action, it was found that both religious and non-religious participants
were equally likely to be consistent (or inconsistent) in their beliefs and actions.
These analyses were completed using a measure of religiosity which relied on a
person?s belief in a personal God. As an additional test of the strength of religiosity in
these relationships, two additional sets of analyses were completed using two different
measures of religiosity. The first was simple church attendance and the second was a
measure of the importance of religion in the person?s life. These additional analyses
replicated the same findings. Thus, based on these results, the authors concluded that
religiosity is not an important factor in predicting the types of delinquent behavior which
are most important to society as a whole (i.e., anti-social behavior).
15
Gannon (1967b) used a random sample of 150 delinquents to further explore the
impact religion might have on behavior. The sample was limited to boys of the Catholic
faith who were 14 to 16 years of age. A self-report survey was administered to boys on
the first day of their intake at a juvenile detention center. The survey included measures
of social and religious background; religious beliefs, values, and practices; and attitudes
toward stealing, sex, fighting, peer group values, and God. In addition, subsequent to the
survey, an hour-long in-depth survey was conducted with each boy to further explore
these areas.
Using frequencies, Gannon made the following findings: 1) the vast majority
(79.8%) reported attending mass once a month or more; 2) only 4 of the 150 boys said
they had never prayed while over 40% reported praying often; 3) a large majority (69%)
indicated that they had a favorable or positive attitude toward God; 4) a large majority
reported understanding and even agreeing with the church?s standards against delinquent
practices, while at the same time, participating in these activities anyway. Based on these
findings, the author concluded that religious commitment, as evidenced in the lives and
responses of these delinquent boys, was not an important deterrent to delinquency, but
rather had controlling influence only when supported by other factors of a more important
nature to the adolescent (see also Miller, 1965, who understood that juvenile delinquents
were very outwardly religious and offered possible explanations for this). According to
this study, delinquent boys had knowledge of what was wrong and understood such
behavior to be sinful in the eyes of their church, yet this knowledge did not have any real
impact on their way of living. The author concluded that this inconsistency between belief
16
and practice was characteristic of religious people in general, and that, consequently,
religiosity played a role in behavior only when convenient for the actor.
In another similar study, Gannon (1967a) studied a group of Catholic boys in a
youth correctional institution. An intensive treatment program involving weekly religious
instructional classes and monthly individual counseling sessions was implemented. Two
hundred and seventy-two boys were pretested at the beginning of the program, and 128 of
these were still available to be posttested at the end of the program. Thus, the author was
able to collect data over time indicating whether a specific intervention had been effective
in promoting religious control.
Though results seemed to indicate that the intervention program had been
effective in increasing the boys? religious knowledge and attitudes, this change did not
seem to have much impact on actual behavior. The author concluded that delinquent boys
were able to ?compartmentalize? their religious beliefs, thus effectively separating them
from their daily life. These boys were interested in stressing ritual observance of religious
practices, but were not interested in translating those observances into their actual
behavior.
Studies Concluding There Is An Effect
In contrast to these early studies which found no positive influence of religion on
behavior, others have found that such an influence does exist. For instance, Carr-
Saunders, Mannheim, and Rhodes (1944) compared delinquent and non-delinquent youth
in England in both rural and urban settings. In the city, delinquents regularly attended
religious institutions 29% of the time while non-delinquents (from the same
17
neighborhoods) attended 39% of the time. The same 10% difference was discovered
when comparing delinquents and non-delinquents in the country.
In a more extensive study, Wattenberg (1950) used official government and police
data to find out more about the impact of church attendance on delinquent boys. In a first
round of data collection in 1946, all boys picked up for offenses were interviewed.
According to these self-reports, over two-thirds of these delinquents reported attending
church regularly or occasionally, a seemingly high number. However, follow-up data
collected one year later in order to identify recidivism indicated that there was a
statistically significant difference in the patterns of church attendance between repeat and
non-repeat offenders. These data seemed to indicate that the more delinquent a young
person was, the less likely they were to be church attenders.
Travers and Davis (1961) conducted a two-part study to investigate the effect of
religion on delinquency. Their sample included 223 male, white boys. The subjects were
all Catholic and ranged in age from 10 to 17 years old. There were 120 nondelinquents
and 103 delinquents. One of the interesting aspects of this study was that religion was not
operationalized simply as church affiliation or attendance. Rather, the author developed a
measure of religious motivation. Each item in the survey presented a scenario where a
delinquent act (e.g., stealing) was imminent (e.g., a boy?s friends were about to steal
something and he was asked to join in). The multiple choice responses for each item each
presented a different motivation for completing or not completing this act. There were six
motivational concepts reflected in the responses including religion, civil authority, family
loyalty, peer group loyalty, community loyalty, and practical consequences, thus enabling
18
each boy to identify his thinking processes and the motivation behind what he would or
would not do. This represented a significant advance in the measurement of religion for
the purposes of identifying its impact on deviant behavior.
Results from the first part of this study revealed that boys classified as high in
religious motivation were significantly less likely to be delinquent. The second part of the
study more closely examined the boys classified as delinquent in an attempt to more fully
understand which factors may contribute to their attitudes toward religion. Results
showed that neither age, lack of school achievement, seriousness of offense, grade level,
nor mother?s work status contributed to the delinquents? tendency to be more or less
religious.
In an extensive study on several factors associated with delinquency, Cortes
(1965) questioned 100 delinquents and 100 nondelinquents concerning, among other
things, their religiosity. In order to control for several additional factors (age, achievement
motivation, parental occupational level, parental educational level, family structure, race,
and residence location), a smaller subset of 40 (20 from each group) subjects was selected
by matching them on these factors. This study was unique in that the parents (both
mothers and fathers) of these boys participated as well by separately completing the same
measures.
Results indicated that delinquents were significantly more likely to score lower on
the religious measures than nondelinquents. In addition, both mother and father reports
corroborated this evidence, namely that delinquent youth were significantly less religious.
19
In fact, in the entire study, the measure of religiosity was the most powerful discriminator
between delinquents and nondelinquents.
Allen and Sandhu (1967) compared 179 institutionalized delinquent boys to 198
high school attending nondelinquent boys on measures of religiosity. The samples were
fairly well-matched regarding age, family income, and race. No differences were found
between the groups on religious categorization (i.e., denominational affiliation) or
frequency of church attendance. However, findings indicated that delinquent boys
reported significantly weaker religious feelings (one item asking what kind of feelings
they had about religion, with responses ranging from ?very strong? to ?none?). An
additional finding provided some very interesting information. In multivariate analyses,
results showed that nonparticipation in church contributed to higher delinquency rates,
but for controls, the opposite was found, namely that church participation contributed to
higher delinquency rates. Though this may have been a measurement issue where
delinquent boys? delinquency was measured according to official criminal status, while
nondelinquent boys? delinquency was self-reported, the reasons for this counterintuitive
finding were unclear.
Summary of Early Research (Pre-1969)
Early research on whether or not religiosity had a measurable impact on deviant
behavior was strongly divided, with some evidence affirming the negative deterrent effect
of religiosity and other evidence not supporting it. Most of this early research was done
with legally delinquent youth (often in a comparison with non-delinquents), usually males
only, frequently Catholic. Also, most of this work relied on univariate frequencies
20
employing group comparisons rather than multivariate techniques. Studies that found
there was no effect of religiosity on delinquency used single-item measures, such as
church attendance or church membership/affiliation, while studies that employed scalar
measurement of religiosity found such an effect.
A Landmark Study: Hirschi and Stark (1969)
As is clearly seen, there has been considerable debate on this topic among
academics for at least a century now. The most common strain of this argument centers
around a study published by Hirschi and Stark (1969). In this study, the authors were
concerned with what they believed to be misguided confidence in the widely accepted
assumption that participation in religious activities provided a powerful social control
against deviancy. This belief, they felt, had been ?supported by clergymen? (p. 202) and
was ?congruent with the social scientific view that religious sanctioning systems play an
important role in ensuring and maintaining conformity to social norms? (p. 202; see also
Davis, 1948, who made a similar connection). Hirschi and Stark were not so much
challenging the idea of social controls (after all, Hirschi authored his own form of social
control theory), but questioning whether the ?Christian sanctioning system of hellfire for
sinners and heavenly glory for the just? (p. 203) qualified as such an effective social
control.
Hirschi and Stark?s (1969) sample included a little over 4,000 junior and senior
high school students from public schools in California (i.e., the Richmond Youth Study).
Their measures included a 6-item self-report index of delinquency, an official measure of
delinquency based on police records, a single-item measure of church attendance, 2 items
21
measuring morality, 2 items measuring attitudes toward the law, and a 2-item self-report
index of belief in the supernatural. Analyses utilizing these measures employed a simple
frequency technique. Cross-tabulations of each independent variable with church
attendance were performed and the gamma statistic (a non-parametric measure of
bivariate association) was used to determine statistical significance of the results.
Results showed that church attenders were not significantly more likely to
condemn statements of questionable morality, were ?only slightly? significantly more
likely to express respect for police and the system of law, and were significantly more
likely to hold supernatural beliefs. In turn, results indicated that those who endorsed
questionable morality and/or had less respect for the legal system were significantly more
likely to be delinquent. Furthermore, findings revealed that those who held supernatural
beliefs were just as likely to commit delinquent acts as those who did not. In a final
analysis, church attendance was found to be completely unrelated to delinquency. In other
words, those who attended church were just as likely to commit acts of delinquency as
those who did not.
These results provided unequivocal evidence, in the minds of the authors, that
?Participation in religious activities and belief in a supernatural sanctioning system have
no effect on delinquent activity? (p. 211). This was the ?shot heard ?round the world? of
social science. For many, if not most, this was considered the final word on the matter.
For some reason, though earlier empirical work on this subject had produced mixed
results, Hirschi and Stark?s findings received greater prominence. After this, Hirschi and
Stark (1969) became the standard citation proving the idea that religiosity had no effect
22
on deviant behavior. However, both antecedent and subsequent literature provided a
different picture.
Specific Responses to Hirschi & Stark (1969)
Since the publication of this influential study, researchers have been consistently
retesting the relationship between religion and deviance using various types of samples
and a wide variety of measures. Many of these studies have been published in direct
response to Hirschi and Stark by tying in to the key word from their title (i.e., ?hellfire?).
For instance, Burkett and White (1974) attempted to take ?another look? at the ?hellfire?
hypothesis. These authors hypothesized that the reason Hirschi and Stark found no
relationship was because they measured more serious forms of delinquency
(e.g., vandalism and assault) which are condemned not only by religion but also by
society at large (basically a restatement of Middleton & Putney?s [1962] original
?antiascetic hypothesis?). Therefore, according to these authors, religiosity would be no
more effective as a deterrent to such activities than any other social control. However,
these authors hypothesized that religion might be more important in the deterrence of
delinquent acts which are condemned by churches yet condoned by secular leaders
(e.g., alcohol and marijuana use).
In this study, 855 high school seniors (all Caucasian) in the Pacific Northwest
were surveyed using the exact same measures of morality, worldly authority, supernatural
sanctions, and delinquency (specifically larceny, vandalism, and assault) from Hirschi and
Stark (1969). In addition, frequency of alcohol and marijuana use were measured. Results
of this study indicated that church attendance was only weakly related to delinquency as
23
measured by Hirschi and Stark (i.e., larceny, vandalism, and assault) but was more
closely related to alcohol and marijuana use. For these authors, the evidence seemed to
suggest that religiosity was a more effective deterrent against victimless crimes than
against victim crimes. This was interpreted as meaning that churches teach a different
form of morality encompassing substance use which was no longer vigorously
condemned by other contemporary social structures.
Higgins and Albrecht (1977) hypothesized that the lack of a finding in Hirschi and
Stark may have resulted simply from geographic location. In fact, both Hirschi and
Stark?s data as well as Burkett and White?s were collected in the Pacific Northwest, an
area that is known to be less religiously oriented. Higgins and Albrecht, on the other
hand, collected their data in Atlanta, Georgia, and proposed that religiosity may be a more
important part of the average Southeast adolescent?s experience.
One thousand four hundred and ten tenth-grade students from six Atlanta high
schools were asked to indicate how often they attended church or religious services. In
addition, delinquent behavior was measured using a self-report checklist of 17 items
ranging from coming home later than midnight to stealing a car. Finally, respect for the
juvenile court system was assessed for both the study participant and their friends with a
four-item scale. Church attendance was found to be strongly negatively related to each of
the delinquency items, both for the total sample (gamma = -.48 for the composite
measure) as well as all four race-by-sex subgroups (white males, nonwhite males, white
females, nonwhite females). Also, church attendance was positively related to both self?s
and friends? respect for the juvenile system which were both, in turn, negatively related to
24
delinquency. These findings provided some evidence that the relationship between
religiosity and deviance varied by geographical region since church attendance, which
was not related to delinquency in California, was strongly related to delinquency in
Atlanta.
It is important to note several differences of this study from previous studies.
First, this study revealed a moderate to high negative correlation between church
attendance using the very same measures of delinquency employed by both Hirschi and
Stark (1969) for which they found no relationship, and Burkett and White (1974) for
which they found only a weak relationship. Second, this strong relationship was found not
only for victimless crimes (e.g., skipping school, reckless driving), but also for several
more serious offenses involving victims (e.g., using force to steal someone?s money,
fighting, vandalism). This was inconsistent with Burkett and White?s findings and
conclusions.
Data from 3,268 Arizona high school students were used by Jensen and Erickson
(1979) to take yet another look at the ?hellfire hypothesis.? A single item measure of
church attendance was utilized along with measures of youth church activity participation
and religious affiliation (Catholic, Protestant, Mormon). Hirschi and Stark?s (1969)
measure of supernatural beliefs was also used as well as a measure of delinquency
involvement.
Their findings suggested that church attendance was consistently related to a wide
variety of self-reported delinquent activities rather than just ?victimless? ones. This study
also found evidence for differential effects of church attendance on delinquency by
25
religious affiliation; namely, those who consistently attended churches with higher ascetic
values were less delinquent. In addition, results seemed to indicate that religiosity played
a lesser role in the deterrence of deviance among metropolitan or urban youth as opposed
to rural youth. Finally, regression analyses showed that religiosity explained additional
variance (from .8% to 3.5%) in delinquency, even after holding constant such social
control items as activity participation, school bonds, parent bonds, and legal bonds. In
fact, this explanatory power was substantially greater for those who were affiliated with
churches with more strict moral teachings (13.6% to 21.5%). As a further test of these
results, Jensen and Erickson obtained Hirschi and Stark?s original data and sought to
replicate some of these findings. Indeed, further analyses revealed that church attendance
was significantly related to drinking, smoking, and truancy in these data. In addition, high
levels of regular church attendance at churches with more strict moral teachings was
found to be significantly negatively related to delinquency.
In Stark, Kent, and Doyle (1982), one of the coauthors of the original 1969 study
revisited the topic. Equipped with insights from intervening literature which has just been
reviewed (e.g., Burkett & White, 1974; Higgins & Albrecht, 1977), Stark and colleagues
hypothesized that ?religious ecology? was the reason for the lack of findings in the earlier
paper. This hypothesis was largely dependent on Higgins and Albrecht?s speculation that
geographical location was the source of the variation in the relationship between
religiosity and delinquency. More specifically, Stark et al. (1982) proposed that church
attendance would only deter deviance to the extent that religious beliefs were pervasive in
the community in which the individual resided. In other words, because the western area
26
of the country (where Hirschi and Stark?s data had been collected) could be considered
largely ?unchurched,? the effect was undetectable, thus reflecting the fact that religious
values are not important to people who live in a nonreligious or largely secularized social
environment.
As predicted, these authors found that 1) church attendance was strongly related to
both official (gamma = -.46) and self-reported (gamma = -.45) delinquency in a sample
from a ?moral community? (Provo, Utah); 2) church attendance and importance of
religion were only weakly related to measures of recent (gammas = -.08, -.14) and
lifetime (gammas = -.13, -.14) deviance in a ?secular community? (Seattle, Washington);
and 3) church attendance, importance of religion, and religious values were all strongly
related (gammas ranging from -.19 to -.29) to three different measures of delinquency in a
national sample of Caucasian boys. As a final test of their hypothesis, Stark and
colleagues isolated three of the secular communities in their data that were from
California and exhibited very similar church attendance rates to the original Richmond
Youth Study data. Using this data, they replicated Hirschi and Starks? original findings of
no effect of church attendance or religiosity on deviance.
The finding that religiosity was strongly related to delinquency in a national
sample was viewed by the authors as evidence that, ?For the nation as a whole, religion
serves to undergird the moral order? (p. 14). However, they interpreted this finding to
mean that personal religious convictions only constrain deviance in a social environment
where religiosity is positively reinforced.
27
Cochran and Akers (1989) responded directly to this line of research when they
examined the more specific relationship between religiosity and substance use (marijuana
and alcohol). In a sample of 3,065 male and female adolescents in the seventh through
twelfth grades from seven school districts in three Midwestern states, it was found that
three measures of religiosity were all significantly negatively related to alcohol use,
explaining close to 9% of the variance after controlling for age, race, gender, and SES,
while two of these measures were significantly negatively related to marijuana use
(explaining over 5% of the variance). These authors concluded that religion has
consistently moderate ?demonstrable potent effects? (p. 221).
Stark (1996) attempted to alleviate the confusion that had developed over the
?contextual theory,? or the idea that the effect of religiosity on deviance varies by
geographical region or sociological context. This time, the sample came from a national
study conducted in 1980 that included close to 12,000 seniors in high school. Indeed,
results indicated that a rather strong relationship existed between church attendance and
trouble with the law in all regions of the country (gammas ranging from -.23 to -.39)
except the Pacific West (gamma = -.02). In addition, this same effect was found for males
(gamma = -.30), females (gamma = -.33), whites (gamma = -.39), and blacks
(gamma = -.29) in the non-Pacific regions.
The literature just reviewed represents only a subset of the literature dealing with
the relationship between religiosity and deviance. Its importance lies in the fact that it is
the most consistent strain of thinking representing an ongoing dialogue among social
scientists concerning the ?Hellfire Hypothesis? as originally proposed by Hirschi and
28
Stark (1969). In parallel with this line of work, there has been a large body of literature
dealing with this same bivariate relationship. Several researchers have undertaken
thorough reviews of this literature. These reviews will be examined in the next section.
Other Important Literature
There have been some important studies which have attempted to investigate and
summarize this entire body of literature. For instance, Tittle and Welch (1983) searched
the literature available at their time (20 years ago now) and found that there were ?fewer
than two dozen research reports in which the effect of individual religiosity on deviant
behavior has been the primary focus of attention? (p. 654). In addition, they found a total
of 65 published studies that had reported some empirical evidence regarding the
relationship between a measure of religiosity and some indicator of rule-breaking. Most
importantly, a perusal of these studies revealed that only 10 of the 65 (15%) failed to
report a significant negative relationship between religiosity and deviance. Based on this
evidence, these authors concluded that ?the evidence seems remarkably consistent in
suggesting that religion is related to deviant behavior. Indeed, only a few variables in
social science have proven to be better predictors of rule breaking? (p. 654).
More recently, Johnson et al. (2000) sought to summarize the extant research
literature utilizing an innovative method called a ?systematic review? (SR). This review
technique adds a quantitative element to the traditional literature review approach, thus
making the review more scientific (i.e., less subjective) as well as replicable. For the
purposes of their review, these authors identified 40 journal articles ?that examined the
effect of religion on juvenile delinquency published from January 1985 to December
29
1997? (p. 36). More specifically, this review included studies that contained at least one
quantified variable, consisted only of juvenile and/or young adults (20 years of age or
under), were published in peer-reviewed journals, and used a sample collected in the
United States.
The measures used in this review were extensive and included the following:
1) quality of research (including such things as use of longitudinal data and specification
of race and/or gender), 2) role of religious measures (peripheral or central to study),
3) effect of religious measures (none, positive, negative, etc.), 4) dimensions of religious
measures (attendance, prayer, Bible study, etc.), 5) number of religious measures, and
6) reliability of measurement. Results indicated that 30 of the 40 studies found a
statistically significant negative effect of religiosity on deviance while only five found no
effect. In addition, only one of the studies found a positive effect and in this study,
religion was used as a control variable. This is rather strong cumulative evidence of a
consistent negative bivariate relationship between religiosity and deviance in the
literature.
This study also provided evidence concerning the importance of sound
measurement. First, results showed that studies which included a multi-dimensional
measure of religiosity were more likely to find a statistically significant negative
relationship with deviance. Second, results documented that when measure reliability was
assessed, a negative effect was always found. In fact, of the 40 studies included in the
review, only 13 checked the reliability of their measures. Interestingly, though, all 13 of
these studies found a significant negative relationship.
30
The results from this study are an important confirmation that, as a whole, the
literature dealing with religiosity and deviance provides more consensus than most have
acknowledged. In their concluding remarks, Johnson and colleagues provide a strong
acknowledgment of this fact as well as a challenge to researchers who have ignored it.
From the above review, it is apparent that the role of religiosity in explaining and
understanding juvenile delinquency has been an overlooked factor in many
studies. Although religion is one of the major social institutions, most studies in
our target population of articles did not include a measure of religious
commitment or religiosity. In general, researchers do not include religious
measures in their study . . . (p. 45).
Another recent study by Baier and Wright (2001) has provided further solid
evidence of the existence of this relationship. In this meta-analysis, the authors used both
database and ancestry searches to identify 60 studies suitable for their analyses (i.e., those
that included a measure of association between religion and criminal behavior). Of these
60 studies (produced between 1962 and 1998), 56 were published while 4 were
dissertations or conference presentations. In order to obtain a uniform statistic for cross-
study comparison, a Pearson?s r correlation coefficient was calculated for any study that
did not originally utilize this statistic.
Several important findings were made in this meta-analytical review of the
literature. First, of the 60 studies identified for analysis, none indicated a positive
relationship between religion and deviance. Second, the mean reported effect size among
these 60 studies was r = -.12. Third, the null hypothesis that the mean effect for religion
31
on deviance equals zero was rejected (t = -11.9). Fourth, almost one-fourth of these
studies (13 of 60 or 22%) found a fairly substantial effect size ranging from r = -.2 to
r = -.5.
Together, these three summary studies reviewed 123 separate empirical
investigations, both published and unpublished. The findings based on this extensive
review of the literature establish that there is a consistently moderate negative
relationship between religiosity and deviance. The fact that so many of the published
studies in this area reveal significant findings is powerful since, in the aftermath of
Hirschi and Stark?s (1969) study, ?the religion-on-crime literature has viewed
nonsignificant findings as important contributions? (Baier & Wright, 2001, p. 7). In other
words, in many, if not most, fields of study, only significant findings are published.
However, because of the considerable impact of a nonfinding in this area, studies
revealing a nonsignificant relationship have been viewed as equally important for
publication. In light of this evidence, it is safe to say that the negative relationship
between religiosity and deviance is an empirically established fact.
Family and Deviance
Family Processes as a Dimensional Construct
Thus, it appears that religion plays an important role in the prediction of deviance.
In addition to religion, family has also been identified as playing an influential role in
understanding deviant behavior. In fact, though researchers have found evidence for
multiple etiological factors related to adolescent deviance, a central role has been
assigned to the influence of family (Barber & Rollins, 1990; Cernkovich & Giordano,
32
1987; Geismar & Wood, 1986; Hirschi, 1969; Loeber & Stouthamer-Loeber, 1986;
Snyder & Patterson, 1987; Zhang & Messner, 1995). Numerous familial factors have
been investigated, some more peripheral than others. For instance, both familial deviance
(Baker & Mednick, 1984; Farrington, Gundy, & West, 1975; Lauritsen, 1993; Moffitt,
1987; Nagin & Farrington, 1992; Rutter & Giller, 1984; Sampson & Laub, 1993; West &
Farrington, 1973, 1977) and family structure (Dornbusch et al., 1985; Free, 1991; Loeber
& Stouthamer-Loeber, 1986; Rankin, 1983; Rebellon, 2002; Steinberg, 1987; Wells &
Rankin, 1985) have been found to have some importance.
However, most of the literature attempting to link family with deviance has
focused on fluid functioning rather than static structure. Researchers have hypothesized
that the most interesting connection between family and deviance emphasizes family
processes involving behavioral interactions between family members. This process is
often conceptualized as consisting of several dimensions. Attempts to explore the
dimensionality of parenting have resulted in a number of conceptualizations (Maccoby &
Martin, 1983). Some of these have contained two variables on a continuum which, when
overlapped, created a four-fold typology (warmth/hostility and control/autonomy,
Schaefer, 1959; warmth (acceptance)/hostility (rejection) and restrictive/permissive,
Becker, 1964; demandingness and responsiveness, Baumrind, 1967, 1971) while others
have emphasized related sets of orthogonal factors (acceptance, firm control, and
psychological control, Schaefer, 1965; consistent discipline, maturity demands,
restrictiveness, and encouragement of independent contacts, Baumrind & Black, 1967).
33
Recent work in this area has continued to rely heavily on typological orientations
while ?unpacking? them in somewhat different ways than originally hypothesized. For
instance, Baumrind (1978) originally used the concepts of demandingness and
responsiveness to identify four types of parenting ? authoritative, authoritarian, indulgent,
and neglecting. Consequently, much of Steinberg?s work has focused on these types, and
more specifically, on authoritative parenting, which is viewed as the most optimal in this
typology (Steinberg, 2001). In recent work, however, Steinberg has sought to enumerate
the contents of authoritative parenting as more than simply ?demandingness? and
?responsiveness.? More specifically, Gray and Steinberg (1999) state that the ?three core
dimensions of authoritative parenting? (p. 574) are acceptance-involvement, strictness-
supervision, and psychological autonomy (or warmth, firmness, and psychological
autonomy granting; see Steinberg, 2001). Even more recently, Steinberg and Silk (2002;
see also Steinberg, 1990) have named these three dimensions autonomy, harmony, and
conflict. In this conceptualization, harmony is defined as warmth, involvement, and
emotional closeness while conflict includes contentiousness and hostility. Autonomy
involves the balance between parental control and adolescent autonomy and is described
as having two sub-dimensions, psychological control (control of opinions, feelings, and
thoughts; see Barber, 1996) and behavioral control (control of activities).
In an important recent test of this dichotomy of autonomy, Pettit, Laird, Dodge,
Bates, and Criss (2001) contributed evidence that may link earlier typological
conceptualizations (e.g., Baumrind) to more recent dimensional advances in the study of
parental socialization of children (e.g., Steinberg & Silk, 2002). They found that parental
34
monitoring (i.e., positive behavioral regulation) was associated with a proactive parenting
style (e.g., authoritative parenting) while psychological control (i.e., negative emotional
manipulation) was associated with a harsh parenting style (e.g., authoritarian parenting).
These findings seem to undergird the validity and efficacy of the dichotomous approach
to parental control, namely, that some types of control (i.e., behavioral control) produce
good and healthy adolescent outcomes while others (i.e., psychological control) result in
harmful internalizing and externalizing problems for adolescents.
This conceptualization emphasizing three important dimensions ? autonomy,
harmony, and conflict ? may provide a helpful outline for capturing the totality of family
functioning. Steinberg and Silk?s (2002) contention is that all research dealing with
parent-adolescent relationships can be categorized into these overarching dimensions.
Vazsonyi, Hibbert, and Snider (2003) recently used this tridimensional categorization
(harmony, autonomy, conflict) to rigorously test a multidimensional measure of family
processes. Their final model contained six parenting dimensions reflecting these three
overarching domains: 1) Harmony measured by closeness and communication,
2) Autonomy measured by support, monitoring, and peer approval, and 3) Conflict
measured by conflict. Confirmatory factor analyses (CFA?s) of this model indicated good
fit for younger and older adolescents as well as males and females. This provides solid
evidence for looking at family processes in this tridimensional manner.
35
Literature in the Autonomy Dimension
Behavioral Control
Before this literature is perused, it is important to establish a baseline of
understanding regarding the usage of terminology in this section. Traditionally, terms
such as ?parental supervision,? ?parental monitoring,? and even ?parental control? have
all been used interchangeably to refer to the role of parents in controlling their children?s
behavior. Recent work by Kerr and Stattin (2000; Stattin & Kerr, 2000) has attempted to
develop greater semantic specificity in this area by distinguishing between ?what parents
know? (knowledge) and ?how they know it? (sources of knowledge). In this line of work,
Kerr and Stattin assign the term ?monitoring? to actual parental knowledge and point out
that this monitoring (i.e., knowledge) may be passive. In other words, from their
perspective, monitoring is a passive role, implying the possession of knowledge from one
of many sources, while control and/or supervision imply activity on the part of the parent.
While such semantic distinctions may be helpful for future research, they do not reflect
the terminology employed in most research to date. Therefore, in this section, the terms
?supervision,? ?monitoring,? ?control,? and even ?discipline? will be used
interchangeably to refer to the general processes whereby parents seek to exert behavioral
control over their adolescent?s activities.
Early research on juvenile delinquency established that discipline and supervision
were significantly lacking in the homes of deviant adolescents (Glueck & Glueck, 1950;
McCord & McCord, 1959; Nye, 1958). Subsequent research has confirmed the veracity of
this robust relationship (Farrington, 1989; Riley & Shaw, 1985; West & Farrington,
36
1977). For example, Galambos and Maggs (1991) found that lack of supervision after
school (self-care) led to high rates of deviant activity, particularly for girls, whose
misbehavior sometimes exceeded that of the boys in such situations (see also Posner &
Vandell, 1994 and Steinberg, 1986). Thus, lack of parental supervision during the critical
developmentally transitional years of adolescence has consistently been found to be
related to increases in problem behaviors (Flannery, Vazsonyi, Torquati, & Fridrich,
1994; Richardson et al., 1989; Vazsonyi & Flannery, 1997) including drifts toward
antisocial peers (Patterson, Reid, & Dishion, 1992).
Work from the Oregon Social Learning Center (Patterson and colleagues) has
upheld this tenet, namely that lack of parental supervision leads directly to childhood
antisocial behavior (e.g., Patterson, Dishion, & Bank, 1984). Patterson has articulated his
theory on how this happens in his coercion model (Patterson, 1982). The core assumption
of this model is that parenting skills, including discipline and monitoring, are the specific
mechanisms by which childhood antisocial behaviors are constructed and maintained
(Patterson, Forgatch, Yoerger, & Stoolmiller, 1998; Stoolmiller, Patterson, & Snyder,
1997). In a test of this model, Forgatch (1991) used structural equation modeling to show
that parental discipline and monitoring accounted for 30%-52% of the variance in
antisocial behaviors in three different at-risk samples. Often, the disrupted parenting
practices resulting in antisocial behavior are a result of ineptness rather than intentionality
(e.g., Wahler & Dumas, 1986). In other words, parents who are incompetent in their
attempts at discipline and erratic in their expressions of anger are unintentionally
promoting antisocial behavior (Kazdin, 1987).
37
Patterson et al. (1998) found that parenting variables (i.e., discipline and
supervision) were a significant proximal influence on deviant behaviors recorded at three
different points of time ? antisocial childhood behavior, early arrest, and chronic juvenile
offending ? thus indicating a trajectory-like processes for which parenting is a primary
socialization factor. Similarly, Vuchinich, Bank, and Patterson (1992) found that parental
practices influenced antisocial behavior even when controlling for previous antisocial
behavior. In fact, the addition of the causal path from parenting to behavior reduced the
estimated stability of antisocial behavior by 16%. The direct causal nature of parenting
with regard to youthful deviance has also been confirmed in prevention and intervention
trials employing random assignment designs (Chamberlain, Ray, & Moore, 1996;
Dishion, Patterson, & Kavanagh, 1992). Similarly, successful clinical trials have been
shown to reduce antisocial behaviors through the strengthening of parental practices
(Dadds, Schwartz, & Sanders, 1987; Eisenstadt, Eyberg, McNeil, Newcomb, &
Funderburk, 1993; Patterson, Chamberlain, & Reid, 1982; Patterson, Dishion, &
Chamberlain, 1993; Webster-Stratton, 1984).
Researchers in the Child Development Project (Bates, Dodge, Pettit, and
colleagues) have also examined the importance of parental monitoring on child
behavioral outcomes (e.g., Pettit, 1997). Their work has found evidence for the fact that
after-school activity lacking in parental (or any adult) supervision is a powerful risk factor
in the development of externalizing problem behaviors (e.g., Pettit, Laird, Bates, &
Dodge, 1997) as is non-parental childcare in general (Bates et al., 1994). Furthermore,
they have documented that parenting practices not only contribute uniquely to
38
externalizing problems, even when child, peer, and sociocultural factors are considered
simultaneously (Deater-Deckard, Dodge, Bates, & Pettit, 1998), but also that parental
monitoring moderates a number of other etiological factors, including peer activity and
neighborhood safety (Pettit, Bates, Dodge, & Meece, 1999). In general, research from this
group has reinforced the idea that negative parenting (e.g., harsh, hostile, punitive)
contributes to youth deviance (e.g., Deater-Deckard & Dodge, 1997; Deater-Deckard,
Dodge, Bates, & Pettit, 1996; Pettit & Dodge, 1993; Weiss, Dodge, Bates, & Pettit, 1992)
while positive parenting (e.g., responsive, affectionate, proactive) is associated with lower
levels of behavioral problems (e.g., Bates, Bayles, Bennett, Ridge, & Brown, 1991; Pettit
& Bates, 1989; Pettit, Bates, & Dodge, 1993; Pettit & Dodge, 1993; Pettit, Harrist, Bates,
& Dodge, 1991).
Psychological Control
The distinction between behavioral and psychological control was highlighted
years ago by Schaefer (1965), but was not subsequently pursued by social scientists until
Steinberg and colleagues revived the subject (Steinberg, 1990; Steinberg, Elmen, &
Mounts, 1989). Most recently, Barber?s (2002) edited book has brought the issue to the
forefront of the study of parenting. In his earlier review of literature related to this
construct, Barber (1996) describes parental psychological control as ?psychologically
controlling processes involv[ing] socialization pressure that is nonresponsive to the
child?s emotional and psychological needs (Maccoby & Martin, 1983), that stifles
independent expression and autonomy (Baumrind, 1965, 1978; Hauser, 1991; Hauser et
al., 1984)? (p. 3299). Though the construct of parental psychological control continues to
39
evolve amidst debate over its conceptualization (Barber, Olsen, & Shagle, 1994),
empirical evidence indicates that it is an important positive predictor of adolescent
problem behavior (Barber, 1996; Steinberg, 1990).
Although the specific components of psychological control are still somewhat
unclear, the general consensus is that it relates to parental intrusiveness on a child?s
emotional development and includes elements of parental support and approval (or the
lack thereof, e.g., love withdrawal; Barber, 1996; Pettit et al., 2001). Thus, an appropriate
lack of psychological control may involve positive evaluations of the teen from the parent
(Burbach & Bourdin, 1986), acceptance of the teen?s ability to make autonomous
decisions (e.g., choice of friends; Hauser, 1991; Hauser et al., 1984), and encouragement
of healthy interactions outside the family (Baumrind, 1965, 1978; Youniss & Smollar,
1985). Such a parenting style  would be more conducive to healthy emotional
development in the adolescent (Steinberg, 1990). On the other hand, a discouraging,
nonaccepting, disapproving parental style has been shown to carry dire behavioral
consequences (Anderson, Hinshaw, & Simmel, 1994).
Research has found that psychological control processes are more strongly related
to internalizing than externalizing behaviors (Barber, 1992, 1996; Baumrind, 1965, 1968,
1978; Becker, 1964; Coopersmith, 1967; Maccoby & Martin, 1983; Steinberg, 1990).
However, among those studies that have undertaken an examination of the relationship
between parental or familial support and adolescent deviant behavior, the overwhelming
majority have shown that growing up in an emotionally supportive family is predictive of
lower risk for delinquency and drug abuse (Brook, Brook, Gordon, Whiteman, & Cohen,
40
1990; Johnson & Pandina, 1991; Kandel, Kessler, & Margulies, 1978; Wills & Vaughn,
1989; Windle & Miller-Tutzauer, 1992). In a recent measurement article, Tolan, Gorman-
Smith, Huesmann, and Zelli (1997) established that familial psychological support played
an important role in the multi-dimensional measurement of family processes and that this
measure, in turn, significantly predicted both externalizing (i.e., aggression) and
internalizing (i.e., depression) problems (see also Dekovic, 1999). In fact, support was
one of only two scales in the 6-scale model that significantly predicted both types of
behavioral problems (see also Vazsonyi et al., 2003).
Research has also been done in this area from a negative perspective. In other
words, the above research focuses on the positive outcomes (less deviance and substance
use) associated with healthy emotional support from parents or family. However, there
also exists research dealing with the harmful effects of parental psychological overcontrol
(Barber, 1992). For example, Stierlin (1974) observed that parents of teens in
psychologically dysfunctional families interfered with the appropriate development of
autonomy by utilizing specific types of interactional strategies designed to hinder healthy
individuation and/or differentiation, thus creating severe behavioral problems (see also
Hauser et al., 1984). In addition, Costanzo and Woody (1985) have suggested that
intrusive psychological control can emotionally and socially cripple an adolescent,
leading ultimately to the necessity of employing alternative strategies (e.g., antisocial
behavior) for the accomplishment of goals. Previous research focusing specifically on the
authoritarian parenting style implies that overcontrolling parents produce children lacking
in social skills (Maccoby & Martin, 1983). For instance, Baumrind (1967; Baumrind &
41
3 Several terms have been used interchangeably to refer to this component,
including affect, warmth, acceptance, and responsiveness (Rothbaum & Weisz, 1994).
Similarly, these terms will be used interchangeably in this paper.
Black, 1967) found that socially withdrawn nursery school children tended to have
parents fitting the authoritarian description.
Literature in the Harmony Dimension
Emotional Closeness3
In addition to parental control, the affective component of the parent-child
relationship has been identified as a possible contributor to the development of
inappropriate behavioral patterns. Low levels of parental affection have been found to be
associated with greater amounts of adolescent deviance (Gove & Crutchfield, 1982;
Henggeler, 1989; Hurrelman, 1990; Tolan & Lorion, 1988). Apparently, adolescents
whose interactions with parents are characterized by high levels of affection develop a
perception of being supported and, consequently, those in supportive families have been
found to be at a lower risk for various types of deviant behavior (Johnson & Pandina,
1991; Kandel, Kessler, & Margulies, 1978; Sokol-Katz, Dunham, & Zimmerman, 1997).
The importance of the affective dimension of parenting on children?s behavior has
been found to emerge in very early life. In fact, Sroufe, Egeland, and Kreutzer (1990)
presented compelling evidence that attachment behaviors such as proximity-seeking and
primary caregiver contact measured in the first two years of life were predictive of
behavioral problems several years later. Also, clinically treated preschool children,
referred because of disruptive behavioral problems, have been found to be more likely to
display behaviors associated with insecure attachment (Greenberg, Speltz, DeKlyen, &
42
Endriga, 1991; Speltz, Greenberg, & DeKlyen, 1990). In addition, Sutton, Cowen, Crean,
and Wyman (1999) found that low parental warmth was significantly related to
aggressive behavior as early as the second and third grades. There seems to be some
indication that insufficient attachment is a greater risk for those children who already
faced other risky factors, such as having a young, single mother (Egeland & Sroufe, 1981;
Erickson, Sroufe, & Egeland, 1985; Sroufe, 1983).
Research has shown that this is a rather stable pathway leading from early infancy
into adolescence and beyond (Greenberg, Speltz, & DeKlyen, 1993). While of school age,
attachment to family has been found to reduce the overall frequency, prevalence, and
intensity of involvement in deviant activities, regardless of neighborhood, ethnicity, or
gender (Dornbusch, Erickson, Laird, & Wong, 2001). It also seems that emotional
attachment to parents may be more salient earlier in life and for younger compared to
older adolescents (Corwyn & Benda, 2001). There is also some evidence for the fact that
the effects of parental warmth on a child?s behavior accumulate over time. For instance,
Scaramella, Conger, and Simons (1999) used longitudinal data from 319 adolescents and
their parents over a five year period to examine the role of parental warmth in inhibiting
the growth trajectories of externalizing behavioral problems. Their results clearly
indicated that affection received from parents provided a buffer against the development
of antisocial behavior. More importantly, this buffering effect seemed to prevent the
accumulation of misbehavior in an additive fashion through the years.
There is also evidence that warmth and involvement from parents is a strong
enough factor in adolescent deviant behavior as to even explain away any negative impact
43
associated with harsh corporal punishment. In a very important study, Simons, Johnson,
and Conger (1994) found that, after the effect of parental involvement was removed, no
detrimental effect of harsh corporal punishment remained on adolescent aggressive
behavior or general delinquency. These findings seemed to indicate that overly punitive
parental discipline was simply a correlate or symptom of a deeper problem, namely lack
of consistent parental warmth. This is very important evidence for the crucial role played
by parental emotional closeness in the deterrence of deviance.
Not only is positive affect an important protective factor against deviant behavior,
but negative affect has also been found to place children and adolescents at a greater risk
for such. Previous research has shown that a predominance of negative parental affect is
associated with such behaviors as defiance or hostile aggression in their children (Grusec
& Lytton, 1988; Olweus, 1980). In a study among African-American female adolescents,
Pittman and Chase-Lansdale (2001) found that those whose mothers were measured as
being low on parental warmth were the most at risk for a variety of externalizing
problems (see also Gray-Ray & Ray, 1990). Specifically in relation to substance abuse,
young people who report strong, positive relationships with their parents also indicate
lower levels of substance use (Barnes, 1984; Barrera, Chassin, & Rogosch, 1993;
McDonald & Towberman, 1993; Selnow, 1987). For example, Volk, Edwards, Lewis,
and Sprenkle (1989) found that non-drug-abusing adolescents reported significantly
higher levels of family cohesion (i.e., lower levels of disengagement) than their drug-
abusing peers, thus indicating that positive affect from parents was, in itself, an important
protective factor against substance use (Fletcher & Jefferies, 1999; White, Johnson, &
44
Buyske, 2000). In a meta-analytic summary of 47 studies, Rothbaum and Weisz (1994)
confirmed that, as a general rule, measures of parental caregiving are negatively
associated with children?s externalizing behaviors. Furthermore, in a unique meta-
analytic factor analysis, a ?pattern of caregiving? tapping into a construct they termed
acceptance-responsiveness was identified that demonstrated an even stronger deterrent
effect on antisocial behaviors.
Thus, the type of home where the appropriate balance of affection and supervision
has been achieved and consistently maintained (i.e., authoritative; Baumrind, 1987) has
been identified as the optimal setting for the deterrence of adolescent misconduct (Loeber
& Dishion, 1984; Maggs & Galambos, 1993; Patterson & Stouthamer-Loeber, 1984;
Steinberg, Mounts, Lamborn, & Dornbusch, 1991; Vuchinich et al., 1992). For example,
Baumrind (1991) found that parents who were nonauthoritarian yet still directive (that is,
they maintained an adequate balance of love and discipline) were able to successfully
shield their children from dysfunctional, risk-taking behavior more so than other types of
parents (see also Baumrind, 1989).
Communication
Because a large amount of the literature dealing with family processes and
deviance has examined the role of a typological parenting style classification system in
adolescent behavioral outcomes (Durbin, Darling, Steinberg, & Brown, 1993; Lamborn,
Mounts, Steinberg, & Dornbusch, 1991; Steinberg, Lamborn, Darling, Mounts, &
Dornbusch, 1994; Steinberg, Lamborn, Dornbusch, & Darling, 1992; Steinberg et al.,
1991), it is difficult to identify the specific effects on deviance directly attributable to
45
patterns of parent-teen communication. Typically, these studies include measures of
parent-adolescent communication that are either confounded with other parenting
dimensions or are not broken out from the authoritative parenting construct so that unique
contribution may be examined. For example, in an examination of authoritative parenting
on adolescent behavior, Slicker (1998) included measures of behavioral control,
acceptance, and democracy. One of the behavioral control items was, ?How much do
your parents try to know what you do with your free time?,? while another item intended
to measure acceptance read, ?My parents spend time just talking to me,? and a third item
attempted to measure democracy by stating, ?My parents tell me that their ideas are
correct and that I should not question them? (p. 353). It is clear, of course, that all three of
these items contain a very strong communication component and touch on the quality of
communication between parent and adolescent, even though none are specifically
intended to measure a construct called ?communication.?
Thus, in examinations utilizing a typological orientation, specific parenting
behavioral dimensions, such as communication, are not normally investigated
individually because these researchers are interested only in general patterns or climates
of parenting formed by clusters or constellations of behaviors rather than the relationship
of specific dimensions to specific outcomes (Steinberg, Darling, & Fletcher, 1995). This
is also true in clinical research utilizing typologies such as the circumplex model (Olson,
1986, 2000; Olson et al., 1983; Olson, Russell, & Sprenkle, 1983; Olson, Sprenkle, &
Russell, 1979) or the Beavers systems model (Beavers, 1981, 1989; Beavers & Hampson,
1993).
46
Another aspect of this literature which makes it difficult to tease out the effects of
verbal communication on deviant behavior is the use of broad constructs incorporating
aspects of communication as well as other dimensions. For example, Pettit and Bates
(1989) used an observational coding system to identify important mother-child dyadic
interactions. One of their measures, called proactive maternal involvement, combined
both social contact and teaching behaviors. Certainly teaching is facilitated by verbal
communication, as is the initiation of social contact, but non-verbal behaviors would be
inherently included as well. Thus, though their results indicated that children with less
problem behaviors had more positive interactions with their mothers and that this
relationship was predictive over time, it is unclear what specific role maternal
communication played in such proactive involvement patterns. Similarly, a recent study
by Lynch (2001) included a measure called family interaction/bonding. Parental
monitoring, parental caring, and parental communication were all three included in this
single latent construct.
An examination of a study by Forehand, Miller, Dutra, and Watts-Chance (1997)
clearly illustrates this snag in the literature. Their study is one of the few that includes a
separate and distinct measure of parent-adolescent communication. However, their
literature review relies exclusively on studies that did not. For instance, Weintraub and
Gold (1991) studied parental affection, Gray-Ray and Ray (1990) examined parental
rejection, and Smith and Krohn (1995) looked at parental involvement, yet Forehand et
al. called all of these ?communication.?
47
Because of these potentially confounding issues, the literature examining the
unique effect of parent-adolescent communication on deviant behavior is limited. Those
studies that have done so have produced mixed and rather weak results. Forehand et al.
(1997) found that, in four different samples of youth from different geographical
locations, parent-adolescent communication was significantly positively correlated with
parental monitoring and significantly negatively correlated with deviance. However, in a
set of hierarchical regression analyses, only monitoring maintained a significant
relationship with deviance after several demographic variables were controlled. In
addition, no interaction terms (e.g., communication and monitoring, communication and
age, communication and sex) were significant. These findings seem to indicate that,
though communication may be a family processes that is important in the prevention of
deviant behavior, its importance diminishes, or even vanishes, when other, more
important dimensions of family functioning (e.g., monitoring) are taken into account.
However, other studies have presented different findings. For example, a
longitudinal study, such as that conducted by Ellickson and Morton (1999), carries with it
considerable weight concerning the causal role of certain dimensions of family processes
in the formation (or prevention) of deviant behavior. In this study, data concerning the
target adolescents? communication with parents were collected during their seventh grade
year in school. Three years later, in tenth grade, data were collected concerning
behavioral outcomes. Their results presented compelling evidence that parent-adolescent
communication played a significant predictive role in adolescent deviant behavior. More
specifically, adolescents who felt comfortable talking to a parent about personal problems
48
and did so regularly in seventh grade were significantly less likely to begin to use hard
drugs by the time they were in tenth grade. On the other hand, those who did not report
good communication with a parent in seventh grade were at an increased risk for hard
drug usage by tenth grade. Interestingly, this predictive relationship was found for several
different ethnic groups, including White, Hispanic, and Asian youth. In a similar
affirmation of the important role of parent-adolescent communication in the deterrence of
deviance, Leas and Mellor (2000) found that, though a multi-factorial measure of parental
attachment did not significantly predict juvenile delinquency, one of the subscales,
namely parental communication, did.
Literature in the Conflict Dimension
There is a large literature dealing with the fact that parental abuse of children
leads to externalizing (George & Main, 1979; Hoffman-Plotkin & Twentyman, 1984;
Howes, 1988; Howes & Eldredge, 1985) and deviant (Gray, 1988; Straus, 1991; Widom,
1989) behaviors. Indeed, previous research has consistently documented that individuals
who were subjected to physical maltreatment during childhood are at risk for the
commitment of delinquent behaviors in adolescence (Eckenrode, Laird, & Doris, 1993;
Sternberg et al., 1993). Recently, Benda and Corwyn (2002) found that abuse during
adolescence (at the age of 12 or older) was significantly related to violent behavior among
both younger (less than 16 years old) and older (16 years old and older) adolescents and
that this relationship was not mediated by any demographic (gender, race, father?s
education, family structure) or theoretical (five social control, three strain, and four social
learning) variables. In addition, they showed that abuse during childhood (before the age
49
of 12) was also positively related to violent behavior for both age groups and was not
mediated for the younger group. Similarly, Keiley, Howe, Dodge, Bates, and Pettit (2001)
found a direct link between the physical maltreatment of a child by a parent and
externalizing behaviors. More specifically, this longitudinal study used growth modeling
to reveal that the earlier a child was abused, the more likely they were to exhibit
adjustment problems in early adolescence.
In addition, we know that interparental conflict and/or marital discord is also
harmful to adolescent development (Vandewater & Lansford, 1998). Rutter and Giller
(1984) noted that the relationship between ?broken homes? and delinquency may be
largely explained by the presence of marital discord (see also Penning & Barnes, 1982) to
the extent that there may not even be a direct link otherwise (Wilson & Hernstein, 1985).
Also, interparental conflict has been found to be significantly associated with antisocial
behavior in children even when their home was unbroken (Porter & O?Leary, 1980). In
studies involving only children from broken homes, the extent of parental marital discord
has been found to be associated with the likelihood of deviant behavior (Hetherington,
Cox, & Cox, 1979; Wallerstein & Kelly, 1980).
However, child abuse or maltreatment and interparental conflict are not the same
as parent-child conflict. Though early adolescent theorists (e.g., Freud, 1962; Hall, 1904)
viewed adolescence as a period consumed with inevitable parent-child turbulence (?storm
and stress?), more recent work has recognized that conflict with parents during this time
is less frequent and less traumatic than originally felt (e.g., Steinberg, 1990) and normally
involves the mundane issues of daily life (Montemayor, 1983). However, where parent-
50
adolescent conflict does exist, it seems to be a risk factor for negative behavioral
outcomes (Shek & Ma, 2001; Tomlinson, 1991), though there is some disagreement on
this matter (e.g., Buysse, 1997). Research has widely recognized parent-adolescent
conflict as an important etiological factor in the development of psychological
maladjustment (Reich, Earls, & Powell, 1988; Foster & Robin, 1988; Hall, 1987), such as
depression (Forehand et al., 1988) and anxiety (Slater & Haber, 1984). Such conflict has
also been linked to externalizing problems at school (Forehand, Long, Brody, & Fauber,
1986), status offenses (Adams, Gullotta, & Clancy, 1985; Justice & Duncan, 1976), and a
general antisocial tendency (Knight, Broome, Cross, & Simpson, 1998). Kim,
Hetherington, and Reiss (1999) found that conflict with mothers (more so than fathers;
cf. Shek & Ma, 2001 who found that conflict with fathers was more influential than with
mothers) had a strong positive relationship with externalizing behaviors in adolescence
for both boys and girls from both intact and stepfamily homes. In fact, these researchers
noted that such conflict contributed more substantially to adolescent problem behavior
than did parental monitoring.
However, with all of this empirical knowledge, researchers remain somewhat
frustrated at the lack of conclusive evidence regarding the relationship between parent-
child conflict and behavioral disorders. For instance, Rubenstein and Feldman (1993)
stated, ?it is not known to what extent adolescent behavioral and emotional disorders are
a function of the amount of conflict in the family? (p. 43). Some of this frustration stems
from the ?chicken-egg? dilemma. In other words, not enough work has been done to
identify the direction of the influence and/or its reciprocal nature. It may also be that the
51
importance of parent-adolescent conflict lies not with its existence within the family, but
in other, more subtle, moderating factors. For instance, whether such conflict is a
developmentally-enhancing or psychologically-damaging experience has been found to
rely on the manner in which the conflict is resolved (Hill, 1988; Laursen & Collins, 1994;
Montemayor & Hanson, 1985). Thus, in some instances, such conflict may actually be
?good? or helpful in the facilitation of necessary social-cognitive abilities in the context
of relationships (Cooper, 1988). Some researchers have found that the effect of parent-
child conflict on problem behavior is at least partially mediated through such factors as
family involvement, parental monitoring, or peer delinquency (Ary, Duncan, Duncan, &
Hops, 1999; Kim et al., 1999). 
Perhaps the primary source of parent-child conflict during adolescence is the
negotiation of autonomy (Steinberg & Silk, 2002). As young people grow older, new
roles need to be negotiated in the family. Both parents? and teens? views of family
relationships and expectations of each other change and violations of these expectations
can cause conflict (Collins, 1988, 1990; Dekovic, Noom, & Meeus, 1997). Such
developmental expectation mismatches have been found to be most prevalent during early
adolescence (Collins, Laursen, Mortensen, Luebker, & Ferreira, 1997). In fact, the most
common view of parent-child conflict is that it follows an inverted U-shape which peaks
during early adolescence and subsides thereafter (Montemayor, 1983; Steinberg & Silk,
2002). However, some recent work is beginning to show that, though a slight increase in
parent-adolescent conflict can be documented during early adolescence, subsequent
52
decreases are not necessarily empirically verifiable (Collins & Laursen, 1992; Laursen &
Collins, 1994; Laursen, Coy, & Collins, 1998).
Summary of Family and Deviance Literature
First, early empirical research in the autonomy dimension of family processes
established that parental supervision was lacking in the homes of delinquent adolescents
(Glueck & Glueck, 1950). More recent research concurs (Vazsonyi & Flannery, 1997). In
fact, research emanating from more than one ongoing research project (Oregon Social
Learning Center, Forgatch, 1991; Patterson et al., 1998; Vuchinich et al., 1992; Child
Development Project, Deater-Deckard et al., 1998; Pettit et al., 1997) confirms this
simple truth ? less parental behavioral control equals more adolescent problem behavior.
In addition, research shows that psychological control, as originally explicated by
Schaefer (1965) and more recently expanded by Barber (2002) is important as well. More
specifically, having an emotionally supportive family results in lower levels of risk for
deviance among adolescents (Johnson & Pandina, 1991; Windle & Miller-Tutzauer,
1992). Also, intrusive psychological control by parents can drive adolescents to employ
deviant behavior for the accomplishment of otherwise frustrated goals (Costanzo &
Woody, 1985).
Second, research concerning the harmony dimension of family processes has
shown that less parental affection leads to more deviance (Henggeler, 1989; Hurrelman,
1990) while large amounts of affective support protect against deviant activity (Johnson
& Pandina, 1991; Sokol-Katz et al., 1997). A weak emotional attachment very early in
life (first two years) has been shown to be important in the prediction of problem
53
behavior in adolescence (Sroufe et al., 1990) and this trajectory, if uninterrupted, seems to
be quite stable, even into adulthood (Greenberg et al., 1993). There is some evidence that
the earlier parents and children bond, the better (Corwyn & Benda, 2001), partly because
the effects of this warm relationship accumulate over time (Scaramella et al., 1999). In
addition, parental expressions of negative affect have been found to place children and
adolescents at a much greater risk for deviant behavior (Pittman & Chase-Lansdale,
2001). In general, then, positive parental caregiving is negatively associated with
adolescent deviance (Rothbaum & Weisz, 1994). Communication is an aspect of family
harmony that is difficult to study because it is often subsumed under a larger dimensional
conceptualization (e.g., Slicker, 1998) or broader construct (e.g., Lynch, 2001). We do
know that it may be important in the understanding of adolescent deviance (Ellickson &
Morton, 1999), but perhaps not as important as other family processes (Forehand et al.,
1997), though there is disagreement in this area (Leas & Mellor, 2000). 
Third, research in the conflict dimension shows that parental abuse leads to
problem behavior (Eckenrode et al., 1993; Sternberg et al., 1993) and that, even though
abuse as late as adolescence can significantly impact behavior (Benda & Corwyn, 2002),
the earlier abuse is inflicted, the more severe the behavior problems are later (Keiley et
al., 2001). Interparental conflict has also been implicated in the formation of deviant
behavior in adolescents (Porter & O?Leary, 1980; Vandewater & Lansford, 1998). In
addition, parent-child conflict, whether with mothers (Kim et al., 1999) or fathers (Shek
& Ma, 2001), constitutes a serious risk factor for problem behavior, regardless of sex or
family structure (Kim et al., 1999). Additional factors such as absence of a parent from
54
the home (Loeber & Stouthamer-Loeber, 1986; Steinberg, 1987) or parental deviance
(Nagin & Farrington, 1992; Sampson & Laub, 1993) have also been found to contribute
to norm-violating behavior.
Family and Religiosity
A large majority of high school students and young adults attend religious services
at least sometimes and say that religion has at least some importance in their lives
(Bachman et al., 2002; Smith, Denton, Faris, & Regnerus, 2002; Smith, Faris, Denton, &
Regnerus, 2003). Furthermore, Bachman and colleagues suggest that patterns of
religiosity and church attendance are normally established by the end of high school and
remain stable into adulthood. Of course, these patterns are largely transmitted from
parents to teens (Erickson, 1992; Flor & Knapp, 2001). In fact, parental religious beliefs
have been found to be the best predictors of adolescent religious beliefs (Acock &
Bengston, 1978; Dudley & Dudley, 1986; Hayes & Pittlekow, 1993; Hoge, Petrillo, &
Smith, 1982). Thus, it was understandable when Snarey and Dollahite (2001) recently
noted that ?religious faith and family relations are interrelated in positive, statistically
significant, and psychologically interesting ways? (p. 646). In other words, family and
religiosity are powerfully intertwined. Researchers attempting to explore the nature of this
interrelationship have done so in different areas of family life. Some have examined the
direct role religiosity may play in shaping parental behaviors such as parental warmth and
control or parental coping with a child?s developmental or physical difficulties. Others
have been interested in the indirect effect religiosity may have on marital relationships
55
which, in turn, have an influence on children. Most of the literature dealing with family
and religiosity focuses on these areas.
Parenting
Religion has also been found to have an impact on families through the
mechanism of parenting. For example, Brody et al. (1994) reported that both maternal and
paternal religiosity positively influenced marital interaction quality as well as co-
caregiver conflict (i.e., conflict arising between spouses concerning child-rearing issues).
In turn, these aspects of the marital relationship (interaction quality, caregiving conflict)
had a direct impact on parent-child relationship quality. More specifically, high levels of
personal religiosity by parents led them to report better quality interactions in their
marriage and less caregiving conflict between them which then led to better quality
relationships between the parents and their child. 
Studies often highlight the mother-child connection while neglecting the input of
fathers. However, recent findings from the National Study of Youth and Religion confirm
the importance of the role of fathers in the religion-family equation. Using data from the
National Longitudinal Survey of Youth 1997 (NLSY97), Smith and Kim (2002) have
found evidence that youth from religiously active homes are more likely than
nonreligious youth to have a positive relationship with their father. Similarly, Gunnoe et
al. (1999) found that greater personal religiousness led to better authoritative parenting
skills (for both mothers and fathers) and that this, in turn, led to more positive behavioral
outcomes for their adolescent children.
56
Research has found that Christian conservatism is modestly correlated with the
general belief of placing a high priority on child conformity and obedience (Danso,
Hunsberger, & Pratt, 1997; Ellison & Sherkat, 1993b), more favorable attitudes toward
corporal punishment (Danso et al., 1997; Ellison & Sherkat, 1993a; Grasmick, Bursik, &
Kimpel, 1991; Grasmick, Morgan, & Kennedy, 1992; Wiehe, 1990), and more frequent
use of corporal punishment (Ellison, Bartkowski, & Segal, 1996a, 1996b; Gershoff,
Miller, & Holden, 1999). However, this emphasis on parental control by religious parents
does not seem to require the sacrifice of child autonomy (Alwin, 1986).
Though the literature tying religiosity to warmth in family relationships is small,
results have consistently suggested that greater religiousness is related to greater
positivity. For example, in a very important longitudinal study on religion and family,
Pearce and Axinn (1998) showed that higher religious importance scores at time 1 were
significantly related to positive mother-child relationships at time 2 (five years later).
Parent-child congruence on religious matters also positively predicted later relationship
warmth. Wilcox (1998) even found that parental endorsement of theologically
conservative religious views about the Bible was related to greater parental warmth
(i.e., hugging and praising) as was church attendance.
There is a line of research dealing with the use of religion by parents to cope with
children with developmental or physical disabilities. As many as three-quarters of parents
with disabled or severely sick children use religion to cope (Coulthard & Fitzgerald,
1999). In general, it has been found that religious coping strategies are associated with
better parental health (Coulthard & Fitzgerald, 1999), less parental stress and depression
57
(Friedrich, Cohen, & Wilturner, 1988; Rogers-Dulan, 1998), more family cohesion
(Weisner, Beizer, & Stolze, 1991), greater social support (Barbarin & Chesler, 1984), and
less parental depression (Rogers-Dulan, 1998). One specific religious coping strategy is
using benevolent cognitive reappraisals (e.g., ?My child is a gift from God?) to reframe
the child?s problems and the parents? role in caregiving (Haworth, Hill, & Glidden, 1996;
Skinner, Bailey, Correa, & Rodriguez, 1999; Weisner et al., 1991). Parents of disabled or
sick children have also reported receiving psychological benefits, such as hope, strength,
and peace, from religious practices such as prayer, church attendance, or holy pilgrimages
(Bailey, Skinner, Rodriguez, Gut, & Correa, 1999).
Greater maternal, paternal, or family religiosity has also been linked to fewer
internalizing problems for youth (Brody et al., 1996). In a very important ten-year
longitudinal study, Miller, Warner, Wickramaratne, and Weissman (1997) presented
evidence for the fact that maternal religiosity protected children against depression. Some
have speculated that religion will lead to excessively harsh disciplinary practices which
will, in turn, lead to child psychopathology. It is notable that such a path has not yet been
identified. Rather, in this study, stronger religious beliefs and practices were found to
result in significantly less psychopathology.
Marriage
Several studies have linked religion and family by investigating the impact of
religious beliefs and behaviors on various aspects of marriage.
58
Marital Dissolution
Having a religious affiliation (as opposed to having none; Bock & Radelet, 1988;
Breault & Kposowa, 1987; Chan & Heaton, 1989), attending church services (Bahr &
Chadwick, 1985; Call & Heaton, 1997; Clydesdale, 1997; Heaton & Goodman, 1985),
and denominational homogamy between spouses (Bahr, 1981; Lehrer & Chadwick, 1993)
have all been documented as protective factors against marital dissolution. Similarly,
marital satisfaction has been positively linked to religious affiliation (Bahr & Chadwick,
1985; Bock & Radelet, 1988; Larson & Goltz, 1989) and religious participation
(i.e., church attendance; Bahr & Chadwick, 1985; Booth et al., 1995; Hatch, James, &
Schumm, 1986) as well as religious importance (Dudley & Kosinski, 1990; Mahoney et
al., 1999; Young, Denny, Luquis, & Young, 1998). Denominational homogamy has also
been found to predict marital satisfaction (Heaton & Pratt, 1990; Shehan, Bock, & Lee,
1990), presumably as a result of conflict resulting from different religious beliefs between
spouses. In addition, church attenders and couples who share common religious beliefs
(Larson & Goltz, 1989; Mahoney et al., 1999; Wilson & Musick, 1996) have been found
to be more strongly committed to their marital relationships.
Marital Conflict
The benefit of religiosity for marriage does not seem to be in preventing conflict,
since research in this area has not been able to document this kind of effect (Bahr, 1982;
Booth et al., 1995; Mahoney et al., 1999; Scanzoni & Arnett, 1987). However, there is
some evidence that greater religiousness is associated with a higher quality of
communication (Brody et al., 1994) and greater skill in handling disagreements (Mahoney
59
et al., 1999). Some have hypothesized that greater religious involvement would be related
to higher levels of domestic violence, but efforts to substantiate this claim have been
unsuccessful (Brinkerhoff, Grandin, & Lupri, 1992). In fact, there is evidence for the
reverse (Ellison, Bartkowski, & Anderson, 1999; Fergusson, Horwood, Kershaw, &
Shannon, 1986).
Marriage and Parenting
Brody et al. (1994) added the dimension of marital relationships to their
investigation of religiosity and family processes in order to explore potential operative
mechanisms. Their research involved African-American families and found that both
maternal and paternal religiousness predicted less co-parenting conflict and better marital
quality. In fact, marital quality and parenting skills mediated the effect of parental
religiousness on the quality of the parent-child relationship. Furthermore, in a second
follow-up study, Brody et al. (1996) found that paternal religiousness was related to
greater observed family cohesion and fewer child internalizing problems while both
maternal and paternal religiousness predicted less marital conflict and fewer child
externalizing problems. In this study, results from a structural equation model indicated
that parental religiousness produced positive child behavioral outcomes by promoting
family cohesiveness and lowering marital conflict.
Summary of Family and Religiosity Literature
First, religiosity seems to influence family processes through the channel of
parenting. Parents who show higher levels of personal religiosity exhibit stronger
authoritative parenting skills (Gunnoe et al., 1999; Smith & Kim, 2002) and report less
60
parent-child conflict (Brody et al., 1994). Conservative religious beliefs lead parents to
emphasize child obedience and corporal punishment (Danso et al., 1997; Gershoff et al.,
1999), though not at the loss of child autonomy (Alwin, 1986) or at the cost of abuse
(Mahoney et al., 2001). At the same time, it has been found that religion leads to more
positive parent-child relationships (Pearce & Axinn, 1998) characterized by more
physical affection and greater warmth (Wilcox, 1998). Parents with very sick or disabled
children report using religion as an effective coping mechanism (Bailey et al., 1999;
Coulthard & Fitzgerald, 1999; Rogers-Dulan, 1998) which seems to lead, in turn, to less
internalizing problems for their children (Brody et al., 1996; Miller et al., 1997).
Second, there is evidence that religiosity impacts family processes through
spousal interaction. Religious affiliation, religious participation, and denominational
homogamy have all been found to predict both marital satisfaction and marital dissolution
(Bahr & Chadwick, 1985; Heaton & Pratt, 1990; Lehrer & Chadwick, 1993) as well as
commitment to marriage (Mahoney et al., 1999). Greater religiousness has also been
found to be related to better marital communication (Brody et al., 1994) and conflict
management (Mahoney et al., 1999).
Third, there has been some evidence that religiosity interacts with both marriage
and parenting simultaneously. Both marital quality and parenting skills have been found
to mediate the relationship between parental religiousness and quality of parent-child
relationship (Brody et al., 1994; Gunnoe et al., 1999). In fact, paternal religiosity has been
found to be especially important in predicting less marital conflict and greater family
cohesion leading to more positive child outcomes (Brody et al., 1996).
61
Theoretical Considerations
The search to understand the relationship between religiosity and deviance should
not be void of theory, since theory should provide structure and meaning for future
exploration in this area. Despite the fact that Hirschi and Stark?s (1969) study did not
provide support for the relationship between religiosity and deviance, Hirschi?s (1969)
social control theory provides an ideal theoretical framework for such study.
Though Hirschi is not the first (see e.g., Paternoster & Bachman, 2001, who state
that ?there was a rich history of control theories of crime by the mid-1960's,? or, in other
words, before Hirschi published his; p. 73) nor the only (e.g., there have been Matza,
Nye, Reckless, Toby, Briar, Piliavin, and Reiss, to name only a few control theorists,
whose ideas may be attributed to the likes of Hobbes and Durkheim; Hirschi, 1969, p. xv)
control theorist, he has been labeled ?the spokesperson of the microsociological
perspective [of social control theory]? (Adler, Mueller, & Laufer, 1995, p. 61). Thus, it is
his explication of social control that has guided and justified research on a broad range of
topics, not the least of which is the relationship between religiosity and deviance.
The basic premise of Hirschi?s social control theory is that ?delinquent acts result
when an individual?s bond to society is weak or broken? (1969, p. 16). So the explanation
of deviance from a control theory perspective requires an accounting of the reason for
societal conformity. The important thing to explain, then, according to this perspective, is
not why people deviate from societal norms, but rather, why they conform to them. In
other words, while many are looking for an explanation of why people deviate, control
theory assumes that deviation will happen and looks instead for a reason for conformity.
62
Therefore, the important question to answer, from a social control perspective, is not
?Why do people disobey the rules of society?? but rather, ?Why do people obey the rules
of society??
Hirschi answers this question by hypothesizing that it is a ?bond to society? that
causes people to conform and, consequently, when this bond is weak or broken,
delinquent acts result. Subsequently, Hirschi delineates the nature of this ?bond,? stating
that it has four specific elements: 1) attachment to conventional people, 2) commitment to
conventional action, 3) involvement in conventional activities, and 4) belief in
conventional values. Concerning the attachment to conventional people which is a
necessary part of this bond, Hirschi specifies three contexts ? parents, school, and peers ?
where bonds with people are important and presents data (from the Richmond Youth
Study) to support their role as controls against deviance. Here, he could have easily
discussed the church (i.e., ?religiosity?) as another parallel control mechanism, but
presumably he chose not to because of his own findings that religiosity was not important
in this context.
The important thing to understand about deviant behavior is what social structures
or institutions may successfully persuade and enable an individual to conform to societal
norms, it becomes obvious how religiosity may fit into social control theory.
An important social institution not included by Hirschi is religion. Bonds to
religious organizations may inhibit drug use in several ways. First, individuals
may become attached to a church and the people they associate with in the
church. Because of that attachment, individuals who attend a church may be less
63
likely to use various drugs than might individuals who do not attend a church.
Second, involvement in religious activities may leave less time available for drug
experimentation. Involvement may also provide a network of support which may
insulate people from opportunities to use drugs. Third, commitment to a religious
organization and its goals may provide meaning to life that will make drug use
appear less attractive. Fourth, the belief system of most religious groups is against
drug use, and those teachings may reinforce personal beliefs against drug use.
(Bahr, Hawks, & Wang, 1993, p. 448)
Thus, religiosity can conceivably exhibit all four elements of a bonding mechanism,
making it a potentially important control factor in any control-based empirical
investigation of adolescent deviant behavior. Not surprisingly, religiosity has been
frequently conceptualized in this manner (e.g., Cochran et al., 1994; Evans et al., 1996;
Ross, 1994, as only a small sampling).
In addition to its utility in conceptualizing religiosity as a control mechanism for
the deterrence of deviance, social control theory offers the ability to identify some
etiological factors as more distal and others as more proximal. Such flexibility lends itself
nicely to a search for the more complex relationships between potential causal factors.
For instance, it is well documented that family control factors, such as attachment to
parents, are important in the explanation of adolescent deviance. In addition, we know
that religiosity has, at the very least, a moderate effect in controlling deviance. A social
control perspective allows us to go beyond these simplistic bivariate associations in the
identification of a more complex model of the mechanisms by which these processes may
64
operate. It may be that religiosity mediates the relationship between family processes and
deviance. Just as easily, however, the inverse could be true, that is, family processes may
mediate between religiosity and deviance. Another intriguing possibility is that one or the
other moderates the relationship so that high levels of one produce a more favorable
relationship, while low levels result in a relationship which is less favorable.
All of these possibilities exist within a social control framework and can be
empirically tested. Therefore, this theoretical perspective is employed in this paper in
order to facilitate an exploration of the interplay between two specific control
mechanisms, family processes and religiosity, in the explanation of adolescent deviant
behavior.
Family, Religiosity, and Deviance
It is evident, based on the preceding material, that significant bivariate
relationships between religiosity and deviance, family and deviance, and family and
religiosity have been empirically documented. However, the interplay between all three of
these has not been widely studied. Discovering the specific mechanisms by which these
three important factors operate is an important next step in this area of research. Research
concerning other intervening factors in the relationship between religion and deviance has
already been conducted. Most frequently, this research has used a social control
perspective and has examined the mediating role of several important control factors.
The purpose for such an examination lies in the widely held idea among control
theorists that the frequently observed and well-documented relationship between
religiosity and deviance is spurious in the sense that sufficient controls have not been
65
utilized. In this view, there are two likely scenarios under which the spuriousness of this
relationship may be exposed. First, more proximal control factors may mediate the effect
of religiosity, which, being more distal or antecedent, has no direct linkage to delinquency
(e.g., Elifson, Peterson, & Hadaway, 1983). Second, religiosity may be one of multiple
sources of social control over deviance which, being comparatively weaker than the
others, loses its impact when the others are simultaneously considered. In either case,
according to this school of thought, the inclusion of additional, ?more important? sources
of social control would reduce the inverse bivariate relationship between religion and
deviance to nonsignificance.
As an example of an empirical test of this social control perspective, Evans,
Cullen, Dunaway, and Burton (1995) investigated the possibility that secular social
controls may attenuate the effect of religion on crime. Three separate measures of
religiosity (religious activity, religious salience, and religious beliefs) were included as
well as two measures of social control (informal social constraints and formal legal
deterrents). Their findings suggested that religious activity had a direct effect on
criminality even in the presence of the secular controls. It is important to note, concerning
this study, that the sample was composed of adults, not adolescents. Furthermore, the
measure of deviance in this study was limited to self-report items concerning criminal
behavior. In other words, a wider variety of non-criminal deviance was not examined.
Since Hirschi?s (1969) social control theory advocates the central role of family
processes as a bonding mechanism, most studies conducted from a social control
perspective have included some measure of parenting. For instance, in a very recent
66
study, Benda and Toombs (2000) tested several mediation and moderation models in a
test of the robustness of the religiosity-deviance relationship. Both attachment to mother
and attachment to father were included as potentially intervening social control elements.
Their sample consisted of 600 men in a boot camp program ranging in age from 15-47
years. Two separate measures of religiosity were included in the study. The first was a
multi-item scale of religiosity, including such activities as prayer, Bible study, and
attempts to convert others. The second was a single-item measure of church attendance.
Using hierarchical logistic regression to predict a dichotomous indicator of violent
behavior, results showed that church attendance was not related to violence while the
religiosity scale was significantly related ($ = -.67). As additional variables were added to
the equation (sociodemographic, social control, strain, social learning; a total of 22
variables/scales), the effect of the religiosity scale was not diminished. More specifically,
neither attachment to mother or father nor any of 22 other theoretically-oriented variables
explained away the effect of religiosity on deviance. These results present strong evidence
for the fact that the effect of religiosity is not mediated, but rather is quite robust. Similar
findings were made by Chadwick and Top (1993) where, in a multiple regression
including several measures of peer and family influence on various types of deviant
behavior, the individual religious behavior measure retained its significant impact on all
four measures of deviance (victim offenses, victimless offenses, property offenses, and a
total deviance measure).
Benda (1995) made some more complex findings. His study sample consisted of
1,093 public high school students (grades 9-12) attending six different schools in both
67
rural and urban areas of Oklahoma, Maryland, and Arkansas. Results from this study
indicated that the effect of an 8-item religiosity scale on three measures of delinquency
(property crime, personal crime, status offenses) and four measures of substance use
(marijuana, amphetamine, alcohol, heavy alcohol) was differentially affected by the entry
of intervening measures of parenting (attachment to mother, attachment to father, parental
monitoring). On the one hand, parenting mediated the relationship of religiosity with
marijuana use, while, at the same time, a non-significant relationship with person crimes
became significant upon entry of the parenting variables, indicating a possible suppressor
effect. Also, though religiosity was never significantly related to amphetamine use, it
maintained a significant relationship with property crimes, status offenses, and both types
of alcohol use, even after controlling for parenting variables.
In a similar follow-up study, Benda and Corwyn (1997) conducted an
investigation using three separate multi-item personal religiousness scales (church
attendance, religiosity, and evangelism) and two multi-item social control scales
(attachment to mother and father). The findings of this study indicated that religiosity and
church attendance (but not evangelism) were significantly related to the commission of
status offenses (e.g., drinking alcohol), but that this relationship was reduced to non-
significance when the social control variables (maternal and paternal attachment) were
added to the model. However, the reverse was true when criminal offenses were
examined, namely, neither religiosity nor church attendance were related at all to crime,
but evangelism maintained a strong significant relationship, even after the addition of
68
parental attachment. Together, these results are somewhat confusing and difficult to
interpret.
Brody et al. (1996) tested a model that hypothesized a mediating role for family
processes between religiosity and adolescent behavioral outcomes. One of the unique
aspects of this study was its sample. Only African American families (90) with a 9- to
12-year-old firstborn child and two parents living in a rural setting were recruited for the
study. A measure of formal religiosity was included for both mothers and fathers based
on two items (church attendance, importance of church attendance). There were also two
measures of family processes (conflict-harmony scale, engagement scale) based on coded
video observations. Interparental conflict was measured with a 10-item self-report
instrument. Finally, externalizing behavior was measured in two ways (Revised Behavior
Problem Checklist, Self-Control Inventory) separately by mothers, fathers, and teachers.
Latent variable path analysis with partial least squares (LVPLS) estimation procedures
were used to test a model that hypothesized that the effects of parental religiosity would
be mediated through both interparental conflict and family cohesion, thus indirectly
affecting child externalizing problems. Indeed, results indicated a good fit for this model
(see also Brody & Flor, 1998, where similar findings were made). Unfortunately, this
study is only partially helpful since it focuses on parental religiosity rather than the
personal religious commitment of the target child.
The above studies found that (at least to a certain extent) the effects of religiosity
were not explained away by parental control elements. However, Cochran et al. (1994)
found something different. Data for this study were collected from 1,591 ninth through
69
twelfth grade students attending school in five purposively selected school districts in
Oklahoma. The measures included five delinquency scales (assault, theft, vandalism,
illicit drug use, licit drug use) and one additional delinquency item (truancy), three
religiosity items (religious participation, religious salience, and religious affiliation), and
two measures of parental control (parental supervision scale, broken home item). Results
indicated that, though both religious participation and salience were significantly related
to the delinquency variables, the addition of social control variables, such as parental
supervision and broken home status, reduced this relationship to non-significance for all
but licit drug use.
Similarly, Evans et al. (1996) used a multi-item, multi-dimensional measure of
personal religiosity to assess the efficacy of religion as a deterrent in the presence of other
important control factors such as attachment to parents. Results indicated that the
moderately significant impact of personal religiosity on general delinquency was reduced
to non-significance upon the addition of parental attachment and other control variables.
Ross (1994) found that a multi-dimensional measure of religiosity failed to increase the
amount of variance explained when added to a regression equation which already
included such control items as attachment to family. The problem with this study, though,
was that neither the unidimensional (single item measuring church attendance) or
multidimensional (five scales with a total of 12 items measuring religiosity) religiosity
measures had even an initial significant bivariate relationship with deviance.
In another study, Elifson et al. (1983) found very strong zero-order correlations
between a measure of religious salience and several measures of deviant behavior
70
(gammas ranging from -.15 to -.41), yet when other etiological factors, such as closeness
to mother, were added to the regression, its effect was reduced to nonsignificance. Using
structural equation modeling (with LISREL), Bahr et al. (1993) found that, when other
important factors (e.g., parental monitoring, family drug use) were in the model, the
contribution of both cohesion with parents and religious salience was insignificant, and,
therefore, both were removed completely from a final model attempting to predict
adolescent substance use.
These and similar studies emphasize the mediating role of family processes (and
other social control elements) in the relationship between religiosity and deviance. As can
be seen, the results are mixed and no definitive conclusion has yet been reached in the
literature. Another way of looking at the relationship between these three key constructs
(religiosity, family processes, and deviance) may be to hypothesize an intervening role for
religiosity. For example, Litchfield, Thomas, and Li (1997) examined the possibility that
religiosity mediated the relationship between parenting and adolescent deviant behavior.
In 1989, self-report data were collected from 1,715 Mormon adolescents (11-15 years
old). Data were subsequently collected from the same youth in 18-month increments at
three more points in time ? 1,492 in 1990, 1,369 in 1992, and 1,133 in 1993. Three multi-
item measures of family processes (behavioral regulation, maternal connection, family
religious observance), three multi-item measures of religiosity (public religiosity, private
religiosity, future religious plans), and three multi-item indicators of deviant behavior
(sex/substances, aggression, vulgarity) were all included in a structural model which
hypothesized that the effects of the more distal family variables on deviance would be
71
mediated through the religiosity variables. Results of LISREL analyses using early data to
predict later data (i.e., family processes in 1989 predicting religious behavior in
1990/1992 leading to deviance in 1993) indicated that this model fit the data well. It
should be noted that this sample may not be representative since it consisted of
?religiously active participants from highly educated, intact families? (Litchfield et al.,
1997, p. 215).
Another way in which religiosity may intervene between family processes and
deviance is in a moderating role. Previous research has shown that religiousness can
moderate the effects of stressful life events (Pargament, 1990), coping with adversity
(Balk, 1983; Palmer & Noble, 1986; Seligman, 1991), and adjustment to bereavement
(McIntosh & Spilka, 1990). Clearly, religious people can be distinguished from non-
religious people on a number of behavioral outcomes. Therefore, it would make sense
that religiosity could similarly condition the relationship between family processes and
adolescent deviant behavior. In fact, there is some evidence that religiosity may moderate
the relationship between family processes and various types of deviant behavior (Bahr et
al., 1998; Thomas, 1988). Finally, there is also the possibility that the interrelationship of
family processes, religiosity, and deviant behavior is best explained by a two-way
interaction between the predictor variables (i.e., family processes and religiosity).
Other Important Intervening Factors
The primary importance of the current investigation lies in its attempt to explain
the relationship between family processes and religiosity in their prediction of adolescent
deviant behavior. Thus, the preceding review of literature has focused on those constructs
72
that hold the position of central importance in this examination. However, it would be
shortsighted for the present study to completely ignore other important issues that have
emerged from decades of research in this area. Such issues, though not central in focus,
are still intriguing and, more importantly, may require further empirical clarification
regarding their role in understanding family, religiosity, and deviance.
As was stated earlier, Hirschi and Stark?s (1969) findings prompted many
researchers to examine the conditionality of the relationship between religiosity and
deviance. Consequently, several important intervening factors have been tested and
identified. For example, there has been some evidence that religiosity may deter only
certain types of deviant behavior. This evidence has been inconsistent and tenuous at best,
thus requiring a more solid conclusion on the matter. In addition, developmental issues,
such as the role of sex or age, have sometimes been examined, but their role still remains
quite unclear, particularly in relation to all of these variables of interest (i.e., family
processes, religiosity, deviance) simultaneously. Finally, measurement itself (the use of
single-item, variance-restricted measures versus multi-item, reliable scales) has come to
the forefront of research in this area. Therefore, advancing the field by providing
conclusive evidence that more reliable measures of religiosity produce more robust
results concerning the role of religion in the development of adolescents in general and,
more specifically, in their choice to participate or refrain from deviant behavior continues
to be a matter of priority for researchers.
These three issues (i.e., type of deviance, developmental factors, measurement)
have been discovered by previous researchers to be important in this area of study.
73
Though they are not central to the focus of the current study, they represent
conditionalities which, if left unclarified, may cast doubt upon the present findings.
Therefore, because the data for the current study allow for an examination of these issues,
they will be further investigated.
Type of Deviance
A rather large consensus has developed among researchers of religion and
deviance that this relationship is conditioned by the type of deviant behavior under
examination. More specifically, some feel the evidence suggests that religiosity is only an
effective deterrent to deviant behavior that is not condemned by society in general.
Middleton and Putney (1962) were perhaps the first to advocate this perspective when
they argued the following:
Much of the confusion surrounding the relation between religion and morality
derive [sic] from failure to distinguish two different types of ethical standards ?
the ascetic and the social. Social standards proscribe actions which in general are
harmful to the social group, and, we hypothesize, tend to be shared by the
religious and the nonreligious alike as a part of a general social ideology. The fact
that religious ideology may also proscribe these actions is incidental; . . . In
contrast, ascetic standards ? abstinence from sensual indulgences, gambling, and
the like ? derive from an ascetic religious tradition. Within the context of religious
ideology violations of ascetic standards may be held spiritually harmful to the
perpetrator. But since such violations are usually not directly or obviously harmful
to the social group ? at least in moderation ? ascetic standards have less
74
persuasiveness to the secularly oriented individual. He is therefore more likely to
violate them.? (pp. 142-143).
Because of Middleton and Putney?s use of the term ?ascetic standards,? this
perspective has been labeled ?The Antiascetic Hypothesis.? In their study using simple
frequency analyses, Middleton and Putney found evidence in support of this hypothesis,
namely that religious persons are more likely to believe antiascetic actions (gambling,
smoking, drinking, and sexual activity) are wrong and they are less likely to engage in
them while religious and non-religious persons do not differ in their beliefs about the
morality of antisocial actions or in their participation in such.
Since the proposal of this antiascetic hypothesis, several studies have attempted to
test it, with very mixed results. Part of the problem has been in the interpretation of what
it means to be ?antiascetic? and in the operationalization of this concept. One of the
earliest studies to test Middleton and Putney?s antiascetic hypothesis was Burkett and
White (1974) and their interpretation/operationalization shaped much of the subsequent
literature. The sought to identify a type of deviant behavior that would be viewed as more
acceptable or less serious by society than by religious people and concluded that
?victimless? crimes, especially marijuana and alcohol use, were the best example of such.
Consequently, when they tested this hypothesis using alcohol and marijuana use as
operational measures of ?victimless crime? and contrasted these behaviors with ?more
serious? crimes, such as larceny, vandalism, and assault, they found that religious
participation was not related to the more serious category of criminal activities but was,
indeed, related to alcohol and marijuana use. Similarly, Albrecht, Chadwick, and Alcorn
75
(1977) found that measures of religiosity were more strongly related to victimless than
victim deviance (see also Elifson et al., 1983; Fernquist, 1995).
After Burkett and White (1974) and Albrecht et al. (1977) laid the groundwork for
testing the antiascetic hypothesis, the majority of researchers adopted a similar position
and, therefore, the central focus of tests of the antiascetic hypothesis has been on
substance use as a type of deviant behavior that is condemned more strongly by churches
than other social institutions. However, this is a problematic approach since drug and
alcohol use is clearly not ?condoned? by society. For example, Corwyn and Benda (2000)
found that religiosity had a significant negative impact on ?hard drug? use (e.g., cocaine,
heroin, hallucinogens, solvents), even in the presence of several other control variables
such as attachment to mother, attachment to father, and parental supervision. Others who
have made similar findings have concluded that their results supported the antiascetic
hypothesis since avoidance of drugs has often been considered a particularly religious
teaching. However, based on these findings, Corwyn and Benda concluded that ?these
results contradict the antiascetic hypothesis? since ?there would seem to be little
ambiguity in societal values and norms regarding use of ?hard drugs? . . . among
adolescents? (p. 253). Thus, there is disagreement and confusion in the literature as to
what constitutes antiascetic behavior and what its relationship to religiosity really is.
In spite of the fact that substance use is not necessarily an appropriate measure of
antiascetic behavior, numerous studies have focused exclusively on substance use as the
behavioral outcome of interest in tests of the deterring influence of religiosity according
to the antiascetic hypothesis. For instance, in a test of the effect of religiosity on 18
76
different deviant behaviors, Jensen and Erickson (1979) found that the strongest
relationship was with alcohol and marijuana use. In addition, Jensen and Erickson also
tested their idea in Hirschi?s original Richmond Youth Study data and found that, in this
very data set, church attendance was significantly negatively related to drinking, smoking,
and truancy, deviant behaviors which the authors considered antiascetic. Tittle and Welch
(1983) also tested this same hypothesis in a somewhat broadened form by hypothesizing
that religious proscriptions are enhanced when the community at large is characterized by
moral ambiguity and moral guidelines are, therefore, unavailable. They found evidence
supporting this hypothesis.
However, empirical evidence regarding the antiascetic hypothesis has not at all
been consistent. For example, Chadwick and Top (1993) found religiosity to be
significantly negatively correlated with both person and property offenses as well as
victimless offenses. Also, in Jensen and Erickson?s (1979) study mentioned above, it was
found that religiosity had a significant negative relationship with 14 of the 18 deviant
behaviors measured, including several more serious offenses such as grand theft,
vandalism, and assault. In addition, Evans et al. (1996) found that, though their multi-
item measure of personal religiosity was significantly negatively related to a measure of
general delinquency, it had only a weak, non-significant, negative relationship to a more
specific measure including only antiascetic delinquency items.
The empirical evidence in general indicates that religiosity has a negative
relationship with both drug and alcohol use (Brownfield & Sorenson, 1991; Cochran,
1993; Cochran & Akers, 1989; Dudley, Mutch, & Cruise, 1987; Ellis & Thompson, 1989;
77
Free, 1994; Linden & Currie, 1977; McLuckie, Zahn, & Wilson, 1975; Pearlman et al.,
1972; Perkins, 1987; Turner & Willis, 1979; Vener, Zaenglein, & Stewart, 1977;
Whitehead, 1970). In the influential Monitoring the Future study, considerable
longitudinal evidence has been gathered from a nationally-representative sample in
support of the consistently deterrent effect of religiosity on substance use of all kinds. In a
summary of eight data points over an eight-year period of time, Bachman et al. (2002)
found that young people who regularly attended church (once or more a week) were
consistently less involved in cigarette smoking, alcohol drinking, use of marijuana, and
use of cocaine. This pattern held true in each of these categories, not only in baseline
data, but also in every year for each of the subsequent seven years of follow-up data
collection.
Developmental Issues
Much of the religion-deviance literature has completely ignored, or at least treated
in a peripheral manner, such developmental issues as age and gender. However, it seems
important to understand whether the relationship between religiosity, family processes,
and deviant behavior changes as an adolescent gets older or varies depending on whether
one is male or female.
Age
Age is an important developmental construct and, therefore, the interplay between
age and the variables of interest in the current study is an important developmental
question. Previous empirical research has established that participation in religious,
family, and deviant activities varies somewhat by age. It would follow that potential
78
changes in the interrelationship between these variables due to increases in age should
also be examined.
Age and Religiosity
Participation in and importance of religion has been found to be age dependent.
For instance, Bachman et al. (2002) found that the percentage of people reporting that
religion was ?very important? gradually increased between the ages of 18 and 32, from
24% to 29% for males and from 33% to 41% for females. However, during the same time
period, the percentage who reported ?rarely? or ?never? attending church increased (44%
to 60% for males; 39% to 49% for females) and the percentage of regular church
attenders (once a week or more) declined (40% to 25% for males; 46% to 33% for
females). Though these findings indicate definite developmental trends in religious
participation associated with age, they are also somewhat contradictory and require
further exploration.
For the most part, empirical research has shown a persistent overall decline in
religiousness during adolescence (Donahue & Benson, 1995). Potvin, Hoge, and Nelson
(1976) observed declines in prayer, feelings of closeness to God, and orthodoxy during
adolescence (see also Benson, Donahue, & Erickson, 1989; Roehlkepartain & Benson,
1993), while Benson, Yeager, Wood, Guerra, and Manno (1986) reported lowering scores
among adolescents on both intrinsic and extrinsic religiousness (see also Allport & Ross,
1967). Unfortunately, these empirical findings are somewhat inconsistent with
developmental theories which postulate that issues regarding the meaning and purpose of
79
life (such as religion) become more salient with age (e.g., Erickson 1968; Kohlberg,
1969). Clearly, greater elucidation is necessary in this area.
Age and Family
The classic view of changes in parent-adolescent relationships during adolescence
emphasizes the need for adolescents to emotionally disengage from the family and views
those adolescents who remain emotionally attached to parents as somewhat inferior to
those who are able to detach themselves (Erickson, 1959; Marcia, 1980).
Developmentalists have characterized adolescence as a time when a childlike unilateral
trust in parents is replaced by a more autonomous view of authority (Kohlberg, 1969;
Piaget, 1965). Some scholars, however, have shown concern that this view may promote
agency at the expense of connectedness (Baumrind, 1987, 1991; Gilligan, 1987;
Sampson, 1988).
In any case, empirical research has shown that family processes, including parent-
adolescent relationships, is fluid during adolescence and different in adolescence than
childhood. Many scholars have argued that parent-adolescent relationships follow a
curvilinear pattern where they deteriorate rapidly at early adolescence, remain poor during
middle adolescence, and finally begin improving during late adolescence
(e.g., Montemayor, 1983). However, in a recent meta-analysis of studies examining
changes in parent-adolescent conflict during adolescence, Laursen et al. (1998) found that
the aggregated evidence did not support this common assertion. Rather, evidence
accumulated from 37 empirical studies indicated an identifiable linear decline in parent-
adolescent conflict. Conflict was found to be greater in early adolescence than mid-
80
adolescence and greater in mid-adolescence than late adolescence. There was some
evidence, however, for the fact that conflict in mid-adolescence, though less in frequency,
may be more heated than in early adolescence.
Age and Deviance
Remarkable stability has been observed in antisocial behavior (Elliot, Huizinga, &
Ageton, 1985; Farrington, 1995; Huesmann, Eron, Lefkowitz, & Walder, 1984; Loeber &
Dishion, 1983; Olweus, 1979) beginning in early childhood (Lerner, Hertzog, Hooker,
Hassibi, & Thomas, 1988) and moving into adolescence and adulthood (Kazdin, 1987;
Robins, 1978). In fact, compelling evidence is emerging for a time-ordered sequence of
events which may almost universally predict an early-onset trajectory for juvenile
offending (Patterson et al., 1998). Progress down this pathway may begin as early as
fourth grade (Farrington & Hawkins, 1991) or even first grade (Kellam, Brown, Rubin, &
Ensminger, 1983) and leads to later juvenile and adult deviance (Loeber, 1982; Moffitt,
1983; Patterson, 1995; Zhang, Wieczorek, & Welte, 1997).
However, research has also documented that deviance varies considerably with
age. More specifically, a population-based pattern has been well-established showing a
surge in deviant behavior during adolescence followed by a steady decline beginning in
early adulthood and continuing throughout life (Blumstein & Cohen, 1987; Farrington,
1986; Hirschi & Gottfredson, 1983; for an example of this pattern, compare Evans et al.,
1996, who found that delinquency increased with age in a high school sample
[mean age = 16.4], with Evans et al., 1995, who found that crime decreased with age in
an adult sample [median age = 41]). This pattern of gradual change does not seem
81
consistent with the previous thought of stability. Moffitt (1983) has attempted to
reconcile this apparent conflict between stability and change by proposing two separate
types of deviance trajectories ? adolescence-limited and life-course-persistent. More
simply, this trajectory which reflects change should be viewed as stable in itself.
Age, Religiosity, Family, and Deviance
As evidenced in the brief discussions above, age has a documented effect on all
three of the variables of interest in the current study. However, research dealing with
more than one of these areas simultaneously while also including a measure of age is
more rare. Age has sometimes been included in studies involving the link between
religion and delinquency. Most of the time, age has simply been controlled in such
analyses. For instance, Evans et al. (1996), when controlling for age, found that it was the
strongest predictor of delinquent behavior (being older meant being more delinquent)
among all demographic variables, was a stronger predictor than even religiosity, and was
second only to conventional beliefs in a model including demographic and social control
elements. In a similar earlier study, Evans et al. (1995) controlled for age and results
indicated that it was a stronger predictor of crime than any theoretical or religiosity
measures.
Benda and Toombs (2000) provided a more interesting opportunity to examine the
effects of age on the religion-deviance relationship when they tested whether age
moderated the effect of religiosity on violent behavior. As expected, their results
indicated that age did indeed moderate this relationship. More specifically, they found
that the inhibitory effects of religiosity on violence were greater for older subjects (over
82
24 years old) than for younger. However, an opposite finding was made by Rohrbaugh
and Jessor (1975) whose study suggested that the effect of religion as a personal control
against deviant behavior declined with age.
Age has also been examined in its relationship to family processes and deviance,
resulting in mixed findings. On the one hand, Miller, Cowan, Cowan, Hetherington, and
Clingempeel (1993) reported that the younger children in their study exhibited stronger
associations between familial variables and various measures of externalizing behavior
than the older children in the study. They attributed this finding to the fact that older
children were achieving greater levels of independence and were interacting with people
other than just parents (e.g., peers, teachers, others) and, consequently, greater autonomy
coupled with multiple types of outside influence resulted in a reduction of the importance
of family processes in the determination of behavior.
On the other hand, Rothbaum and Weisz (1994) conducted an extensive meta-
analysis where the results from 47 empirical studies were quantitatively aggregated and
subsequently statistically analyzed. They found that older children/adolescents (6 years
and older) exhibited a stronger connection between parental caregiving and externalizing
behaviors than younger children (5 years and younger). This was exhibited by the fact that
their mean effect size was larger (r = .31 vs. r = .21) and the percentage of the effects that
were significant in these studies was significantly greater (61% vs. 33%). After
controlling for type of measure (self-report vs. observation), these differences became
slightly more pronounced (mean effect size: older, r = .37; younger, r = .21; percentage of
significant effects: older, 64%; younger, 33%). These results seemed to indicate that
83
family processes (i.e., parental caregiving) was a more important deterrent to deviance
(i.e., externalizing behavior) among older children/young people. This may be viewed as
support for the assertion of reciprocity theory that parent-child mutuality may take years
to develop and that, consequently, the parent-child bond is stronger at older compared to
younger ages (e.g., Patterson, 1982). Though the results of the meta-analysis may seem to
outweigh those from a single study (Miller et al., 1993), it still remains that such
conflicting evidence may require further elaboration.
In a more recent study, Amato and Fowler (2002) presented evidence that harsh
parental punishment was consistently positively related to problem behavior for young
people regardless of their age. In addition, they found that parental support was important
for preventing behavior problems among adolescent youth, but not among preadolescent
children. In this study, parental monitoring seemed to have no significant impact on
problem behavior for youth of any age.
Summary of Age, Religiosity, Family, and Deviance
Church attendance seems to decline with age while at the same time the salience
of personal religiosity increases (Bachman et al., 2002), though there are conflicting
findings in this area (Benson et al., 1986). There is also evidence that adolescents?
participation in family processes diminishes as they grow older (Laursen et al., 1998).
Deviant behavior shows a corresponding increase during the adolescent years (Farrington,
1986; Hirschi & Gottfredson, 1983). In fact, this pattern of increased deviant activity
during adolescence has been found to both mediate (Evans et al., 1995, 1996) and
moderate (Benda & Toombs, 2000; Rohrbaugh & Jessor, 1975) the effect of religiosity on
84
deviance. In addition, some have found that the association between family processes and
deviance was stronger for younger than older adolescents (Miller et al., 1993) while
others have made the opposite finding (Rothbaum & Weisz, 1994). Recently, a case has
been made for the possibility that such relationships do not change over time, but are, in
fact, the same regardless of age (Amato & Fowler, 2002).
Sex
Sex is another important developmental construct and it is important
developmentally speaking to determine whether the interrelationship between the
variables of interest in the current study is conditioned by sex. Previous research has
found that religious participation, family relationships, and deviant behavior all vary
somewhat by sex, but it is not yet accurately known whether we should expect their
simultaneous relationship to be different for males and females.
Sex and Religiosity
Empirical research has provided consistent evidence that females exhibit more
religious behavior than males (Donahue & Benson, 1995; Fiese & Tomcho, 2001;
Sullivan, 2001). This pattern seems to hold throughout the lifespan (Argyle & Beit-
Hallahmi, 1975; Batson, Schoenrade, & Ventis, 1993; Francis, 1985; Roehlkepartain &
Benson, 1993; Snarey & Dollahite, 2001) though the effect size is normally weak
(i.e., < .10). There is also evidence that parental religiosity, which is one of the most
potent predictors of adolescent religiosity, has a stronger impact on girls than boys
(Erickson, 1992).
85
Sex and Deviance
The fact that males are much more deviant in their behavior has been well-
established empirically (Blumstein, Cohen, Roth, & Visher, 1986; Elliott, Huizinga, &
Menard, 1989; Hickman & Piquero, 2001; Maccoby, 1986). In fact, ?maleness? is
normally considered a risk factor in the etiology of deviant behavior (Huselid & Cooper,
1994; Zahn-Waxler, 1993). Donahue and Benson (1995) found that ?femaleness? was
negatively associated with every form of deviance they tested, including alcohol use,
binge drinking, marijuana use, sexual intercourse, and violence, while Deater-Deckard et
al. (1998) found ?maleness? to be positively associated with three separate measures of
externalizing behavior (see also Carlucci, Genova, Rubackin, Rubackin, & Kayson, 1993;
Engs & Hanson, 1990; Evans et al., 1995; Evans et al., 1996).
Sex, Religiosity, Family, and Deviance
Empirical research has identified a role for sex in the understanding of adolescent
religious and deviant behavior as well as processes within adolescents? families.
However, the function of sex as a moderator between these variables in their
simultaneous relationships is still largely unknown. For example, in a study involving the
intersection of religiosity and family, Pearce and Axinn (1998) found evidence that the
more important religion was to a mother, the more likely her child was to report a higher
quality relationship with her. This touches on sex since mothers are specified rather than
fathers, but not in the way we would hope. First, fathers were not included in the study at
all, therefore, no contrasts could be made between the effects of maternal and paternal
86
religiosity. Second, the sex of the child was not distinguished in this relationship,
therefore, it is left undetermined whether this effect is different for males than females.
In a meta-analytic investigation of the role of gender in the relationship between
family and deviance, Rothbaum and Weisz (1994) investigated whether parental
caregiving was more salient in the deterrence of externalizing behavior for male or female
children. Initial analyses involving all aggregated studies found statistical trends (p < .10
for both mean effect size and percentage of significant effects) but no statistical
differences between boys and girls. In an additional set of analyses, significant sex
differences were found among preadolescent children living with their single mothers.
These statistically significant findings indicated that maternal caregiving was more
important for the deterrence of deviance in preadolescent boys than girls. This was
consistent with Zaslow (1989) who had previously made an even more specialized
finding, namely that these associations were stronger for boys than girls only among
preadolescents being raised by divorced, unremarried mothers. This may or may not be
consistent with Baumrind?s (1991) finding that an intact family structure is more
important for prohibiting problem behavior for girls than boys.
Only a few studies have examined the linkage between religiosity and deviance
through a gender lens. When Rohrbaugh and Jessor (1975) did so, they found a very
similar relationship by gender. Religiosity correlated similarly by sex with marijuana use
(r = -.29 for males, r = -.31 for females, both statistically significant) and general deviant
behavior (r = -.16 for males, r = -.22 for females, both statistically significant) among
high school students and the same general pattern held up for college students as well.
87
On the other hand, Heath et al. (1999) found the deterrent effect of religiosity on
drug, cigarette, marijuana, and alcohol use to be stronger among adolescent girls than
boys. Along the same line, Albrecht et al. (1977) found some interesting gender
differences when testing the effect of religiosity on deviance. Though a measure of
religious behavior was significantly negatively related to measures of total and victimless
deviance for both males and females, a measure of religious attitudes was only related to
these deviance constructs for males. In contrast, neither religious attitudes nor religious
behavior were related to victim deviance for males though they were both related for
females. Similarly, Chadwick and Top (1993) also identified some interesting gender
patterns in their examination of this relationship. Their measure of private religious
behaviors was significantly negatively correlated with three different types of deviant
behavior (person offenses, property offenses, and victimless offenses) for both males and
females and a measure of church attendance was found to provide constraint against
victimless offenses for both as well. However, religious beliefs predicted property
offenses and religious integration predicted both person and property offenses, but both
only for males. Spiritual experiences, on the other hand, predicted person offenses only
for females.
In their meta-analytic study of the religion-deviance literature, Baier and Wright
(2001) tested for gender differences as well. They found that a quantitative variable
representing the male proportion of the samples included in their review positively
predicted an effect by religion on deviance. Though this finding was non-significant, it
implied that, as a whole, the impact of religion on deviance may be slightly stronger for
88
males than for females. Consequently, the only thing that is clear from a review of the
literature detailing the role of sex in the relationship between religiosity and deviance is
that this role is still unclear. Some say the relationship does not differ by sex, others say it
is stronger for girls, and still others say it is stronger for boys. Obviously, more
clarification is required.
Summary of Sex, Religiosity, Family, and Deviance
Females tend to be more religious than males (Sullivan, 2001) throughout the
lifespan (Snarey & Dollahite, 2001) and also may be more strongly influenced by the
religion of their parents (Erickson, 1992). On the other hand, males exhibit much higher
levels of deviant behavior (Deater-Deckard et al., 1998). Maternal religiosity seems to
lead to a much stronger mother-child bond (Pearce & Axinn, 1998). There is evidence
that the developmental processes involving parental caregiving and child externalizing
behavior (Rothbaum & Weisz, 1994) as well as the relationship between religiosity and
deviance (Rohrbaugh & Jessor, 1975) do not differ by sex. However, there have also been
studies showing significantly complex patterns of sex moderation (Albrecht et al., 1977;
Chadwick & Top, 1993). In meta-analytic work, some results have supported the idea of
no sex differences (Rothbaum & Weisz, 1994) while other results have shown that
perhaps the effect of religiosity on deviance is slightly stronger for males than females
(Baier & Wright, 2001).
Measurement
Though measurement is almost always a point of contention among researchers in
various fields of study, it has played a particularly consequential role in the study of
89
religion and deviance (e.g., Gorsuch & McFarland, 1972). Several of the subsequent
criticisms of Hirschi and Stark?s (1969) seminal work have been methodological in focus.
In particular, their measure of religiosity has come under considerable fire (Knudten &
Knudten, 1971). More specifically, they employed a single-item measure assessing
frequency of church attendance. They defended this choice by stating that the usual
measure of personal religiosity ?in studies of the effects of religion on delinquency is a
measure of church attendance. In our opinion, this is as it should be. The view that church
attendance should reduce delinquency is accepted by sociologists, layman [sic], and the
clergy . . .? (p. 205).
Subsequent research, however, has shown that single-item measures of church
attendance, though sometimes still effective (e.g., Burkett & White, 1974; Higgins &
Albrecht, 1977; Jensen & Erickson, 1979, to name only a few), can often be insufficient
for capturing the essence of the type of individual religious commitment that results in
lifestyle changes and behavioral control.
It has been cogently argued in the psychology of religion literature that, given the
complexity of religious commitment, single measures simply are inadequate
(Fulton, 1997; Gorsuch, 1988; Newcomb & Bentler, 1989). Stated succinctly,
church attendance, like attendance in a college classroom, is at best a vicarious
indicator of commitment and a poor measure of performance. Church attendance,
especially among adolescents, is often a result of parental pressure, social gains,
and a belief that attendance leads to eternal security rather than a deep personal
90
commitment to doctrine that restrains behavior like drug use (Corwyn & Benda,
2000, p. 253).
Based on their review of the literature, these same authors concluded that ?studies
that fail to find support for a relationship between religion and delinquency . . . almost
invariably operationalize religiosity with single-item measures? (Corwyn & Benda, 2000,
p. 253). This conclusion finds empirical support in Johnson et al.?s (2000) systematic
review of the literature. First, they found that studies utilizing multiple measures of
religiosity (i.e., four or more) always reported a significant negative relationship with
deviance. Second, they found that studies assessing the reliability of their measure(s) of
religiosity (i.e., multi-item scales) always reported a significant negative effect as well.
Apparently, multi-item, multi-dimensional measurement of religiosity produces more
valid, more reliable, and more consistent results, namely that religion is an effective
deterrent against deviant behavior.
In attempts to clarify this issue, some researchers have used both single-item
variables and multi-item scales in a side-by-side comparison within the same study,
hoping to find that one is more effective than the other or that both are equally effective
(e.g., Benda & Toombs, 2000; Evans et al., 1995). Some have even found that other
single-item measures (e.g., religious salience) are better than (e.g., Elifson et al., 1983) or
at least equal to (e.g., Sloane & Potvin, 1986) a single-item measure of church attendance.
In any case, many researchers have self-acknowledged the deficiencies of their single-
item measures for tapping the full domain of the concept of religiosity (Cochran, 1989;
Higgins & Albrecht, 1977; Rohrbaugh & Jessor, 1975; Tittle & Welch, 1983). More
91
recent studies employing more sophisticated designs and improved measurement
frequently highlight the weakness of the use of single-item measures of religion in
previous literature (Benda & Corwyn, 1997; Benda & Toombs, 2000; Corwyn & Benda,
2000; Johnson et al., 2001).
Summary of Other Intervening Factors
The current investigation seeks to examine three potentially important peripheral
issues ? type of deviance, developmental issues, and measurement. First, some have
advocated that religiosity carries with it a unique capacity for social control only in
regards to behaviors that are labeled deviant by churches or religious organizations
(Middleton & Putney, 1962). In addition, this antiascetic hypothesis has been
reoperationalized to mean that antiascetic deviance is behavior which is not harmful to
others, but is, rather, victimless (Burkett & White, 1974). There has been considerable
confusion, however, as to whether this hypothesis is correct, with large quantities of
evidence for both sides (e.g., Jensen & Erickson, 1979, contains evidence for both within
one study).
Second, age and sex are already known to be important developmental variables
in the study of religiosity (Bachman et al., 2002; Fiese & Tomcho, 2001), family (Laursen
et al., 1998), and deviance (Deater-Deckard et al., 1998; Farrington, 1986). Additionally,
some researchers have investigated their role in studies involving more than one of the
key variables in the present examination (e.g., Benda & Toombs, 2000; Pearce & Axinn,
1998). There have even been meta-analyses that have sought to include age and sex in an
effort to better understand their multivariate role (Baier & Wright, 2001; Rothbaum &
92
Weisz, 1994). However, apparently no study has brought together all of these variables ?
religiosity, family processes, deviant behavior ? and tested their developmental salience
by age or sex.
Third, work in this area has largely utilized single-item measures of church
attendance or some other religious construct (e.g., Hirschi & Stark, 1969). Unfortunately,
formal church participation is not a very good measure of personal religious commitment
(Corwyn & Benda, 2000). Along with this, multi-item measurement has been shown to be
more reliable and powerful in uncovering the effect of religiosity on deviance (Johnson et
al., 2000; Johnson et al., 2001). Because of the inconsistencies and unanswered questions
relating to these three areas, the current investigation will attempt to include them in
analyses with the hopes that clarification can be acquired.
Rationale for the Current Investigation
It is clear from the literature that religiosity acts consistently as a moderate
deterrent to deviant behavior. However, it is not clear from the literature under what
circumstances this relationship may be diminished or magnified. Part of the problem has
been lack of sufficient measurement and lack of sophisticated analytical techniques. In
addition, no clarity exists for how this relationship may vary by age or gender. Therefore,
the current study will attempt to add to the literature in the following ways:
1. By providing a side-by-side comparison of various single- and multi-item
measures of religiosity in order to separately assess their independent predictive strength
regarding adolescent deviant behavior;
93
2. By using a multi-dimensional conceptualization of family processes, involving
several multi-item subscales;
3. By using a multi-dimensional conceptualization of deviant behavior, involving
several multi-item subscales;
4. By conducting a test of the mediation role of family processes between
religiosity and deviance;
5. By conducting a test of the moderation effect of religiosity between family
processes and deviance;
6. By conducting a test of a possible interaction between family processes and
religiosity with deviance;
7. By conducting these tests separately by age group and gender in order to seek
some developmental clarification regarding these relationships;
8. By conducting these tests using structural equation modeling in order to apply a
more sophisticated analytical lens to these tests (only a limited set of studies have utilized
this analytical approach ? Benda & Whiteside, 1995; Brody & Flor, 1998; Brody et al.,
1996; Brody et al., 1994; Cochran et al., 1994; Johnson et al., 2001).
9. By testing several competing structural models in order to ascertain which may
be the best representation of the data.
Hypotheses
Based on this review of literature, the following hypotheses were developed and
tested in this study.
94
Hypothesis 1: Bivariate Relationships
It was expected that (A) all religious measures would have a significant negative
bivariate relationship with deviant behavior (Baier & Wright, 2001; Johnson et al., 2000)
and that (B) this deterrent effect would be stronger for male adolescents (Baier & Wright,
2001) and (C) older adolescents (Benda & Toombs, 2000).
Hypothesis 2: Measurement
It was expected that the religiosity scales would perform better than single-item
measures of religiosity in the prediction of deviant behavior.
Hypothesis 3: Type of Deviance
It was expected that (A) the religiosity scales would be more strongly related to
victimless deviance, but not unrelated to victim deviance (Chadwick & Top, 1993; Jensen
& Erickson, 1979), and that (B) this relationship would be true for both younger and older
adolescents (Rothbaum & Weisz, 1994). Furthermore, it was expected that (C) the
religiosity scales would be related to victimless deviance for both male and female
adolescents (Albrecht et al., 1977), but that (D) religious salience would be related to
victimless deviance only for male adolescents (Albrecht et al., 1977), while (E) both the
religiosity scales and religious salience would be related to victim deviance for females
only (Albrecht et al., 1977). It was also expected that (F) the religiosity scales would be
more strongly related to antiascetic deviance, but not unrelated to secular deviance
(Jensen & Erickson, 1979), and that (G) this relationship would be true for both male and
female adolescents (Rothbaum & Weisz, 1994), as well as (H) younger and older
adolescents (Rothbaum & Weisz, 1994).
95
Figure 6. Model showing family processes mediating religiosity single-item measures
(Hypothesis 4A) and ritualistic participation (Hypothesis 4B) where there is a significant
negative pathway from family processes to deviance and no significant direct pathway
from religion to deviance.
Hypothesis 4: Mediation
It was expected that (A) family processes would mediate the relationship between
all single-item religiosity measures and deviant behavior (Benda & Toombs, 2000; see
Figure 6), and (B) would also mediate the relationship between ritualistic participation
and deviant behavior (Benda & Toombs, 2000; see Figure 6), but (C) would not mediate
the relationship between relational practice and deviant behavior (Benda & Toombs,
2000; see Figure 7), or (D) between total religiosity and deviant behavior (Benda &
Toombs, 2000; see Figure 7), and that (E) this same mediation pattern would be true for
both male and female adolescents (Rothbaum & Weisz, 1994), as well as (F) younger and 
96
Figure 7. Model showing the partially mediating role of family processes between
relational practice and religiosity where negative direct and indirect pathways between
religion and deviance both exist.
older adolescents (Rothbaum & Weisz, 1994). In addition, it was expected that (G) the
mediating role of family processes between relational practice and total religiosity and
deviant behavior would be clarified as that of partial mediation with the religiosity scales
having both indirect (through family processes) and direct pathways to deviance (Johnson
et al., 2001; see Figure 7), and that (H) this would be equally true for both male and
female adolescents (Rothbaum & Weisz, 1994), as well as (I) younger and older
adolescents (Rothbaum & Weisz, 1994).
Hypothesis 5: Moderation
It was expected that (A) religiosity would moderate the relationship between
family processes and deviant behavior (Bahr et al., 1998; Brody et al., 1996; see
97
Figure 8. Model showing religiosity moderating between family processes and deviance
where the predictive pathways for the two groups (low and high religiosity) are not equal.
Figure 8), and that (B) this pattern of moderation would be true for both male and female
adolescents (Rothbaum & Weisz, 1994), as well as (C) younger and older adolescents
(Rothbaum & Weisz, 1994).
98
CHAPTER 3: METHOD
Sample
Data for this study were collected from a total of N = 877 adolescents who
attended a high school in the southeastern region of the United States. Twelve of the
participants were missing data regarding their sex and, therefore, had to be dropped from
the current study, resulting in a final sample of N = 865 (mean age = 16.4, sd = 1.2)
students. Four hundred and thirty-two of the participants were female (49.9%,
mean age = 16.3, sd = 1.2) and 433 were male (50.1%, mean age = 16.4, sd = 1.2). The
sample was predominantly Caucasian (72.4%) with 20.9% African-Americans and 6.1%
other ethnicities (.6% were missing ethnicity data). The majority of sample participants
were from two biological parent homes (65.7%), while 16.2% were living with one
biological parent only and 13.0% were in a stepfamily arrangement. The families of these
participants were mostly middle and upper middle class (74.2% reported family income
of $35,000 or more per year) whose primary wage-earner held a white-collar job (75.0%
reported a primary wage earner whose job was ?semi-professional? or higher). In
addition, their parents were well educated (64.1% of fathers and 58.6% of mothers had
college and/or graduate degrees). This high level of education and income was largely due
to the fact that data were collected in a university town.
99
Procedures
The data collection for this study was part of the International Study of
Adolescent Development (ISAD; Pickering & Vazsonyi, 2002; Vazsonyi & Pickering,
2000, 2003; Vazsonyi, Pickering, Belliston, Hessing, & Junger, 2002; Vazsonyi,
Pickering, Junger, & Hessing, 2001). A standard data collection protocol, approved by a
university Institutional Review Board (IRB), was followed that consisted of a self-report
data collection instrument which included instructions on how to complete the survey, a
description of the ISAD project, and assurances of anonymity and confidentiality. These
instructions were read verbatim in each classroom before the surveys were distributed.
The total student population at the high school was 1,134. All students in the school were
invited to participate and 920 (81% of the school population) did so. Forty-three (4% of
the school population, 5% of the surveyed sample) of the surveys were dropped from the
sample because they were incomplete, leaving a final sample of 877 (77% of the school
population, 95% of the surveyed sample) high school students.
Measures
Background Variables
Sex
Subjects were asked to indicate their sex on a single item: ?What is your gender??
Responses were given as 1 = male and 2 = female.
100
Age
Participants were asked to indicate the month (?In which month were you born??)
and year (?In what year were you born??) of their birth. Subjects? specific ages were
calculated based on the fifteenth day of the month.
Grade
Subjects were asked to indicate their grade level in school and responses were
provided as 1 = 9th grade, 2 = 10th grade, 3 = 11th grade, and 4 = 12th grade.
School Age
Initial frequencies involving crosstabulations of the age and grade variables
indicated the presence of a common problem, namely that students in each school grade
level represented a wide range of ages. For example, in the tenth grade, the age range
Table 1
Crosstabulations of age and grade variables
School Grade
9th grade 10th grade 11th grade 12th grade
13 year olds 1 1 0 0
14 year olds 118 0 0 0
15 year olds 123 111 0 0
16 year olds 26 84 123 2
17 year olds 1 13 89 100
18 year olds 0 1 6 53
19 year olds 0 0 1 12
101
spanned from 13 to 18 years old. A similar broad age range was evident in all of the
grades (see Table 1).
This produces a dilemma for conducting age-related analyses. Chronological age,
of course, represents, at the very least, the physical aspects of maturation, such as onset of
puberty. So, a 17 year old girl who is in ninth grade will be quite different physically from
Table 2
Frequencies and Descriptives of Dichotomous Age and School Age Variables
Age Variable School Age Variable
Younger
(n = 451)
Older
(n = 414)
Younger
(n = 435)
Older
(n = 430)
Grade 9 (.59) 11 (.73) 9 (.49) 11 (.64)
   9 259 10 269 0
   10 171 39 166 44
   11 21 198 0 219
   12 0 167 0 167
Age 15.4 (.60) 17.4 (.68) 15.4 (.61) 17.4 (.71)
   13 2 0 2 0
   14 118 0 118 0
   15 234 0 234 0
   16 97 138 80 155
   17 0 203 1 202
   18 0 60 0 60
   19 0 13 0 13
Mean School Age 21.4 (8.04) 57.3 (12.79)
Note. First numbers in title rows represent means (age and school age) or medians
(grade) and second number (in parentheses) represents sd.
102
4 There were 36 cases that would have been classified differently by using the age
variable rather than the school age variable to create the dichotomous (younger, older)
age variable. These cases ranged in age from 16.1 to 17.3. If the age variable had been
used, a select group of 26 adolescents who were 16.1 to 16.4 years of age would have
been placed into the younger group. However, all 26 of these adolescents were in the
tenth and eleventh grades; consequently, when the school age variable was used, they
were placed into the older group. Along the same line, a separate group of 10 adolescents
her 13 year old classmates. In addition, though she attends classes with ninth graders, she
may not associate with them outside of school, choosing rather to have friends that match
her chronological age. Thus, combining them (13-17 year olds) into one analytical group
(i.e., ?ninth graders?) seems inappropriate.
At the same time, however, her developmental experience (e.g., social
development, emotional development, etc.) will inevitably be shaped by the fact that she
shares the daily routine and circumstances of ninth graders which are very different from
those of most 17 year olds. In addition, she interacts with these ninth graders for hours at
a time every day, is treated like a ninth grader by the teachers and administration of the
school, and she may even ?hang out? with other ninth graders in her extracurricular social
life. Thus, including her in an analytical group composed of others of her chronological
age (i.e., ?17 year olds?) seems equally inappropriate. To solve this dilemma, a variable
called school age was computed as an interaction (i.e., product term) between age and
grade. This new school age variable was correlated r = .91 with the ?age-only? variable.
In order to create a categorical variable (i.e., older and younger adolescents) for use in
developmental analyses, a median split of the school age variable was used. When a
comparative categorical variable was computed using the ?age-only? variable, frequencies
indicated only a few minor differences with the school age variable.4
103
who were 16.5 to 17.3 years of age would have been placed into the older group, using
the age variable, but, because they were all in ninth grade, they were placed into the
younger group when the school age variable was used.
Race
Participants were asked to give only one response to the following question:
?What is your ethnic/racial background or heritage?? Responses were given as
1. Caucasian (White), 2. African American, 3. Native American,  (e.g., Cherokee, Ute,
Cheyenne), 4. Asian American (e.g., Chinese, Japanese, Vietnamese, Korean),
5. Hispanic (e.g., Puerto Rican, Mexican, Cuban), 6. Pacific Islander (e.g., Hawaiian,
Samoan, Tongan). For regression analyses, dummy-coded race variables were used.
SES
Two indicators of social status (family income and primary wage earner
occupation) were combined to form a measure of socioeconomic status (SES).
Family Income
The income level of each participant?s family was measured by asking them to
?Please pick one of the following choices describing your family?s approximate total
annual income? using the following response options: 1. $20,000 or less, 2. $20,000 to
$35,000, 3. $35,000 to $60,000, 4. $60,000 to $100,000, 5. $100,000 or more.
Primary Wage Earner?s Occupation
Subjects were asked to indicate the type of work performed by the primary wage
earner in their family. Six categories, collapsed from Hollingshead?s (1975) original nine
categories, were specified. Each category contained descriptions of sample jobs which
would fit into each of them. Responses were given by indicating the number of the
104
category which contained the closest or most accurate description of the family?s primary
wage earner?s job. The categories, listed here with condensed descriptions, were as
follows: 1 = owner of a large business, executive; 2 = owner of a small business,
professional; 3 = semi-professional, skilled laborer; 4 = clerical staff; 5 = semiskilled
laborer; and 6 = laborer or service worker. Responses for this item were reverse-coded so
that higher numbers represented higher social class standing.
SES Scale
An established procedure was used to combine these two measures (e.g., Capaldi,
Stoolmiller, Clark, & Owen, 2002; Ruiz, Roosa, & Gonzales, 2002). First, both items
were correlated. Results indicated that the bivariate correlation between family income
and primary wage earner?s occupation was r = .54 (p = .000). Second, because this
analysis indicated a significant relationship between these items, these two items were
standardized and combined. Cronbach?s alpha reliability statistics for this scale were as
follows: for male adolescents, " = .69; for female adolescents, " = .71; for younger
adolescents, " = .72; for older adolescents, " = .69; for total sample, " = .70.
Religiosity
Because the literature in this area has involved considerable debate over the use of
single- versus multi-item measures (e.g., Corwyn & Benda, 2000; Denton & Smith, 2001,
Gorsuch & McFarland, 1972; Johnson et al., 2000), the current investigation included
both. Therefore, 8 item level measures were used as well as 2 multi-item subscales and
one total religiosity scale.
105
Single Item Measures
Church Attendance
Church attendance was measured by asking participants to respond to the
question, ?How often do you attend the regularly scheduled services of a church (not
including weddings, funerals, baptisms, or any other special occasions of a religious
nature)?? using the following answers: 1 = practically never, 2 = only on holidays or holy
days, 3 = only occasionally, 4 = once a week, 5 = more than once a week.
Church Status
An item measuring church status combined the concepts of church membership
and church attendance by asking, ?What is your current status in relation to church??
Responses to this item were given as 1 = no association with church, 2 = attending
nonmember, 3 = nonattending member, 4 = attending member. Because actual attendance
at church services was considered a better indicator of commitment to a religious
organization than simple membership, the second and third choices were reverse-coded
so that higher values on this item would reflect incrementally greater levels of
commitment.
Religious Involvement
An item asking, ?What is your current level of involvement with a church or
religious organization?? was used to measure the level of involvement in religious
volunteer work. Possible responses to this item were: 1 = I never participate in
church/religious voluntary work/ministry, 2 = I rarely participate in church/religious
voluntary work/ministry, 3 = I occasionally participate in church/religious voluntary
106
work/ministry, 4 = I frequently participate in church/religious voluntary work/ministry,
5 = I am a paid staff member of a church/religious organization.
Bible Reading
A measure of Bible reading was included which asked, ?How often do you read
the Bible or any other book of religious faith?? Possible responses to this item were:
1 = never, 2 = infrequently, 3 = sometimes, 4 = quite a bit, 5 = every day.
Praying
The amount that participants engaged in prayer was assessed by asking, ?How
often do you pray?? Respondents could answer: 1 = never, 2 = infrequently,
3 = sometimes, 4 = quite a bit, 5 = every day.
Music Listening
Participants were asked to indicate how much they listened to religiously-oriented
music by responding to the question, ?How often do you listen to Christian music
(e.g., Steven Curtis Chapman, Michael W. Smith, DC Talk, Petra, Glad, Twila Paris, etc.)
for personal pleasure?? with any of the following answers: 1 = never, 2 = infrequently,
3 = sometimes, 4 = quite a bit, 5 = every day.
Religious Conviction
A measure of religious conviction (originally used by Allport and Ross, 1967, and
called ?extrinsic religiosity?) was included in the current study. The item states,
?Occasionally I find it necessary to compromise my religious beliefs in order to protect
my social and economic well-being.? Responses to this item were made on a 4-point
Likert-type scale: 1 = strongly disagree, 2 = disagree, 3 = agree, 4 = strongly agree. Scores
107
on this item were reverse-coded so that higher values would represent higher levels of
religiosity (i.e., lower levels of compromise of religious beliefs).
Religious Salience
A measure of religious salience (also used by Allport and Ross, 1967, and called
?intrinsic religiosity?) was also included. Participants were asked to respond to the
statement, ?I try hard to carry my religion over into all my other dealings in life? using the
same 4-point Likert-type scale: 1 = strongly disagree, 2 = disagree, 3 = agree, 4 = strongly
agree.
Multi-Item Scales
Two subscales, each measuring distinct dimensions of religiosity, were included
as well as a total religiosity scale. The conceptualization of these two subscales represent
the potential dimensional nature of religiosity, conceiving of religiosity as either
ritualistic or relational.
Ritualistic Participation
The church attendance, Bible reading, and religious salience items were summed
to form a ritualistic participation variable. This score represented an individual?s focus
on religion as a status to be maintained that has, associated with it, certain ritualistic
practices to be observed. Cronbach?s alpha reliability statistics for this scale ranged from
" = .67 to " = .74 among the study groups (see Table 3).
Relational Practice
The church status, religious involvement, and praying items were combined to
form a relational practice score, representing an individual?s conception of religion as a
108
Table 3
Reliability Statistics for Study Scales by Sex, Age, and for Total Sample
Total Males Females Younger Older
SES .70 .69 .71 .72 .69
Family Processes
     Maternal
          Closeness .84 .85 .83 .84 .84
          Support .84 .82 .85 .82 .85
          Monitoring .79 .77 .79 .74 .83
          Communication .85 .82 .87 .85 .86
          Conflict .85 .85 .85 .84 .85
          Peer Approval .83 .82 .83 .84 .81
Family Processes
     Paternal
          Closeness .89 .90 .88 .89 .90
          Support .82 .82 .84 .82 .82
          Monitoring .89 .87 .90 .87 .90
          Communication .89 .87 .90 .90 .88
          Conflict .86 .83 .88 .86 .86
          Peer Approval .87 .87 .89 .86 .89
Religiosity
     Ritualistic Participation .71 .67 .74 .68 .74
     Relational Practice .75 .74 .75 .74 .76
     Religiosity .86 .86 .87 .85 .87
Deviance
     Vandalism .89 .90 .85 .90 .88
     Alcohol Use .89 .91 .88 .88 .90
     Drug Use .92 .93 .90 .92 .92
     School misconduct .86 .87 .84 .86 .86
     General .88 .89 .83 .89 .87
     Theft .87 .88 .84 .87 .88
     Assault .82 .83 .78 .82 .82
     Total .97 .97 .96 .97 .97
     Victim .91 .91 .87 .90 .91
     Victimless .96 .96 .95 .96 .96
     Antiascetic .94 .95 .94 .94 .94
     Secular .96 .96 .94 .96 .96
109
relationship with a supernatural being that is accompanied by certain heartfelt practices.
Cronbach?s alpha reliability statistics for this scale ranged from " = .74 to " = .76 among
the study groups (see Table 3). For SEM analyses involving tests of moderation, the
relational practice scale was used to form a dichotomous variable. A median split of the
standardized residualized relational practice variable (after controlling for SES and race)
was used to create two groups (i.e., low and high religiosity).
Religiosity
Seven items, including the six contained in the above two subscales plus the
music listening item, were included in a total religiosity scale. Cronbach?s alpha
reliability statistics for this scale ranged from " = .85 to " = .87 among the study groups
(see Table 3).
Family Processes
Family processes were measured using the Adolescent Family Process (AFP)
measure (Vazsonyi, Hibbert, & Snider, 2003; see Appendix A). The instrument includes a
total of 50 items in 6 subscales consisting of 25 identical items worded in a sex-specific
manner (i.e., ?she? or ?he?) in order to assess both maternal and paternal family
processes. The first three subscales (Closeness, Support, Monitoring) were rated on the
following 5-point Likert-type response scale: 1 = strongly disagree, 2 = disagree,
3 = neither disagree nor agree, 4 = agree, 5 = strongly agree. The other three subscales
(Communication, Conflict, Peer Approval) were rated on the following 5-point Likert-
type response scale: 1 = never, 2 = occasionally, 3 = sometimes, 4 = often, 5 = very often.
110
Closeness
The closeness subscale consisted of 6 items (sample item: ?My mother [father]
gives me the right amount of affection?) measuring the perceived amount of affective
warmth received from parents. Previous work has shown this scale to have good
reliability for both mothers (" = .85) and fathers (" = .89; Vazsonyi et al., 2003). In the
current study, Cronbach?s alpha reliability statistics for this scale ranged from " = .83 to
" = .85 for maternal closeness and from " = .88 to " = .90 for paternal closeness among
the study groups (see Table 3).
Support
The support subscale consisted of 4 items (sample item: ?My mother [father] puts
me down in front of other people?) assessing the perceived amount of psychological
support received from parents. Responses for this subscale were reverse coded so that a
higher score indicated a higher level of support. Previous work has shown this scale to
have good reliability for both mothers (" = .83) and fathers (" = .82; Vazsonyi et al.,
2003). In the current study, Cronbach?s alpha reliability statistics for this scale ranged
from " = .82 to " = .85 for maternal closeness and from " = .82 to " = .84 for paternal
closeness among the study groups (see Table 3).
Monitoring
The monitoring subscale consisted of 4 items (sample item: ?When I am not at
home, my mother [father] knows my whereabouts?) indicating the perceived amount of
behavioral and activity supervision provided by parents. Previous work has shown this
scale to have good reliability for both mothers (" = .78) and fathers (" = .89; Vazsonyi et
111
al., 2003). In the current study, Cronbach?s alpha reliability statistics for this scale ranged
from " = .74 to " = .83 for maternal closeness and from " = .87 to " = .90 for paternal
closeness among the study groups (see Table 3).
Communication
The communication subscale consisted of 5 items (sample item: ?How often do
you talk to your mother [father] about things that are important to you?) measuring the
perceived ability to discuss matters of a personal nature with parents. Previous work has
shown this scale to have good reliability for both mothers (" = .88) and fathers (" = .91;
Vazsonyi et al., 2003). In the current study, Cronbach?s alpha reliability statistics for this
scale ranged from " = .82 to " = .87 for maternal closeness and from " = .87 to " = .90
for paternal closeness among the study groups (see Table 3).
Conflict
The conflict subscale consisted of 3 items (sample item: ?How often do you have
disagreements or arguments with your mother [father]??) assessing the perceived amount
of conflict with parents. Previous work has shown this scale to have good reliability for
both mothers (" = .83) and fathers (" = .86; Vazsonyi et al., 2003). In the current study,
Cronbach?s alpha reliability statistics for this scale ranged from " = .84 to " = .85 for
maternal closeness and from " = .83 to " = .88 for paternal closeness among the study
groups (see Table 3).
Peer Approval
The peer approval subscale consisted of 3 items (sample item: ?How often does
your mother [father] approve of your friends??) indicating the perceived amount of
112
parental approval for the adolescent?s choice of friends. Previous work has shown this
scale to have good reliability for both mothers (" = .83) and fathers (" = .87; Vazsonyi et
al., 2003). In the current study, Cronbach?s alpha reliability statistics for this scale ranged
from " = .81 to " = .84 for maternal closeness and from " = .86 to " = .89 for paternal
closeness among the study groups (see Table 3).
Deviance
Based on various conceptualizations of deviant behavior in previous literature,
deviance was operationalized in six different ways in the current study.
Normative Deviance
Lifetime deviance was measured by the 55-item Normative Deviance Scale (NDS;
Vazsonyi & Killias, 2001; Vazsonyi & Pickering, 2000; Vazsonyi et al., 2001; see
Appendix C). The scale was conceptualized to measure acts of normative deviance in
general adolescent populations and, therefore, examined a broader spectrum of deviant
activities than just status and index offenses. This conceptualization of deviance is
consistent with results from nationally representative data sets (e.g., the National Youth
Survey, Huizinga, Menard, & Elliott, 1989) which report that over 90% of sampled males
and females indicate having committed at least one delinquent act at some time in their
life. Very few such self-report scales that include multi-item subscales with psychometric
properties have been developed (single-item measures are more common; that is, one
item measuring vandalism, for example).
The current investigation utilized all seven subscales of the NDS (vandalism,
alcohol, drugs, school misconduct, general deviance, theft, and assault). Responses for all
113
items in the NDS were given on a 5-point Likert-type scale and identified lifetime
frequency of specific deviant behaviors (1 = never, 2 = one time, 3 = 2-3 times, 4 = 4-6
times, and 5 = more than 6 times). In addition, all 55 items were summed then averaged
to compute a measure of total deviance.
Vandalism
The vandalism subscale consisted of 8 items (sample item: ?Have you ever
committed acts of vandalism when coming or going to a football game or other sports
event??) measuring participation in vandalism behaviors at home, school, work, and/or
other public places. Previous work has shown this scale to have good reliability (" = .84;
Vazsonyi et al., 2001). In the current study, Cronbach?s alpha reliability statistics for this
scale ranged from " = .85 to " = .90 among the study groups (see Table 3).
Alcohol Use
The alcohol use subscale consisted of 7 items (sample item: ?Have you ever
consumed alcoholic beverages [e.g., beer, wine, or wine coolers] before you were 21??)
assessing consumption of soft and hard liquor, drunkenness, and illegal acquisition of
alcohol. Previous work has shown this scale to have good reliability (" = .84; Vazsonyi et
al., 2001). In the current study, Cronbach?s alpha reliability statistics for this scale ranged
from " = .88 to " = .91 among the study groups (see Table 3).
Drug Use
The drug use subscale consisted of 9 items (sample item: ?Have you ever used
?soft? drugs such as marijuana (grass, pot)??) examining tobacco, marijuana, and ?hard?
drug usage as well as attendance at school, work, concerts, or parties while under the
114
influence of a substance, acquisition of a substance at a party, and selling of a substance.
Previous work has shown this scale to have good reliability (" = .89; Vazsonyi et al.,
2001). In the current study, Cronbach?s alpha reliability statistics for this scale ranged
from " = .90 to " = .93 among the study groups (see Table 3).
School Misconduct
The school misconduct subscale consisted of 7 items (sample item: ?Have you
ever been in trouble at school so that your parents received a phone call about it??)
dealing with cheating, class or school discipline (e.g., suspension or expulsion), and
truancy. Previous work has shown this scale to have good reliability (" = .76; Vazsonyi et
al., 2001). In the current study, Cronbach?s alpha reliability statistics for this scale ranged
from " = .84 to " = .87 among the study groups (see Table 3).
General Deviance
The general deviance subscale consisted of 11 items (sample item: ?Have you
ever shaken/hit a parked car just to turn on the car's alarm??) measuring a wide range of
deviant activities including traffic moving violations (e.g., speeding, running a red light),
trespassing, under-age nightclub attendance, and other general mischief. Previous work
has shown this scale to have good reliability (" = .81; Vazsonyi et al., 2001). In the
current study, Cronbach?s alpha reliability statistics for this scale ranged from " = .83 to
" = .89 among the study groups (see Table 3).
Theft
The theft subscale consisted of 7 items (sample item: ?Have you ever stolen,
taken, or tried to take something worth $10 or less (e.g., newspaper, pack of gum, mail,
115
money, etc.)??) assessing petty and major theft (including grand theft auto) from family,
other people, or of public property. Previous work has shown this scale to have good
reliability (" = .83; Vazsonyi et al., 2001). In the current study, Cronbach?s alpha
reliability statistics for this scale ranged from " = .84 to " = .88 among the study groups
(see Table 3).
Assault
The assault subscale consisted of 6 items (sample item: ?Have you ever hit or
threatened to hit a person??) examining several different physical violence scenarios,
including gang violence. Previous work has shown this scale to have good reliability
(" = .76; Vazsonyi et al., 2001). In the current study, Cronbach?s alpha reliability
statistics for this scale ranged from " = .78 to " = .83 among the study groups (see Table
3).
Total Deviance
A total deviance scale was also used that included all 55 deviance items. Previous
work has shown this scale to have good reliability (" = .95; Vazsonyi et al., 2001). In the
current study, Cronbach?s alpha reliability statistics for this scale ranged from " = .96 to
" = .97 among the study groups (see Table 3).
Victim Deviance
In order to assess deviant behavior harmful to other persons, the items in the theft
and assault subscales were combined to form a measure of victim deviance. Cronbach?s
alpha reliability statistics for this scale ranged from " = .87 to " = .91 among the study
groups (see Table 3).
116
Victimless Deviance
Deviant behavior that is not directly harmful to other persons was measured by
combining the items in the alcohol use, drug use, school misconduct, and general
deviance subscales. Cronbach?s alpha reliability statistics for this scale ranged from
" = .95 to " = .96 among the study groups (see Table 3).
Antiascetic Deviance
To form a measure of deviant behavior that is condemned more strongly by
churches and religious organizations than by secular institutions, the items in the alcohol
use and drug use subscales were combined. Cronbach?s alpha reliability statistics for this
scale ranged from " = .94 to " = .95 among the study groups (see Table 3).
Secular Deviance
A measure of deviant behavior that is condemned equally by both religious and
secular organizations was formed by combining the items in the vandalism, school
misconduct, general deviance, theft, and assault subscales. Cronbach?s alpha reliability
statistics for this scale ranged from " = .94 to " = .96 among the study groups (Table 3).
Plan of Analysis
Initial Analyses
Scale Construction
As a first step, 28 scales (SES, 12 family processes, 3 religiosity, 12 deviance)
were developed. Some of these scales (i.e., AFP, NDS) were already established, others
were reconfigurations of existing reliable scales (i.e., victim/victimless deviance,
antiascetic/secular deviance), and still others were newly constructed for this study
117
(i.e., SES, religiosity). The results of reliability analyses for all scales are found in Table
3.
Measurement Models
Subsequent SEM analyses included two elements (i.e., family processes and
deviance) that called for the initial construction of measurement models.
First, based on the statistical evidence that both the maternal and paternal
measures of each family processes subscale were strongly related (ranging from r = .32 to
r = .54, average r = .43, all significant at p < .001 level) and the conceptual assumption
that, together, they may represent an average family processes score based on input from
both parents (Buysse, 1997; Florsheim, Tolan, & Gorman-Smith, 1998; Tolan et al.,
1997), each of the family processes subscales was conceptualized in the structural model
as a latent construct being indicated by two manifest variables ? the maternal and paternal
elements of each subscale (for rationale concerning the aggregation of data across
sources, see Gorman-Smith, Tolan, Zelli, & Huesmann, 1996, and Tolan et al., 1997; for
an example of such in latent modeling, see Duncan, Duncan, Biglan, & Ary, 1998).
Consequently, it was not possible to test a measurement model for the family processes
elements of the structural model since this measurement model was unidentified due to
negative degrees of freedom (5 distinct sample moments, 6 distinct parameters to be
estimated, df = 5-6 or -1). However, this did not pose a problem since preliminary
correlations between maternal and paternal subscales provided sufficient justification for
this approach.
118
Second, a measurement model for the deviance construct was tested. This
measurement model included all 7 of the deviance subscales (vandalism, alcohol use,
drug use, school behavior, general deviance, theft, assault) as manifest indicators of a
latent deviance construct based on previous work (Vazsonyi et al., 2001; see Figure 12).
An initial test of this model showed poor fit to the data (P2 = 435.60, df = 14, p = .000,
CFI = .895, RMSEA = .20). In such cases, the use of conceptually-substantiated
correlated errors is viewed as acceptable in order to improve model fit (Peter Bentler,
personal communication, November 10, 1998; Byrne, 1994; Hayduk, 1987; Hoyle, 1995). 
Since this model represented a 55-item measure that formed 7 subscales, each of which
was simply a variant of deviance and all of which were part of the total deviance measure,
and since all of these items were obtained through mono-method assessment (i.e., all
were self-report survey items), there was an a priori expectation that error terms would
inevitably be related, thus providing sufficient conceptual justification to allow all 7 of
the subscales to covary (a total of 21 correlated error terms). However, in the interest of
simplicity, modification indices were used to add 6 correlated errors to the model (alcohol
use [e2] to drug use [e3]; theft [e6] to assault [e7]; alcohol use [e2] to school misconduct
[e4]; drug use [e3] to school misconduct [e4]; school misconduct [e4] to assault [e7];
vandalism [e1] to assault [e7]; see Table 4). These modifications were made in a stepwise
procedure, one at a time, implementing the modification that would result in the most
statistically significant improvement to the overall model fit, then running the model
again and looking at the modification indices again, etc. The introduction of these
modifications to the model resulted in a final measurement model with excellent fit
119
5 Modification indices (MI = 5.74, EPC = .02) indicated that the freeing of one
more parameter (e2 to e5) would result in not only another significant change in chi-
square (model P2 = 10.75, model df = 7, )P2 = 7.25, )df = 1, p for change = < .01), but
also in an overall final model that would have excellent fit in every way, including a
nonsignificant chi-square (model CFI = .999, model RMSEA = .027, p for model = .15).
However, for simplicity sake, this additional parameter was not added to the model.
(P2 = 18.00, df = 8, p = .021, CFI = .998, RMSEA = .042). As can be seen in the table,
each incremental change to the model represented a significant chi-square change, as did
the difference between the initial (P2 = 435.60) and final (P2 = 18.00) model
()P2 = 417.60, )df = 6, p < .001).5
Analysis Preparation
Because there is an ongoing debate concerning the role of SES and/or race in
adolescent development (see Amato & Fowler, 2002, pp. 704-706 for a concise summary
of this debate), it was important to investigate the potential impact of these demographic 
Table 4
Fit Indices Showing Stepwise Model Improvement for Deviance Measurement Model
Using Correlated Errors Recommended by Modification Indices
Parameter MI EPC
Model
P2
Model
df
Model
p
Model
CFI
Model
RMSEA )P2 )df
p for
change
e2 ? e3 252.77 .33 144.88 13 .000 .967 .118 290.80 1 <.001
e6 ? e7 43.70 .06 97.06 12 .000 .979 .099 47.82 1 <.001
e2 ? e4 26.11 .08 70.14 11 .000 .985 .086 26.92 1 <.001
e3 ? e4 18.43 .05 37.28 10 .000 .993 .061 32.86 1 <.001
e4 ? e7 7.84 .03 28.74 9 .001 .995 .055 8.54 1 <.01
e1 ? e7 9.77 .03 18.00 8 .021 .998 .042 10.74 1 <.01
Note. MI = Modification Index; EPC = Estimated Parameter Change
120
Table 5
Summary of Study Variables
Background Variables Family Processes
(Maternal and Paternal)
Sex
Age
Grade
School Age
Race
SES
Closeness
Support
Monitoring
Communication
Conflict
Peer Approval
Religiosity Deviance
Church Attendance
Church Status
Religious Involvement
Bible Reading
Praying
Music Listening
Religious Conviction
Religious Salience
Ritualistic Participation
Relational Practice
Religiosity
Vandalism
Alcohol Use
Drug Use
School misconduct
General Deviance
Theft
Assault
Total Deviance
Victim Deviance
Victimless Deviance
Antiascetic Deviance
Secular Deviance
variables in the current study. In addition, the current investigation proposed to study
several key issues regarding the relationship between religiosity, family processes, and
deviance separately for male and female and younger and older adolescents. Bivariate
correlations were used to establish whether there was any important relationship between
these demographic variables (SES, race, sex, age) and both the independent and
dependent variables. Race was tested as a dichotomous variable (Caucasian/African
American) since there were too few participants from other racial groups. Results of these
121
analyses are found in Tables 6, 7, and 8, and indicate that these demographic variables are
sufficiently associated with both independent and dependent variables to warrant utilizing
them as either controls or grouping variables to more closely examine their effects on the
relationships between measures of religiosity, family processes, and deviance.
Initial Descriptive Statistics
As a final initial step, several descriptive statistics were computed. The results of
these initial sample-descriptive analyses are found in Tables 9 and 10. In addition, within-
construct correlation matrices for measures of religiosity, family processes, and deviance
can be found in Appendices D, E, and F. Finally, correlations between family processes
and both religiosity and deviance can be found in Appendices G and H.
Analyses Related to Hypotheses
Because initial analyses indicated that SES and race were associated with several
key study variables, in all subsequent analyses, both SES and race were used as controls.
Hypothesis 1
This hypothesis was tested using bivariate correlations. Part A was tested by
computing Pearson?s r correlations for all 11 measures of religiosity with all 8 measures
of normative deviance. Parts B and C were tested by computing the same correlations
separately by group (male and female adolescents, younger and older adolescents) then
using the Fisher?s r to z transformation statistical test to compare whether the male 
adolescents? correlations were stronger than the females? and whether the older
adolescents? correlations were stronger than the younger adolescents?.
122
Table 6
Correlations Between Demographic Variables and Measures of Religiosity
m (sd) SES Race Sex Age
Single-Item Measures
   Church Attendance 3.2 (1.4) .09** .07 .06 -.01
   Church Status 3.0 (1.2) .13*** .10** .11** -.05
   Religious Involvement 2.7 (1.2) .11** .04 .11** -.01
   Bible Reading 2.5 (1.2) .09* .05 .06 -.01
   Praying 3.3 (1.4) .03 .07 .09** -.05
   Music Listening 2.0 (1.3) -.09* .27*** .09* -.01
   Religious Conviction 2.1 (1.0) -.12** .14** -.05 -.07*
   Religion Salience 2.6 (1.0) .01 .08* .08* -.02
Multi-Item Measures
   Ritualistic Participation 2.8 (1.0) .08* .09* .08* -.02
   Relational Practice 3.0 (1.0) .10** .07* .12*** -.04
   Religiosity 2.8 (0.9) .07 .13** .12** -.03
Note. * (p < .05), ** (p < .01), or *** (p < .001).
Hypothesis 2
This hypothesis was also tested by using bivariate correlations. The total deviance
scale was correlated with all 11 religiosity measures. Correlations between each
religiosity measure and the total deviance scale were compared using the Fisher?s r to z
test to determine whether the multi-item scalar measures were more strongly related to
deviant behavior than the single-item measures.
123
Table 7
Correlations Between Demographic Variables and Measures of Family Processes
m (sd) SES Race Sex Age
Maternal
     Closeness 3.9 (0.9) .15*** -.03 .09** -.02
     Support 3.7 (1.1) .08* -.13*** .04 .10**
     Monitoring 3.8 (0.9) .05 .03 .25*** -.10**
     Communication 3.2 (1.0) .07* -.07 .20*** .02
     Conflict 2.8 (1.0) -.02 .03 .09* -.12**
     Peer Approval 3.9 (1.0) .20*** -.20*** .13*** .03
Paternal
     Closeness 3.6 (1.0) .29*** -.25*** .02 -.05
     Support 3.8 (1.1) .08* -.08* .02 .05
     Monitoring 3.2 (1.1) .14*** -.12** .10** -.10**
     Communication 2.8 (1.1) .22*** -.15*** .01 .00
     Conflict 2.5 (1.1) .01 .02 .06 -.07*
     Peer Approval 3.6 (1.1) .21*** -.26*** .05 .00
Note. * (p < .05), ** (p < .01), or *** (p < .001).
Hypothesis 3
This hypothesis was also tested by using bivariate correlations. Part A was tested
by checking whether the relationship between the religiosity scales and victimless
deviance was stronger than the relationship between the religiosity scales and victim
deviance (using Fisher?s r to z transformation test), though the latter relationship should
also be statistically significant. Part B involved running the same tests separately for
124
Table 8
Correlations Between Demographic Variables and Measures of Deviance
m (sd) SES Race Sex Age
Vandalism 1.6 (0.8) -.03 -.00 -.31*** .09**
Alcohol Use 2.2 (1.2) .00 -.11** -.08* .17***
Drug Use 1.8 (1.1) -.05 -.07 -.13*** .13***
School Misconduct 2.1 (1.0) -.08* .00 -.19*** .13***
General 1.8 (0.8) -.06 -.02 -.25*** .12***
Theft 1.5 (0.8) -.08* .01 -.23*** .05
Assault 1.6 (0.8) -.16*** .08* -.29*** .06
Total Deviance 1.8 (0.8) -.07* -.01 -.24*** .14***
Victim Deviance 1.5 (0.7) -.13*** .05 -.28*** .06
Victimless Deviance 2.0 (0.9) -.05 -.06 -.18*** .16***
Antiascetic Deviance 2.0 (1.1) -.02 -.09** -.11** .16***
Secular Deviance 1.7 (0.7) -.09** .03 -.29*** .11**
Note. * (p < .05), ** (p < .01), or *** (p < .001).
younger and older adolescents. Part C was tested by checking to confirm whether the
religiosity scales were significantly related to victimless deviance for both male and
female adolescents, Part D was tested by checking to confirm whether religious salience
was significantly related to victimless deviance for male adolescents but not female, and
Part E was tested by checking to confirm whether religious salience as well as all the
religiosity scales were significantly related to victim deviance for female adolescents but
not male. Parts F, G, and H involved similar tests as in Parts A and B, namely, checking
whether the relationship between the religiosity scales and antiascetic deviance was
125
stronger than the relationship between the religiosity scales and secular deviance (using
Fisher?s r to z transformation test), though the latter relationship should also be
statistically significant and subsequently completing the same test separately for male and
female adolescents (Part G) and younger and older adolescents (Part H) to examine
whether these relationships were similar regardless by sex and age group.
Hypothesis 4
This hypothesis was tested by using structural equation modeling. Part A was
tested by estimating 48 separate structural models with each model using one of eight
different single-item religiosity measures as the exogenous predictor variable, deviant
behavior as the endogenous dependent variable, and one of six family processes as an
intervening variable. Part B involved inserting ritualistic participation as the exogenous
variable in the same model while Part C placed relational practice into that role and Part
D placed religiosity into that role. Parts A, B,C, and D were tested using procedures
outlined in Holmbeck (1997). Part E used an SEM multi-group analysis approach to test
whether this pattern of mediation was moderated by sex while Part F conducted a similar 
test of moderation for age. Both Parts E and F were tested using procedures outlined in
Mounts (2002; see also Parker, Baltes, & Christiansen, 1997). Part G involved the model
parameters for both the direct and indirect pathways from religiosity to deviant behavior.
Parts H and I involved a similar examination of pathway parameters in mediating models
completed separately by sex and/or age group.
126
Table 9
Frequency Statistics for Demographic Variables
Variable Total Males Females Younger Older
Sex P2(1) = .84, p = .36
     Males 433(50.1) 211(48.5) 222(51.6)
     Females 432(49.9) 224(51.5) 208(48.4)
Race P2(5) = .44, p = .99 P2(5) = 6.37, p = .27
     Caucasian 626(72.4) 315(72.7) 311(72.0) 312(71.7) 314(73.0)
     African-American 181(20.9) 89(20.6) 92(21.3) 100(23.0) 81(18.8)
     Native-American 7(.8) 4(.9) 3(.7) 1(.2) 6(1.4)
     Asian-American 39(4.5) 20(4.6) 19(4.4) 17(3.9) 22(5.1)
     Hispanic 2(.2) 1(.2) 1(.2) 1(.2) 1(.2)
     Pacific Islander 5(.6) 2(.5) 3(.7) 2(.5) 3(.7)
     Missing 5(.6) 2(.5) 3(.7) 2(.5) 3(.7)
School Grade P2(3) = 1.58, p = .66
     9th Grade 269(31.1) 138(31.9) 131(30.3)
     10th Grade 210(24.3) 104(24.0) 106(24.5)
     11th Grade 219(25.3) 114(26.3) 105(24.3)
     12th Grade 167(19.3) 77(17.8) 90(20.8)
Family Income P2(4) = 4.32, p = .37 P2(4) = 2.13, p = .71
     $20,000 or less 61(7.1) 32(7.4) 29(6.7) 33(7.6) 28(6.5)
     $20,000-$35,000 98(11.3) 41(9.5) 57(13.2) 48(11.0) 50(11.6)
     $35,000-$60,000 234(27.1) 123(28.4) 111(25.7) 108(24.8) 126(29.3)
     $60,000-$100,000 261(30.2) 129(29.8) 132(30.6) 135(31.0) 126(29.3)
     More than $100,000 146(16.9) 79(18.2) 67(15.5) 71(16.3) 75(17.4)
     Missing 65(7.5) 29(6.7) 36(8.3) 40(9.2) 25(5.8)
Primary Wager Earner?s Occupation P2(5) = 6.36, p = .27 P2(5) = 8.94, p = .11
     Laborer 10(1.2) 7(1.6) 3(.7) 7(1.6) 3(.7)
     Semiskilled 42(4.9) 16(3.7) 26(6.0) 18(4.1) 24(5.6)
     Clerical 55(6.4) 23(5.3) 32(7.4) 34(7.8) 21(4.9)
     Semiprofessional 127(14.7) 59(13.6) 68(15.7) 61(14.0) 66(15.3)
     Professional 247(28.6) 122(28.2) 125(28.9) 111(25.5) 136(31.6)
     Executive 274(31.7) 143(33.0) 131(30.3) 144(33.1) 130(30.2)
     Missing 110(12.7) 63(14.5) 47(10.9) 60(13.8) 50(11.6)
Note. For each cell, the first number is the frequency and the second number is the percentage.
127
Table 10
Descriptive Statistics for Demographic Variables
Total Males Females Younger Older
m sd m sd m sd m sd m sd
Age 16.35 1.19 16.43 1.21 16.28 1.17 15.37 .61 17.36 .71
School Age 39.26 20.90 38.94 20.78 39.58 21.03 21.40 8.04 57.33 12.79
SES -.02 .91 .03 .91 .03 .90 -.02 .93 -.01 .88
Note. Means for male and female adolescents were compared using t-tests and findings
were as follows: Age, t = 1.85, p = .07; School Age, t = -.44, p = .66; SES, t = 1.46,
p = .15; means for younger and older adolescents were compared using t-tests and
findings were as follows: SES, t = -.18, p = .86.
Hypothesis 5
This hypothesis was tested by using structural equation modeling. Part A utilized
the same procedure as described above (Mounts, 2002; Parker et al., 1997) to test the
moderating role of religiosity (i.e., using the dichotomous religiosity variable described
earlier) between family processes and deviant behavior. Part B was tested using the same
procedure to compare male and female adolescents while Part C involved using the same
procedure for younger and older adolescents.
6 Religious conviction was found to be a poor indicator of religiosity as a deterrent
to deviance, since they were consistently positively rather than negatively correlated.
Consequently, it is not included in calculations of average r?s.
128
CHAPTER 4: RESULTS
Hypothesis 1: Bivariate Relationships
Part A
Hypothesis 1A proposed that the measures of religiosity would be statistically
related to the measures of deviance, even after controlling for race and SES. Results from
initial correlation analyses (see Table 11) indicated that all eight of the religiosity items
were significantly related to all seven of the deviance subscales, with only two exceptions
? religious conviction and vandalism were not related (r = .05, p = .19) nor were religious
conviction and assault (r = .05, p = .20). In addition, all eight of the religiosity items were
significantly related to the total deviance measure (average r = .216). Also, the three
religiosity scales were all significantly related to all seven of the deviance subscales
(ritualistic participation average r = -.18; relational practice average r = -.25; religiosity
average r = -.24) as well as the total deviance scale (ritualistic participation, r = -.22;
relational practice, r = -.31; religiosity, r = -.29).
Table 11
Second-Order Correlations of Religiosity Measures with Deviance Measures for Total Sample
Vandalism Alcohol Drugs School General Theft Assault Deviance
Single-Item Measures
   Church Attendance -.16 -.18 -.23 -.14 -.15 -.15 -.15 -.20
   Church Status -.19 -.22 -.27 -.18 -.20 -.20 -.21 -.25
   Religious Involvement -.18 -.23 -.25 -.17 -.18 -.21 -.21 -.24
   Bible Reading -.09* -.16 -.16 -.12** -.08* -.08* -.08* -.13
   Praying -.20 -.21 -.27 -.19 -.20 -.20 -.20 -.25
   Music Listening -.15 -.23 -.25 -.17 -.14 -.13 -.12** -.20
   Religious Conviction .05ns .10** -.08* .08* .08* .08* .05ns .09*
   Religious Salience -.14 -.21 -.23 -.15 -.16 -.18 -.14 -.21
Multi-Item Measures
   Ritualistic Participation -.16 -.22 -.25 -.17 -.15 -.17 -.16 -.22
   Relational Practice -.23 -.27 -.32 -.22 -.24 -.25 -.25 -.31
   Religiosity -.22 -.28 -.32 -.22 -.22 -.22 -.22 -.29
Note. Values in cells represent Pearson's r correlations between variables after controlling for race and SES; all correlations are
significant at p < .001 unless otherwise noted with ns (non-significant), * (p < .05), or ** (p < .01).
130
Part B
Because evidence from these correlation analyses indicated that the religiosity and
deviance measures were statistically related at the item level, further discussion will focus
on relationships at the scalar level since these scales are more reliable measures.
Hypothesis 1B proposed that the negative relationship between the religiosity scales and
deviance would be stronger for male adolescents than for female. Results indicated that
this was not the case (see Table 12). Though initial inspection appeared to indicate
differences by sex, results of Fisher?s r to z transformation tests indicated that only two of
these correlation pairs were significantly different. The relationship between religious
involvement and assault was stronger for male adolescents (r = -.24 versus r = -.10,
p = .026, one-tailed test) while the relationship between music listening and alcohol was
stronger for female adolescents (r = -.30 versus r = -.16, p = .018, one-tailed test). Thus,
results indicated that the relationship between religiosity and deviance did not differ by
sex.
Part C
Hypothesis 1C proposed that the negative relationship between the religiosity
scales and deviance would be stronger for older than younger adolescents. The results did
not support this prediction (see Table 13). Though initial inspection appeared to indicate a
pattern of differences between the age groups, only four were significantly different and,
for all four of these, younger adolescents indicated a stronger relationship than older
adolescents, contrary to the hypothesis (for religious salience and alcohol, r = -.29 versus
r = -.14, p = .03; for religious salience and drug use, r = -.30 versus r = -.16, p = .04; for
Table 12
Second-Order Correlations of Religiosity Measures with Deviance Measures by Sex
Vandalism Alcohol Drugs School General Theft Assault Deviance
M F M F M F M F M F M F M F M F
Single-Item Measures
   Church Attendance -.13** -.17** -.17** -.19 -.22 -.23 -.12* -.13** -.13* -.15 -.15** -.14** -.16** -.11* -.18 -.21
   Church Status -.14** -.20 -.19 -.25 -.24 -.27 -.17** -.16 -.16** -.21 -.17** -.20 -.18 -.18 -.21 -.26
   Religious Involvement -.15** -.16** -.23 -.20 -.26 -.21 -.15** -.14 -.16** -.15 -.23 -.13* -.24 -.10* -.24 -.20
   Bible Reading -.06ns -.11* -.10* -.22 -.11* -.21 -.07ns -.16** -.07ns -.06ns -.05ns -.09ns -.04ns -.11* -.08ns -.18
   Praying -.17** -.21 -.19 -.21 -.24 -.29 -.18 -.17** -.21 -.15** -.18 -.20 -.16** -.23 -.22 -.26
   Music Listening -.12* -.14** -.16** -.30 -.20 -.28 -.15** -.17** -.11* -.13* -.14** -.07ns -.13* -.05ns -.16** -.23
   Religious Conviction .02ns .06ns .07ns .13** .05ns .10ns .05ns .11* .06ns .07ns .07ns .08ns -.00ns .08ns .05ns .11*
   Religious Salience -.10ns -.19 -.17** -.25 -.18 -.28 -.11* -.19 -.13* -.18** -.17** -.17** -.12* -.13* -.16** -.26
Multi-Item Measures
   Ritualistic Participation -.12* -.19 -.17** -.27 -.21 -.29 -.12* -.20 -.13* -.16** -.15** -.16** -.14** -.14** -.17 -.26
   Relational Practice -.19 -.23 -.25 -.27 -.31 -.32 -.20 -.19 -.22 -.21 -.24 -.22 -.24 -.21 -.28 -.30
   Religiosity -.17** -.23 -.23 -.31 -.29 -.34 -.19 -.22 -.19 -.20 -.21 -.19 -.20 -.18 -.25 -.31
Note. M = Males, F = Females; values in cells represent Pearson's r correlations between variables after controlling for race and SES; all correlations are
significant at p < .001 unless otherwise noted with ns (non-signist ficant), * (p < .05), or ** (p < .01); correlation pairs that differed statistically are shaded.
Table 13
Second-Order Correlations of Religiosity Measures with Deviance Measures by Age
Vandalism Alcohol Drugs School General Theft Assault Deviance
Y O Y O Y O Y O Y O Y O Y O Y O
Single-Item Measures
   Church Attendance -.21 -.10ns -.23 -.14** -.28 -.19 -.13** -.14** -.18 -.11* -.18 -.13* -.19 -.12* -.24 -.16**
   Church Status -.20 -.18 -.24 -.21 -.26 -.28 -.12* -.24 -.16** -.24 -.16 -.24 -.19 -.22 -.23 -.28
   Religious Involvement -.18 -.17** -.27 -.19 -.30 -.21 -.16** -.17** -.18 -.17** -.22 -.20 -.20 -.22 -.27 -.22
   Bible Reading -.10ns -.09ns -.23 -.11* -.20 -.13** -.14** -.11* -.10* -.06ns -.13* -.03ns -.10ns -.07ns -.15** -.11*
   Praying -.19 -.21 -.25 -.17** -.28 -.25 -.17** -.21 -.15** -.25 -.21 -.19 -.20 -.21 -.24 -.26
   Music Listening -.16** -.14** -.24 -.22 -.28 -.21 -.16** -.18 -.15** -.13* -.16** -.10ns -.14** -.10* -.21 -.19
   Religious Conviction .04ns .05ns .11* .12* .08ns .09ns .11* .06ns .08ns .08ns .08ns .07ns .04ns .05ns .08ns .10ns
   Religious Salience -.20 -.09ns -.29 -.14** -.30 -.16** -.21 -.10ns -.20 -.12* -.21 -.15** -.18 -.11* -.27 -.15**
Multi-Item Measures
   Ritualistic Participation -.22 -.10* -.31 -.15** -.33 -.19 -.20 -.14** -.20 -.11* -.22 -.12* -.20 -.12* -.28 -.16**
   Relational Practice -.24 -.23 -.31 -.23 -.35 -.30 -.19 -.25 -.21 -.27 -.25 -.25 -.24 -.26 -.30 -.31
   Religiosity -.25 -.18 -.34 -.22 -.38 -.27 -.22 -.22 -.23 -.20 -.25 -.19 -.24 -.20 -.32 -.26
Note. Y = Younger, O = Older; values in cells represent Pearson's r correlations between variables after controlling for race and SES; all correlations are
significant at p < .001 unless otherwise noted with ns (non-significant), * (p < .05), or ** (p < .01); correlation pairs that differed statistically are shaded.
133
ritualistic participation and alcohol, r = -.31 versus r = -.15, p = .02; for ritualistic
participation and drug use, r = -.33 versus r = -.19, p = .04). Therefore, the relationship
between religiosity and deviance appeared to be largely invariant by age group, taking
into account the noted exceptions.
Hypothesis 2: Measurement
The second hypothesis in this study proposed that multi-item scales of religiosity
would be stronger predictors of deviance than single-item measures. Results provided
some support for this hypothesis (see Table 14). The value of the Pearson?s r correlation
coefficients representing the relationship between the single-item measures of religiosity
and the total deviance measure ranged from r = .08 to r = -.25 (all coefficients statistically
significant), while these same coefficient values ranged from r = -.22 to r = -.30 for the
multi-item scales (all coefficients statistically significant). Though some of the single-
item measures (church status, r = -.25; religious involvement, r = -.24; praying, r = -.25)
appeared stronger than the ritualistic participation scale (r = -.22), none of these
differences reached statistical significance (for church status, z = -.67, p = .25; for
religious involvement, z = -.50, p = .31; for praying, z = -.61, p = .27). The ritualistic
participation scale had a significantly larger relationship with deviance than one of the
single-item measures (religious conviction). The relational practice scale had a
significantly larger relationship with deviance than 5 of the 8 single-item measures
(church attendance, Bible reading, music listening, religious conviction, religious
salience), while the religiosity scale had a significantly larger relationship with deviance
134
Table 14
Comparison of Second-Order Correlations of Religiosity Measures (Single- and Multi-
Item) with Total Deviance
Single-Item Measures r p
Multi-Item Measures
Ritualistic
(r = -.22, p = .00)
Relational
(r = -.30, p = .00)
Religiosity
(r = -.28, p = .00)
z p z p z p
   Church Attendance -.20 .000 .41 .34 2.22 .01 1.85 .03
   Church Status -.25 .000 -.67 .25 1.13 .13 .77 .22
   Religious Involvement -.24 .000 -.50 .31 1.30 .10 .93 .18
   Bible Reading -.13 .000 1.80 .04 3.60 .00 3.23 .00
   Praying -.25 .000 -.61 .27 1.20 .12 .83 .20
   Music Listening -.20 .000 .35 .36 2.15 .02 1.79 .04
   Religious Conviction .08 .018 5.96 .00 7.76 .00 7.40 .00
   Religious Salience -.21 .000 .21 .42 2.00 .02 1.64 .05
Note. A one-tailed test was used to test whether the relationship between multi-item
religiosity measures and deviance was significantly larger than the relationship between
single-item religiosity measures and deviance.
than 4 of these single-item measures (church attendance, Bible reading, music listening,
religious conviction).
Hypothesis 3: Type of Deviance
Part A
Hypothesis 3A proposed that religiosity would be more strongly related to
victimless deviance, though still related to victim deviance. Results did not support this
hypothesis (see Table 15). Though all three of the religiosity scales appeared to be more
strongly related to victimless deviance than to victim deviance (for victim deviance,
135
Table 15
Second-Order Correlations of Religiosity Measures with Different Types of Deviance for
Total Sample
Victim
Deviance
Victimless
Deviance
Antiascetic
Deviance
Secular
Deviance
Single-Item Measure
   Religious Salience -.17 -.22 -.23 -.18
Multi-Item Measures
   Ritualistic Participation -.17 -.23 -.25 -.18
   Relational Practice -.27 -.30 -.31 -.27
   Religiosity -.24 -.30 -.31 -.25
Note. Values in cells represent Pearson's r correlations between variables after controlling
for race and SES; all correlations are significant at p < .001
average r = -.23; for victimless deviance, average r = -.28), none of these differences
were statistically significant. All religiosity measures were significantly related to both
victim and victimless deviance.
Part B
Hypothesis 3B proposed that this same relationship (i.e., that religiosity would be
more strongly related to victimless deviance while still being related to victim deviance)
would be found for both male and female adolescents. Results suggested that this same
pattern of association was found for both male and female adolescents (see Table 16).
Though all three of the religiosity scales appeared to be more strongly associated with
victimless deviance for both male (for victim deviance, average r = -.21; for victimless
deviance, average r = -.24) and female (for victim deviance, average r = -.21; for
136
Table 16
Second-Order Correlations of Religiosity Measures with Different Types of Deviance by
Sex
Victim
Deviance
Victimless
Deviance
Antiascetic
Deviance
Secular
Deviance
M F M F M F M F
Single-Item Measures
   Religious Salience -.16** -.16** -.17** -.26 -.18 -.28 -.14** -.20
Multi-Item Measures
   Ritualistic Participation -.15** -.17** -.18 -.27 -.20 -.30 -.14** -.21
   Relational Practice -.25 -.24 -.28 -.29 -.29 -.31 -.25 -.25
   Religiosity -.22 -.21 -.25 -.32 -.27 -.34 -.22 -.24
Note. Values in cells represent Pearson's r correlations between variables after controlling
for race and SES; all correlations are significant at p < .001 unless otherwise noted with
ns (non-significant), * (p < .05), or ** (p < .01).
victimless deviance, average r = -.29) adolescents, none of these differences reached
statistical significance based on Fisher?s r to z transformations. Again, all religiosity
scales were statistically related to both victim and victimless deviance for both groups.
Part C
Hypothesis 3C stated that this same relationship (i.e., that religiosity would be
more strongly related to victimless deviance while still being related to victim deviance)
would be found for both younger and older adolescents. Results suggested that this same
pattern of association was found for both younger and older adolescents (see Table 17).
Though all three of the religiosity scales appeared to be more strongly related to
victimless deviance for both younger (for victim deviance, average r = -.26; for victimless
137
Table 17
Second-Order Correlations of Religiosity Measures with Different Types of Deviance by
Age
Victim
Deviance
Victimless
Deviance
Antiascetic
Deviance
Secular
Deviance
Y O Y O Y O Y O
Single-Item Measures
   Religious Salience -.21 -.13** -.28 -.15** -.31 -.16** -.22 -.13*
Multi-Item Measures
   Ritualistic Participation -.23 -.12* -.30 -.17** -.33 -.18 -.23 -.14**
   Relational Practice -.27 -.28 -.30 -.30 -.35 -.28 -.25 -.29
   Religiosity -.27 -.21 -.34 -.26 -.38 -.26 -.26 -.23
Note. Values in cells represent Pearson's r correlations between variables after controlling
for race and SES; all correlations are significant at p < .001 unless otherwise noted with
ns (non-significant), * (p < .05), or ** (p < .01).
deviance, average r = -.31) and older (for victim deviance, average r = -.20; for victimless
deviance, average r = -.24) adolescents, Fisher?s r to z tests indicated that none of these
differences reached statistical significance. Again, all religiosity scales were statistically
related to both victim and victimless deviance for both groups.
Part D
Hypothesis 3D stated that the measure of religious salience would only be related
to victimless deviance for male adolescents. Results did not support this hypothesis (see
Table 16). In fact, religious salience was related to victimless deviance for both male
(r = -.17, p = .00) and female (r = -.26, p = .00) adolescents, and this association did not
differ by sex based on Fisher?s r to z transformations.
138
Part E
Hypothesis 3E proposed that both the religiosity scales and religious salience
would be related to victim deviance only for female adolescents. Results did not support
this hypothesis (see Table 16). Rather, it was found that all three of the religiosity scales
were significantly related to victim deviance for both male and female adolescents
(average r = -.21 for both groups), and religious salience was also significantly related to
victim deviance in an equal manner for both groups (r = -.16 for both).
Part F
Hypothesis 3F stated that religiosity would be more strongly related to antiascetic
deviance, though still not unrelated to secular deviance. Results did not support this
hypothesis (see Table 15). Though all three of the religiosity scales appeared to be more
strongly related to antiascetic deviance (average r = -.29) than to secular deviance
(average r = -.23), none of these differences were statistically significant. All religiosity
measures were significantly related to both antiascetic and secular deviance.
Part G
Hypothesis 3G proposed that this same relationship (i.e., that the religiosity
measures would be more strongly related to antiascetic deviance than to secular deviance)
would be found for both male and female adolescents. Results suggested that the same
pattern of association found in the total sample was apparent for both male and female
adolescents (see Table 16). Though all three of the religiosity scales appeared to be more
strongly related to antiascetic deviance for both male (for antiascetic deviance, average
r = -.25; for secular deviance, average r = -.20) and female (for antiascetic deviance,
139
average r = -.32; for secular deviance, average r = -.23) adolescents, none of these
differences reached statistical significance. Again, all religiosity scales were statistically
related to both antiascetic and secular deviance for both groups.
Part H
Hypothesis 3H stated that this same relationship (i.e., that the religiosity measures
would be more strongly related to antiascetic deviance than to secular deviance) would
also be found for both younger and older adolescents. Results suggested that the same
pattern of association found in the total sample was apparent for both younger and older
adolescents (see Table 17). Thus, though most of the religiosity scales appeared to be
more strongly related to antiascetic than to secular deviance, none of these differences
reached statistical significance based on r to z transformations. Again, all religiosity
scales were statistically related to both antiascetic and secular deviance for both groups.
Hypothesis 4: Mediation
The most commonly utilized method of testing for mediation is Baron and
Kenny?s (1986) traditional regression approach. However, Baron and Kenny caution that
the regression approach to testing mediation requires the assumption ?that there be no
measurement error in the mediator? (p. 1177). They further concede, however, that ?the
mediator . . . is likely to be measured with error? (p. 1177). Consequently, they
recommend the use of the multiple indicator approach or latent-variable structure
modeling to overcome these inherent weaknesses in the regression analytic approach.
Several authors concur with both this caution and recommendation (Hoyle & Smith,
1994; Judd & Kenny, 1981; Kenny, Kashy, & Bolger, 1998). Thus, structural equation 
140
Figure 10. Direct model, testing effect of religiosity on deviance with no intervening
variable.
modeling offers a superior alternative for tests of mediation because of its ability to avoid
problems of over- or underestimation (i.e., Type I and Type II error) by controlling for
measurement error (Hoyle & Smith, 1994). Consequently, the current investigation
utilized the SEM approach to testing mediation. When testing models using SEM, there is
a wide variety of statistics that can be used to assess the adequacy of model fit. The most
common methods (and, consequently, the ones utilized in the current investigation)
Figure 11. Overall model testing three key relationships (independent-mediator,
mediator-dependent, independent-dependent) simultaneously.
141
7 Though Holmbeck (1997) specifies that a SEM test of mediation includes a
latent predictor variable, the current study utilized a manifest predictor variable. This is
not inconsistent with structural modeling (Hoyle & Smith, 1994), but represents a slight
variation from Holmbeck?s procedure.
include a Comparative Fit Index (CFI) > .90, a Root Mean Square Error of
Approximation (RMSEA) < .05, and a non-significant chi-square.
Holmbeck (1997) outlines a very specific plan of analysis involving the use of
SEM for testing mediated effects (see also Hoyle and Smith, 1994, who present a slightly
different approach to SEM-based mediation testing).7 First, the direct model
(i.e., unmediated; the effect of the independent variable on the dependent variable without
any mediating variable) should be tested (see Figure 10). If this model provides an
adequate fit to the data, the analyses may proceed. Second, the overall model should be
tested with all parameters free to be estimated (see Figure 11). At this point, not only
must this model provide an adequate fit to the data, but also, in order to meet the
statistical criteria for a test of mediation, all of the primary pathways (i.e., A, B, and C on
Figure 11) must be significant. Third, when these criteria are met, the overall model
should be tested with the independent-dependent pathway (i.e., C on Figure 11)
constrained to zero. Finally, the difference between the chi-square statistics from the
constrained and unconstrained models should be statistically tested. If the chi-square
difference statistic is significant, this is an indication that there is no mediation. If, on the
other hand, the chi-square statistic is nonsignificant, this is an indication that there is a
mediation effect. In other words, if the effect of the independent variable on the
dependent variable is being mediated, the addition of the C parameter to the model should
142
Table 18
Parameter Estimates and Fit Indices for Direct Models of Religiosity and Deviant
Behavior
$ p P2 CFI RMSEA
Single-Item Measures
   Church Attendance -.18 .000 39.3 .995 .046
   Church Status -.24 .000 40.3 .994 .047
   Religious Involvement -.22 .000 45.3 .993 .051
   Bible Reading -.11 .004 35.0 .995 .042
   Praying -.24 .000 39.2 .995 .046
   Music Listening -.17 .000 48.7 .993 .054
   Religious Conviction .08 .025 21.2 .998 .024
   Religious Salience -.20 .000 35.1 .995 .042
Multi-Item Measures
   Ritualistic Participation -.20 .000 45.9 .993 .051
   Relational Practice -.27 .000 46.6 .992 .057
   Religiosity -.26 .000 56.5 .991 .059
Note. For all models, df = 14.
not significantly improve its fit. However, if the addition of the C parameter does
significantly improve the fit of the model, this indicates that this effect (i.e., independent
on dependent) is not being mediated.
Part A
Hypothesis 4A suggested that family processes would mediate the relationships
between all of the single-item religiosity measures and deviant behavior. Results did not
Table 19
Parameter Estimates for Overall Models Involving Tests of Mediating Role for Family Processes Between Individual Religiosity Items and Deviant Behavior
Parental
Closeness
Parental
Support
Parental
Monitoring
Parental
Communication
Parental
Conflict
Parental
Peer Approval
R F R D F D R F R D F D R F R D F D R F R D F D R F R D F D R F R D F D
Church Attendance .29(.00) -.04(.30) -.50(.00) .16(.00) -.12(.00) -.39(.00) .26(.00) -.03(.42) -.57(.00) .24(.00) -.10(.01) -.31(.00) -.17(.00) -.11(.00) .41(.00) .25(.00) -.08(.02) -.38(.00)
Church Status .31(.00) -.09(.02) -.48(.00) .23(.00) -.15(.00) -.37(.00) .26(.00) -.09(.02) -.55(.00) .21(.00) -.17(.00) -.29(.00) -.21(.00) -.15(.00) .40(.00) .24(.00) -.15(.00) -.36(.00)
Religious Involvement .26(.00) -.10(.01) -.48(.00) .19(.00) -.15(.00) -.38(.00) .28(.00) -.07(.10) -.56(.00) .26(.00) -.14(.00) -.30(.00) -.18(.00) -.15(.00) .41(.00) .23(.00) -.14(.00) -.37(.00)
Bible Reading .22(.00) -.01(.07) -.51(.00) .16(.00) -.04(.25) -.40(.00) .22(.00) -.03(.51) -.58(.00) .25(.00) -.02(.62) -.33(.00) -.14(.01) -.05(.23) .43(.00) .18(.00) -.03(.39) -.36(.00)
Praying .34(.00) -.07(.07) -.49(.00) .23(.00) -.15(.00) -.37(.00) .24(.00) -.11(.01) -.56(.00) .30(.00) -.14(.00) -.29(.00) -.12(.02) -.19(.00) .41(.00) .29(.00) -.13(.00) -.36(.00)
Music Listening .22(.00) -.06(.10) -.50(.00) .07(.11) -.15(.00) -.40(.00) .28(.00) -.01(.85) -.58(.00) .23(.00) -.10(.01) -.32(.00) -.07(.18) -.14(.00) -.42(.00) .21(.00) -.09(.02) -.38(.00)
Religious Conviction -.03(.47) .07(.06) -.50(.00) -.06(.16) .06(.11) -.40(.00) -.10(.04) .03(.49) -.58(.00) -.07(.08) .06(.09) -.32(.00) .02(.66) .08(.04) .43(.00) -.06(.14) .06(.10) -.40(.00)
Religious Salience .26(.00) -.07(.08) -.49(.00) .15(.00) -.14(.00) -.39(.00) .25(.00) -.05(.21) -.57(.00) .28(.00) -.10(.01) -.31(.00) -.14(.01) -.14(.00) .41(.00) .20(.00) -.12(.00) -.38(.00)
Note. For all models, df = 27. For each cell, top number is $ (standardized coefficient) and bottom number is p value. R F = pathway from religiosity item to
family processes scale; R D = pathway from religiosity item to deviance scale; F D = pathway from family processes scale to deviance scale.
Table 20
Model Statistics for Overall Models Involving Tests of Mediating Role for Family Processes Between Individual Religiosity Items and Deviant Behavior
Parental
Closeness
Parental
Support
Parental
Monitoring
Parental
Communication
Parental
Conflict
Parental
Peer Approval
P2 CFI RMSEA P2 CFI RMSEA P2 CFI RMSEA P2 CFI RMSEA P2 CFI RMSEA P2 CFI RMSEA
Church Attendance 76.3 .990 .046 57.8 .994 .036 65.6 .992 .041 66.8 .992 .041 58.4 .993 .037 85.4 .988 .050
Church Status 78.1 .990 .047 57.8 .994 .036 68.4 .992 .042 64.5 .992 .040 60.0 .993 .038 87.6 .988 .051
Religious Involvement 81.5 .989 .048 62.7 .993 .039 71.2 .991 .044 70.6 .991 .043 67.1 .992 .041 93.9 .987 .054
Bible Reading 73.7 .991 .045 53.9 .995 .034 61.8 .993 .039 57.9 .994 .036 53.8 .994 .034 83.7 .989 .049
Praying 77.8 .990 .047 56.4 .994 .036 66.6 .992 .041 65.2 .992 .040 58.1 .994 .037 85.7 .988 .050
Music Listening 88.0 .988 .051 68.0 .992 .042 73.9 .991 .045 74.4 .991 .045 67.6 .992 .042 96.7 .986 .055
Religious Conviction 59.2 .993 .037 40.2 .997 .024 51.1 .995 .032 44.1 .997 .027 41.8 .997 .025 74.1 .991 .045
Religious Salience 73.8 .991 .045 54.0 .995 .034 62.6 .993 .039 60.9 .993 .038 54.2 .994 .034 84.6 .989 .050
Note. For all models, df = 27. All chi-square statistics are significant at p < .001. CFI = Comparative Fit Index; RMSEA = Root Mean Square Error of
Approximation.
Table 21
Fit Statistics in Comparisons Between Overall Models with Constrained and Unconstrained Pathway Between Single-Item Religiosity Measures and Deviance
Parental
Closeness
Parental
Support
Parental
Monitoring
Parental
Communication
Parental
Conflict
Parental
Peer Approval
C/M U/N )P2 C/M U/N )P2 C/M U/N )P2 C/M U/N )P2 C/M U/N )P2 C/M U/N )P2
Church Attendance $ -.119 -.183 -.103 -.183 -.113 -.183 -.084 -.183P2 68.0 57.8 10.2** 73.9 66.8 7.1** 66.6 58.4 8.2** 90.3 85.4 4.9*
Church Status $ -.091 -.238 -.153 -.238 -.094 -.238 -.172 -.238 -.154 -.238 -.148 -.238P2 83.2 78.1 5.1* 74.0 57.8 16.2*** 73.9 68.4 5.5* 85.0 64.5 20.5*** 74.3 60.0 14.3*** 102.8 87.6 15.2***
Religious Involvement $ -.096 -.223 -.150 -.223 -.142 -.223 -.152 -.223 -.136 -.223P2 87.5 81.5 6.0* 78.8 62.7 16.1*** 84.2 70.6 13.6*** 81.5 67.1 14.4*** 106.9 93.9 13.0***
Bible Reading $P2
Praying $ -.153 -.239 -.105 -.239 -.143 -.239 -.189 -.239 -.128 -.239P2 72.7 56.4 16.3*** 73.5 66.6 6.9** 78.2 65.2 13.0*** 81.6 58.1 23.5*** 96.5 85.7 10.8**
Music Listening $ -.096 -.174 -.088 -.174P2 80.6 74.4 6.2* 102.1 96.7 5.4*
Religious Conviction $P2
Religious Salience $ -.139 -.197 -.102 -.197 -.139 -.197 -.121 -.197P2 67.3 54.0 13.3*** 67.4 60.9 6.5* 66.4 54.2 12.2*** 94.7 84.6 10.1**
Note. Holmbeck (1997) indicates that standardized parameter estimates ($) for the models with and without the mediating variable should be reported; they are
presented here in conjunction with the chi-square statistics from the constrained and unconstrained models; C/M = constrained/mediator present;
U/N = unconstrained/no mediator present; for the constrained model, df = 28, while for the unconstrained model, df = 27; the chi-square critical values at 1 df are
*3.84 (p = .05), **6.64 (p = .01), ***10.83 (p = .001); empty cells represent relationships that did not meet the statistical criteria for a test of mediation.
146
Figure 12. Actual structural model tested.
support this hypothesis (see Tables 18, 19, 20, and 21). Following Holmbeck?s (1997)
procedure, the first step was to test the direct models. Results indicated that models for all
8 of the single-item religiosity measures provided a good fit to the data (Table 18; all
pathways significant, all CFI?s > .99, RMSEA?s ranging from .024 to .054). Thus, the
first of Holmbeck?s criteria was met.
The next step was to test the overall model fits. Results indicated that all of these
mediation models fit the data very well (see Table 20). Though the chi-square statistics in
both the direct and overall model analyses were all significant at the p < .001 level
(theoretically indicating poor model fit), this was to be expected since chi-square is
extremely sample sensitive and always inflated in larger samples (Bentler & Bonett,
1980; Loehlin, 1992; Marsh, Balla, & McDonald, 1988; Marsh & Hocevar, 1985). The
147
fact that the lowest CFI was .986 and all but 7 of the RMSEA?s were less than .05
(highest, .054) was good evidence that these models explained the data in an acceptable
manner. Thus, the first part of the second of Holmbeck?s criteria was met.
The second part of the second of Holmbeck?s criteria required that all model
pathways be statistically significant. Since there were six family processes measures
tested with eight religiosity items, there were a total of 48 sets of pathways (3 pathways in
each set: A, B, and C on Figure 11) that needed to be statistically significant. Of these 48,
22 did not meet the criteria for mediation; that is, at least one of these pathways was
nonsignificant (see Table 19). Therefore, the remaining 26 models met all the criteria for
mediation and could be tested for such.
The next step involved completing analyses that constrained the predictor-
criterion (i.e., C on Figure 11) pathway to zero. Of these 26 models of potential
mediation, the chi-square difference statistic in all 26 was statistically significant and thus
indicated no mediation effect (see Table 21). Thus, in all 26 of the potentially mediated
models, not one was mediated by family processes. More specifically, in all 26 of these
models, the religiosity items maintained a significant relationship with deviance that was
not attenuated by any of the family processes variables.
Part B
Hypothesis 4B predicted that the relationship between the ritualistic participation
scale and deviance would be mediated by family processes. Results did not support this
hypothesis (see Tables 18, 22, 23, and 24). The direct model involving the effect of
ritualistic participation on deviance in the absence of any intervening variable (see Table
148
18) showed good fit (P2 = 45.9; CFI = .993; RMSEA = .051) and the parameter of
importance was statistically significant ($ = -.20, p = .000). Tests of the overall models
including the family processes variables (see Table 23) also showed good fit to the data
(P2 range = 63.5-93.2; CFI range = .987-.993; RMSEA range = .040-.053). Of the 6 sets
of pathways, two (parental closeness and parental monitoring) did not meet the criteria for
mediation (see Table 22). Of the remaining models of potential mediation, the chi-square
difference statistic in all 4 was statistically significant and thus indicated no mediation
effect (see Table 24).
Part C
Hypothesis 4C proposed that the relationship between the relational practice scale
and deviance would not be mediated by family processes. This hypothesis was supported
(see Tables 18, 22, 23, and 24). The direct model involving the effect of relational
practice on deviance in the absence of any intervening variable (see Table 18) showed
good fit (P2 = 46.6; CFI = .992; RMSEA = .057) and the parameter of importance was
statistically significant ($ = -.27, p = .000). Tests of the overall models including the
family processes variables (see Table 23) also showed good fit to the data
(P2 range = 70.5-100.8; CFI range = .986-.991; RMSEA range = .043-.056). All 6 sets of
pathways met the criteria for mediation (see Table 22), and the chi-square difference
statistic in all 6 was statistically significant (see Table 24) and thus indicated no
mediation effect.
Table 22
Parameter Estimates for Overall Models Involving Tests of Mediating Role for Family Processes Between Religiosity Scales and Deviant Behavior for Total
Sample
Parental
Closeness
Parental
Support
Parental
Monitoring
Parental
Communication
Parental
Conflict
Parental
Peer Approval
R F R D F D R F R D F D R F R D F D R F R D F D R F R D F D R F R D F D
Ritualistic Participation .33(.00) -.03(.45) -.50(.00) .20(.00) -.12(.00) -.38(.00) .31(.00) -.02(.70) -.57(.00) .33(.00) -.08(.02) -.31(.00) -.20(.00) -.11(.00) .41(.00) .27(.00) -.09(.01) -.38(.00)
Relational Practice .36(.00) -.12(.00) -.47(.00) .26(.00) -.19(.00) -.36(.00) .31(.00) -.12(.00) -.54(.00) .32(.00) -.19(.00) -.28(.00) -.20(.00) -.21(.00) .39(.00) .31(.00) -.18(.00) -.34(.00)
Religiosity .37(.00) -.08(.04) -.48(.00) .23(.00) -.18(.00) -.37(.00) .34(.00) -.07(.09) -.55(.00) .35(.00) -.16(.00) -.29(.00) -.20(.00) -.18(.00) .40(.00) .31(.00) -.15(.00) -.35(.00)
Note. For all models, df = 27. For each cell, top number is $ (standardized coefficient) and bottom number is p value. R F = pathway from religiosity item to
family processes scale; R D = pathway from religiosity item to deviance scale; F D = pathway from family processes scale to deviance scale.
Table 23
Model Statistics for Overall Models Involving Tests of Mediating Role for Family Processes Between Religiosity Scales and Deviant Behavior for Total Sample
Parental
Closeness
Parental
Support
Parental
Monitoring
Parental
Communication
Parental
Conflict
Parental
Peer Approval
P2 CFI RMSEA P2 CFI RMSEA P2 CFI RMSEA P2 CFI RMSEA P2 CFI RMSEA P2 CFI RMSEA
Ritualistic Participation 83.5 .989 .049 63.5 .993 .040 70.9 .991 .043 72.4 .991 .044 64.7 .992 .040 93.2 .987 .053
Relational Practice 90.7 .987 .052 70.5 .991 .043 79.7 .989 .048 80.8 .989 .048 73.8 .990 .045 100.8 .986 .056
Religiosity 94.3 .987 .054 73.7 .991 .045 80.6 .989 .048 83.7 .989 .049 75.5 .990 .046 103.1 .985 .057
Note. For all models, df = 27. All chi-square statistics are significant at p < .001. CFI = Comparative Fit Index; RMSEA = Root Mean Square Error of
Approximation.
Table 24
Fit Statistics in Comparisons Between Overall Models with Constrained and Unconstrained Pathway Between Religiosity Scales and Deviance for Total Sample
Parental
Closeness
Parental
Support
Parental
Monitoring
Parental
Communication
Parental
Conflict
Parental
Peer Approval
C/M U/N )P2 C/M U/N )P2 C/M U/N )P2 C/M U/N )P2 C/M U/N )P2 C/M U/N )P2
Ritualistic Participation
$ -.119 -.195 -.088 -.195 -.114 -.195 -.093 -.195
P2 73.5 63.5 10.0** 77.2 72.4 4.8* 72.7 64.7 8.0** 99.1 93.2 5.9*
Relational Practice
$ -.084 -.268 -.192 -.268 -.120 -.268 -.193 -.268 -.208 -.268 -.177 -.268
P2 98.2 90.7 7.5** 95.6 70.5 25.1*** 87.7 79.7 8.0** 104.5 80.8 23.7*** 99.6 73.8 25.8*** 121.6 100.8 20.8***
Religiosity
$ -.084 -.263 -.178 -.263 -.158 -.263 -.183 -.263 -.150 -.263
P2 98.2 94.3 3.9* 95.9 73.7 22.2*** 99.1 83.7 15.4*** 95.8 75.5 20.3*** 118.1 103.1 15.0***
Note. Holmbeck (1997) indicates that standardized parameter estimates ($) for the models with and without the mediating variable should be reported; they are
presented here in conjunction with the chi-square statistics from the constrained and unconstrained models; C/M = constrained/mediator present;
U/N = unconstrained/no mediator present; for the constrained model, df = 28, while for the unconstrained model, df = 27; the chi-square critical values at 1 df are
as follows: *3.84 (p = .05), **6.64 (p = .01), ***10.83 (p = .001); empty cells represent relationships that did not meet the statistical criteria for a test of
mediation.
152
Part D
Hypothesis 4D stated that the relationship between the religiosity scale and
deviance would not be mediated by family processes. This hypothesis was supported (see
Tables 18, 22, 23, and 24). The direct model involving the effect of religiosity on
deviance in the absence of any intervening variable (see Table 18) showed good fit
(P2 = 56.5; CFI = .991; RMSEA = .059) and the parameter of importance was statistically
significant ($ = -.26, p = .000). Tests of the overall models including the family processes
variables (see Table 23) also showed good fit to the data (P2 range = 73.7-103.1;
CFI range = .985-.991; RMSEA range = .045-.057). Of the 6 sets of pathways, one
(parental monitoring) did not meet the criteria for mediation (see Table 22). Of the
remaining models of potential mediation, the chi-square difference statistic in all 5 was
statistically significant (see Table 24) and thus indicated no mediation effect.
Part E
Hypothesis 4E suggested that this same pattern of mediation would be found for
both male and female adolescents. In other words, this hypothesis tested the moderating
role of sex for the structural model in which family processes mediated the effects of
religiosity on deviance. This role of moderation was tested using a multiple-group SEM
approach. Results supported the hypothesis that this pattern of mediation would not be
moderated by sex (see Tables 25, 26, 27, and 28).
Several steps are involved in this analytical approach (Mounts, 2002; Parker et al.,
1997). First, a fully variant model was estimated in order to generate path coefficients for
each of the groups (e.g., male and female adolescents). Second, an omnibus test was
153
employed where all relationships were tested simultaneously for both groups (e.g., male
and female adolescents) in order to assess whether the model was equivalent across the
groups. In this omnibus test, all paths were held invariant across the groups. If the chi-
square difference between the fully variant model and the fully invariant model was
statistically significant, this was an indication of moderation, that is, that there were
differences between the groups in the way the data fit the model. If such a significant
difference was found, thus indicating a difference between the groups, then a third
analytical step was necessary. In this third step, a series of univariate tests were conducted
where each path was held invariant while the others were free to be estimated. This
procedure was used to determine exactly which paths were moderated. Because the focus
of these analyses was solely on the mediation of religiosity by family processes, only
these pathways were examined for potential moderation effects in univariate follow-up
analyses.
Results of the omnibus tests comparing the fully variant and fully invariant
models of mediation by sex indicated that, even when all pathways in the model were
constrained to equality for male and female adolescents, this model fit the data well, since
all CFI?s exceeded .99 and RMSEA?s ranged from .018 to .026 (see Table 25). A pattern
was found in these results where, for each of the three religiosity scales tested, the models
involving the same four family processes scales were always found to be moderated by
sex ? closeness, monitoring, communication, and peer approval. Conversely, the models
involving support and conflict never differed by sex. Because significant chi-square
differences were found in four of the six models for all three of the religiosity scales, thus
154
Table 25
Fit Statistics in Comparisons Between Fully Variant and Fully Invariant Models by Sex
Fully Variant Model Fully Invariant Model Change Statistics
Family Processes P2 df CFI RMSEA P2 df CFI RMSEA )P2 )df )CFI )RMSEA
Ritualistic
Participation
Closeness 308.12 135 .99 .022 320.99 138 .99 .023 12.87** 3 0.00 0.001
Support 248.65 135 .99 .018 249.98 138 .99 .018 1.33 3 0.00 0.000
Monitoring 273.91 135 .99 .020 282.13 138 .99 .020 8.22* 3 0.00 0.000
Communication 263.51 135 .99 .019 271.99 138 .99 .019 8.48* 3 0.00 0.000
Conflict 270.33 135 .99 .020 276.18 138 .99 .020 5.85 3 0.00 0.000
Peer Approval 319.65 135 .99 .023 337.86 138 .99 .024 18.21*** 3 0.00 0.001
Relational
Practice
Closeness 331.62 135 .99 .024 346.54 138 .99 .024 14.92** 3 0.00 0.000
Support 274.73 135 .99 .020 277.52 138 .99 .020 2.79 3 0.00 0.000
Monitoring 305.21 135 .99 .022 314.82 138 .99 .022 9.61* 3 0.00 0.000
Communication 293.71 135 .99 .021 302.51 138 .99 .021 8.80* 3 0.00 0.000
Conflict 299.47 135 .99 .022 306.12 138 .99 .022 6.65 3 0.00 0.000
Peer Approval 343.06 135 .99 .024 363.11 138 .99 .025 20.05*** 3 0.00 0.001
Religiosity
Closeness 344.80 135 .99 .024 359.25 138 .99 .025 14.45** 3 0.00 0.001
Support 284.76 135 .99 .021 287.06 138 .99 .021 2.3 3 0.00 0.000
Monitoring 306.90 135 .99 .022 315.67 138 .99 .022 8.77* 3 0.00 0.000
Communication 301.93 135 .99 .022 311.12 138 .99 .022 9.19* 3 0.00 0.000
Conflict 307.67 135 .99 .022 314.10 138 .99 .022 6.43 3 0.00 0.000
Peer Approval 352.52 135 .99 .025 372.24 138 .99 .026 19.72*** 3 0.00 0.001
Note. The chi-square critical values at 3 df are as follows: *7.82 (p = .05), **11.35 (p = .01), and
***16.27 (p = .001).
indicating that there were differences in the way these models operated for male versus
female adolescents, univariate tests were conducted in order to more accurately determine
the specific sources of these differences.
Ritualistic Participation
Results of the univariate tests comparing the individual pathways by sex for the
ritualistic participation set of models indicated that, when examined more closely, the
differences in this model as indicated by the statistically significant chi-square difference
155
in the omnibus test were not attributable to the mediated pathway between religiosity and
deviance (see Table 26). In the models including closeness ()P2 = 10.38, p < .01),
communication ()P2 = 5.62, p < .05), and peer approval ()P2 = 17.98, p < .001), the
difference between male and female adolescents was found in the pathway between the
family processes measure and deviance. In the model including monitoring, the difference
was found in the pathway between the religiosity and family processes measures
()P2 = 4.13, p < .05). Thus, none of these analyses provided evidence for a moderating
effect by sex.
Relational Practice
Results of the univariate tests comparing the individual pathways by sex for the
relational practice set of models indicated that, when examined more closely, the
differences in this model as indicated by the statistically significant chi-square difference
in the omnibus test were not attributable to the mediated pathway between religiosity and
deviance (see Table 27). In all four of the models that yielded a significant chi-square
difference statistic in the omnibus test (closeness, )P2 = 11.20, p < .001; monitoring,
)P2 = 4.25, p < .05; communication, )P2 = 4.57, p < .05), and peer approval,
)P2 = 17.37, p < .001), the difference between male and female adolescents was found in
the pathway between the family processes measure and deviance. Thus, none of these
analyses provided evidence for a moderating effect by sex.
Religiosity
Results of the univariate tests comparing the individual pathways by sex for the
religiosity set of models indicated that, when examined more closely, the differences in
156
Table 26
Path Coefficients and Comparisons by Sex: Ritualistic Participation
Path Total Male Female Variant Invariant )P2
Ritualistic ? Closeness .33(.000) .24(.000) .41(.000) 308.12 310.97 2.85
Ritualistic ? Deviance -.03(.451) -.02(.689) -.07(.282) 308.12 308.18 0.06
Closeness ? Deviance -.50(.000) -.55(.000) -.40(.000) 308.12 318.50 10.38**
Ritualistic ? Monitoring .31(.000) .18(.006) .44(.000) 273.91 278.04 4.13*
Ritualistic ? Deviance -.02(.704) -.05(.329) -.04(.623) 273.91 274.13 0.22
Monitoring ? Deviance -.57(.000) -.57(.000) -.44(.001) 273.91 277.73 3.82
Ritualistic ? Communication .33(.000) .27(.000) .38(.000) 263.51 266.46 2.95
Ritualistic ? Deviance -.09(.024) -.07(.200) -.13(.022) 263.51 263.51 0.00
Communication ? Deviance -.31(.000) -.30(.000) -.26(.000) 263.51 269.13 5.62*
Ritualistic ? Peer Approval .27(.000) .23(.000) .28(.000) 319.65 319.87 0.22
Ritualistic ? Deviance -.09(.012) -.05(.374) -.17(.002) 319.65 320.13 0.48
Peer Approval ? Deviance -.38(.000) -.45(.000) -.23(.000) 319.65 337.63 17.98***
Note. The direct (unmediated) relationship between ritualistic participation and deviance was as follows:
Total, -.20(.000); Male, -.15(.003); Female, -.23(.000). The first number in each cell is the standardized
path coefficient while the second number (in parentheses) is the p value; the chi-square value indicated for
each path represents the chi-square of the model when that particular path is held invariant for the male and
female models. The change in chi-square value represents the difference between the chi-square of the fully
variant model and the chi-square of the model when that particular path is held invariant. The chi-square
critical values at 1 df (fully variant model df = 135; single path invariant model df = 136) are:
*3.84 (p = .05), **6.64 (p = .01), and ***10.83 (p = .001).
this model, as indicated by the significant chi-square difference statistic in the omnibus
test, were not attributable to the mediated pathway between religiosity and deviance (see
Table 28). In the models including closeness ()P2 = 10.70, p < .01), communication
()P2 = 5.09, p < .05), and peer approval ()P2 = 18.03, p < .001), the difference between
male and female adolescents was found in the pathway between the family processes
157
Table 27
Path Coefficients and Comparisons by Sex: Relational Practice
Path Total Male Female Variant Invariant )P2
Relational ? Closeness .36(.000) .32(.000) .39(.000) 331.62 332.87 1.25
Relational ? Deviance -.12(.004) -.09(.095) -.13(.029) 331.62 331.66 0.04
Closeness ? Deviance -.47(.000) -.53(.000) -.37(.000) 331.62 342.82 11.20***
Relational ? Monitoring .31(.000) .21(.001) .38(.000) 305.21 307.35 2.14
Relational ? Deviance -.12(.003) -.15(.000) -.12(.066) 305.21 306.53 1.32
Monitoring ? Deviance -.54(.000) -.55(.004) -.41(.001) 305.21 309.46 4.25*
Relational ? Communication .32(.000) .30(.000) .32(.000) 293.71 295.31 1.60
Relational ? Deviance -.19(.000) -.18(.001) -.20(.000) 293.71 294.66 0.95
Communication ? Deviance -.28(.000) -.26(.000) -.24(.001) 293.71 298.28 4.57*
Relational ? Peer Approval .31(.000) .29(.000) .30(.000) 343.06 343.30 0.24
Relational ? Deviance -.18(.000) -.14(.009) -.21(.000) 343.06 343.11 0.05
Peer Approval ? Deviance -.34(.000) -.42(.000) -.21(.001) 343.06 360.43 17.37***
Note. The direct (unmediated) relationship between relational practice and deviance was as follows:
Total, -.29(.000); Male, -.26(.000); Female, -.28(.000). The first number in each cell is the standardized
path coefficient while the second number (in parentheses) is the p value; the chi-square value indicated for
each path represents the chi-square of the model when that particular path is held invariant for the male and
female models. The change in chi-square value represents the difference between the chi-square of the fully
variant model and the chi-square of the model when that particular path is held invariant. The chi-square
critical values at 1 df (fully variant model df = 135; single path invariant model df = 136) are:
*3.84 (p = .05), **6.64 (p = .01), and ***10.83 (p = .001).
measure and deviance. In the model involving monitoring, though the chi-square
difference statistic from the omnibus test reached significance, the univariate tests
revealed that no specific pairwise pathway could be identified as the source of this
difference. Thus, none of these analyses provided evidence for a moderating effect by sex.
158
Table 28
Path Coefficients and Comparisons by Sex: Religiosity
Path Total Male Female Variant Invariant )P2
Religiosity ? Closeness .37(.000) .29(.000) .43(.000) 344.80 346.88 2.08
Religiosity ? Deviance -.08(.040) -.07(.202) -.11(.086) 344.80 344.82 0.02
Closeness ? Deviance -.48(.000) -.54(.000) -.38(.000) 344.80 355.50 10.70**
Religiosity ? Monitoring .34(.000) .22(.001) .45(.000) 306.90 309.52 2.62
Religiosity ? Deviance -.07(.089) -.10(.046) -.09(.242) 306.90 307.63 0.73
Monitoring ? Deviance -.55(.000) -.56(.000) -.42(.001) 306.90 310.69 3.79
Religiosity ? Communication .35(.000) .30(.000) .38(.000) 301.93 304.35 2.42
Religiosity ? Deviance -.16(.000) -.14(.009) -.18(.002) 301.93 302.37 0.44
Communication ? Deviance -.29(.000) -.27(.000) -.24(.001) 301.93 307.02 5.09*
Religiosity ? Peer Approval .31(.000) .28(.000) .32(.000) 352.52 352.68 0.16
Religiosity ? Deviance -.15(.000) -.10(.048) -.21(.000) 352.52 352.59 0.07
Peer Approval ? Deviance -.35(.000) -.43(.000) -.21(.001) 352.52 370.55 18.03***
Note. The direct (unmediated) relationship between religiosity and deviance was as follows:
Total, -.26(.000); Male, -.23(.000); Female, -.27(.000). The first number in each cell is the standardized
path coefficient while the second number (in parentheses) is the p value; the chi-square value indicated for
each path represents the chi-square of the model when that particular path is held invariant for the male and
female models. The change in chi-square value represents the difference between the chi-square of the fully
variant model and the chi-square of the model when that particular path is held invariant. The chi-square
critical values at 1 df (fully variant model df = 135; single path invariant model df = 136) are:
*3.84 (p = .05), **6.64 (p = .01), and ***10.83 (p = .001).
Part F
Hypothesis 4F predicted that this same pattern of mediation would be found for
both younger and older adolescents. In other words, this hypothesis tested the moderating
role of age for the structural model in which family processes mediated the effects of
religiosity on deviance. As with the tests involving sex, this role of moderation was tested
Table 29
Fit Statistics in Comparisons Between Fully Variant and Fully Invariant Models by Age
Fully Variant Model Fully Invariant Model Change Statistics
Family Processes P2 df CFI RMSEA P2 df CFI RMSEA )P2 )df )CFI )RMSEA
Ritualistic
Participation
Closeness 308.12 135 .99 .022 312.52 138 .99 .022 4.4 3 0.00 0.000
Support 248.65 135 .99 .018 253.66 138 .99 .018 5.01 3 0.00 0.000
Monitoring 273.91 135 .99 .020 279.67 138 .99 .020 5.76 3 0.00 0.000
Communication 263.51 135 .99 .019 269.12 138 .99 .019 5.61 3 0.00 0.000
Conflict 270.33 135 .99 .020 287.49 138 .99 .020 17.16*** 3 0.00 0.000
Peer Approval 319.65 135 .99 .023 324.19 138 .99 .023 4.54 3 0.00 0.000
Relational
Practice
Closeness 331.62 135 .99 .024 332.09 138 .99 .024 0.47 3 0.00 0.000
Support 274.73 135 .99 .020 275.42 138 .99 .020 0.69 3 0.00 0.000
Monitoring 305.21 135 .99 .022 309.97 138 .99 .022 4.76 3 0.00 0.000
Communication 293.71 135 .99 .021 295.28 138 .99 .021 1.57 3 0.00 0.000
Conflict 299.47 135 .99 .022 304.23 138 .99 .022 4.76 3 0.00 0.000
Peer Approval 343.06 135 .99 .024 343.51 138 .99 .024 0.45 3 0.00 0.000
Religiosity
Closeness 344.80 135 .99 .024 346.56 138 .99 .024 1.76 3 0.00 0.000
Support 284.76 135 .99 .021 286.24 138 .99 .021 1.48 3 0.00 0.000
Monitoring 306.90 135 .99 .022 311.17 138 .99 .022 4.27 3 0.00 0.000
Communication 301.93 135 .99 .022 304.30 138 .99 .022 2.37 3 0.00 0.000
Conflict 307.67 135 .99 .022 316.47 138 .99 .022 8.80* 3 0.00 0.000
Peer Approval 352.52 135 .99 .025 354.31 138 .99 .025 1.79 3 0.00 0.000
Note. The chi-square critical values at 3 df are: *7.82 (p = .05), **11.35 (p = .01), and ***16.27 (p = .001).
160
using a multiple-group SEM approach. Results generally supported the hypothesis (see
Tables 29, 30, and 31).
Results of the omnibus tests comparing the fully variant and fully invariant
models of mediation by age indicated that, even when all pathways in the model were
constrained to equality for younger and older adolescents, this model fit the data well,
with all CFI?s equal to .99 and RMSEA?s ranging from .018 to .025 (see Table 29). In the
sets of models involving both the ritualistic participation measure and the religiosity
measure, a significant chi-square difference between the fully variant and fully invariant
models was found for only one of the six (conflict; for ritualistic participation,
)P2 = 17.16, p < .001; for religiosity, )P2 = 8.80, p < .05). In the set of models involving
the relational practice measure, none exhibited a statistically significant difference by age
group.
Ritualistic Participation
Results of the univariate tests comparing the individual pathways by age groups
for the model including ritualistic participation and parental conflict indicated that, when
examined more closely, the differences in this model were not attributable to the
mediated pathway between religiosity and deviance (see Table 30). Rather, it was the
pathway between ritualistic participation and conflict that was found to vary between the
age groups ()P2 = 15.72, p < .001). Thus, none of these analyses provided evidence for a
moderating effect by age group.
161
Religiosity
Results of the univariate tests comparing the individual pathways by age groups
for the model including religiosity and parental conflict indicated that, when examined
more closely, the differences in this model were not attributable to the mediated pathway
between religiosity and deviance (see Table 31). Rather, it was the pathway between
Table 30
Path Coefficients and Comparisons by Age: Ritualistic Participation
Path Total Male Female Variant Invariant )P2
Ritualistic ? Conflict -.20(.000) -.36(.000) -.04(.580) 270.33 286.05 15.72***
Ritualistic ? Deviance -.11(.003) -.11(.062) -.12(.027) 270.33 270.33 0.00
Conflict ? Deviance .41(.000) .40(.000) .45(.000) 270.33 271.90 1.57
Note. The direct (unmediated) relationship between ritualistic participation and deviance was as follows:
Total, -.20(.000); Younger, -.26(.000); Older, -.14(.007). The first number in each cell is the standardized
path coefficient while the second number (in parentheses) is the p value; the chi-square value indicated for
each path represents the chi-square of the model when that particular path is held invariant for the younger
and older adolescent models. The change in chi-square value represents the difference between the chi-
square of the fully variant model and the chi-square of the model when that particular path is held invariant.
The chi-square critical values at 1 df (fully variant model df = 135; single path invariant model df = 136)
are: *3.84 (p = .05), **6.64 (p = .01), and ***10.83 (p = .001).
Table 31
Path Coefficients and Comparisons by Age: Religiosity
Path Total Male Female Variant Invariant )P2
Religiosity ? Conflict -.20(.000) -.31(.000) -.12(.142) 307.67 315.44 7.77**
Religiosity ? Deviance -.18(.000) -.16(.003) -.19(.001) 307.67 307.68 0.01
Conflict ? Deviance .40(.000) .38(.000) .43(.000) 307.67 309.29 1.62
Note. The direct (unmediated) relationship between religiosity and deviance was as follows:
Total, -.26(.000); Younger, -.29(.000); Older, -.24(.000). The first number in each cell is the standardized
path coefficient while the second number (in parentheses) is the p value; the chi-square value indicated for
each path represents the chi-square of the model when that particular path is held invariant for the younger
and older adolescent models. The change in chi-square value represents the difference between the chi-
square of the fully variant model and the chi-square of the model when that particular path is held invariant.
The chi-square critical values at 1 df (fully variant model df = 135; single path invariant model df = 136)
are: *3.84 (p = .05), **6.64 (p = .01), and ***10.83 (p = .001).
162
ritualistic participation and conflict that was found to vary between the age groups
()P2 = 7.77, p < .01). Thus, none of these analyses provided evidence for a moderating
effect by age group.
Part G
The following 3 hypotheses (4G, 4H, 4I) were somewhat peripheral to the central
focus of this section and add very little to the results concerning mediation. Hypothesis
4G proposed that the mediating role of family processes between religiosity and deviance
would be clarified as that of partial mediation, involving both direct and indirect
pathways to deviance. This hypothesis was largely supported (see Table 32). When
parameter estimates were examined (see Table 22), results indicated that both the direct
(religiosity-deviance) and indirect (religiosity-family processes-deviance) pathways were
significant. In tests including the ritualistic participation scale, 2 of the 6 direct pathways
were nonsignificant (parental closeness, $ = -.03, p = .45; parental monitoring, $ = -.02,
p = .70), and in tests including the religiosity scale, 1 of the 6 direct pathways was
nonsignificant (parental monitoring, $ = -.07, p = .09). Tests including the relational
practice scale, however, indicated that all 6 of the direct and indirect pathways were
statistically significant. For the sake of completeness, direct, indirect, and total effects for
the total sample and by sex and age groups are presented in Table 32.
Part H
Hypothesis 4H stated that statistically significant direct and indirect pathways
would be equally present for both male and female adolescents. In essence, this was a test
of moderation (very similar to Hypothesis 4E) that examined whether the pattern of
163
parameter estimates within the mediation models would be invariant by sex. However,
unlike Hypothesis 4E, the focus of this analysis was not only on a single parameter of
interest (e.g., the religiosity-deviance parameter), but rather on the totality of relationships
among the measures of importance (i.e., religiosity, family processes, deviance).
Consequently, using the same multi-group omnibus test to compare fully variant and fully
invariant models of mediation (see Tables 25, 26, 27, and 28), this hypothesis was not
confirmed. Of 16 total mediation models tested, 12 yielded significant chi-square
difference statistics (the same 4 models ? closeness, monitoring, communication, peer
approval ? in all 3 religiosity scale sets), thus indicating that these models were not
invariant by sex, but rather, that the patterns of association within each model were
different for male and female adolescents.
Part I
Similar to Hypothesis 4H, Hypothesis 4I stated that statistically significant direct
and indirect pathways would be equally present for both younger and older adolescents.
Again, the question was similar to (but not the same as) Hypothesis 4F, that is, is the
pattern of parameter estimates within the mediation models invariant by age group?
Using the same multi-group omnibus test, this hypothesis was largely confirmed (see
Tables 29, 30, and 31). Of 16 total mediation models tested, 14 did not yield significant
chi-square difference statistics, thus indicating that almost all of the models were
invariant by age group. For 2 of the models (the conflict model in both the ritualistic
participation and religiosity sets), a significant chi-square difference statistic was found
that indicated that these 2 models were not invariant by age group.
Table 32
Direct, Indirect, and Total Effects of Religiosity Scales on Family Processes Scales in Fully Variant Model by Sex, Age, and for Total Sample
Sex Age
Male Female Younger Older Total
D I T D I T D I T D I T D I T
Ritualistic
Participation
Closeness -.021 -.130 -.151 -.067 -.164 -.231 -.093 -.160 -.252 .029 -.167 -.139 -.030 -.163 -.194
Support -.078 -.074 -.152 -.152 -.080 -.232 -.141 -.112 -.253 -.089 -.051 -.141 -.119 -.076 -.195
Monitoring -.050 -.101 -.151 -.037 -.196 -.232 -.044 -.207 -.251 .022 -.157 -.135 -.016 -.175 -.191
Communication -.068 -.080 -.148 -.132 -.098 -.230 -.176 -.075 -.251 .000 -.137 -.137 -.088 -.102 -.190
Conflict -.025 -.126 -.151 -.157 -.075 -.232 -.111 -.143 -.254 -.121 -.020 -.141 -.114 -.081 -.196
Peer Approval -.046 -.104 -.150 -.167 -.064 -.231 -.157 -.093 -.250 -.031 -.109 -.140 -.093 -.100 -.193
Relational
Practice
Closeness -.091 -.168 -.259 -.132 -.146 -.278 -.108 -.158 -.266 -.120 -.185 -.305 -.116 -.169 -.285
Support -.170 -.088 -.259 -.171 -.106 -.278 -.154 -.112 -.266 -.216 -.091 -.307 -.192 -.094 -.286
Monitoring -.147 -.113 -.260 -.123 -.157 -.280 -.108 -.161 -.269 -.134 -.168 -.302 -.120 -.165 -.285
Communication -.178 -.078 -.256 -.196 -.079 -.275 -.192 -.070 -.263 -.198 -.195 -.303 -.193 -.088 -.281
Conflict -.149 -.111 -.260 -.187 -.090 -.277 -.160 -.107 -.267 -.237 -.070 -.307 -.208 -.079 -.287
Peer Approval -.138 -.120 -.258 -.212 -.063 -.275 -.166 -.095 -.262 -.181 -.126 -.307 -.177 -.106 -.283
Religiosity
Closeness -.069 -.157 -.226 -.109 -.164 -.273 -.112 -.170 -.282 -.053 -.187 -.240 -.084 -.176 -.260
Support -.146 -.080 -.226 -.181 -.093 -.274 -.177 -.108 -.285 -.166 -.076 -.241 -.178 -.085 -.263
Monitoring -.103 -.123 -.226 -.086 -.189 -.275 -.078 -.207 -.285 -.056 -.179 -.235 -.072 -.188 -.260
Communication -.141 -.082 -.223 -.181 -.091 -.272 -.201 -.079 -.280 -.111 -.126 -.237 -.158 -.099 -.257
Conflict -.102 -.105 -.226 -.194 -.079 -.273 -.164 -.120 -.285 -.190 -.052 -.242 -.183 -.080 -.263
Peer Approval -.104 -.121 -.224 -.207 -.065 -.273 -.182 -.098 -.280 -.113 -.128 -.241 -.150 -.110 -.260
Note. Numbers in this table have been transcribed directly from the AMOS tabular output and may reflect rounding error in that the direct and indirect effects
may not add up exactly to equal the total effect.
165
Hypothesis 5: Moderation
Part A
Hypothesis 5A predicted that religiosity would moderate the relationship between
family processes and deviant behavior. In order to test this hypothesis, a median split was
used to create a categorical variable with two levels representing low and high levels of
religiosity. Because the relational practice variable was formulated as a representation of
the best conceptual approach to the measurement of religiosity, the standardized,
residualized (i.e., after controlling for race and SES) relational practice variable was used
to create these religiosity groups. Next, a multi-group omnibus test was completed,
comparing fully variant (i.e., all parameters free to vary) and fully invariant
(i.e., parameter between family processes and deviance constrained to equality for
religious groups) models of the moderating relationship of religiosity between each
family processes variable and deviance (see Table 33). Results of these analyses did not
support the hypothesis. For all six family processes variables, the chi-square difference
statistic was not significant. This was an indication that the relationship between family
processes and deviance was not moderated by religiosity groups.
Part B
Hypothesis 5B proposed that this same pattern of moderation (i.e., religiosity
moderating the relationship between family processes and deviant behavior) would be
found for both male and female adolescents. A multi-group omnibus test conducted
separately for male and female adolescents was used to test this hypothesis. Results
suggested that the same pattern found in the total sample was also apparent for both male 
166
Table 33
Moderation of Family Processes by Religiosity: Path Coefficients and Comparisons
Path Total Low High Variant P2 Invariant P2 )P2 p
Parental Closeness ? Deviance -.51(.000) -.45(.000) -.44(.000) 439.49 440.52 1.03 .31
Parental Support ? Deviance -.41(.000) -.37(.000) -.39(.000) 328.07 329.37 1.29 .26
Parental Monitoring ? Deviance -.58(.000) -.58(.000) -.48(.000) 397.71 400.14 2.43 .12
Parental Communication ? Deviance -.32(.000) -.32(.001) -.22(.045) 321.47 324.03 2.56 .11
Parental Conflict ? Deviance .43(.000) .39(.000) .40(.000) 342.82 345.53 2.71 .10
Parental Peer Approval ? Deviance -.40(.000) -.33(.000) -.42(.000) 453.09 453.20 0.10 .75
Note. The first number in each cell is the standardized path coefficient while the second number (in parentheses) is the
p value; the variant chi-square value indicated for each path represents the chi-square of the fully variant model while
the invariant chi-square value represents the chi-square of the fully invariant model, i.e., when the path is held invariant
for the low and high religiosity groups. The change in chi-square value represents the difference between the variant
chi-square and the invariant chi-square. The chi-square critical values at 1 df (fully variant model df = 220; single path
invariant model df = 221) are: *3.84 (p = .05), **6.64 (p = .01), and ***10.83 (p = .001).
and female adolescents (see Table 34). Among female adolescents, none of the model
parameters were significantly different across the two groups. For male adolescents, 5 of
the 6 chi-square difference statistics did not reach statistical significance. However, a
statistically significant difference (fully variant P2 = 321.47; male invariant P2 = 325.61;
)P2 = 4.14; p = .04) was found between low religiosity male adolescents ($ = -.35,
p = .00) and high religiosity male adolescents ($ = -.16, p = .35) for parental
communication.
Part C
Hypothesis 5C stated that this same pattern of moderation would be found for
both younger and older adolescents. A multi-group omnibus test conducted separately for
younger and older adolescents was used to test this hypothesis. Results suggested that the
same pattern found in the total sample was also apparent for both younger and older
Table 34
Moderation of Family Processes by Sex and Religiosity: Path Coefficients and Comparisons
Variant
P2
Males Females
Path Total Low High
Invariant
P2 )P2 p Low High
Invariant
P2 )P2 p
Parental Closeness ? Deviance -.51(.000) 439.49 -.49(.000) -.53(.000) 439.57 0.07 .79 -.36(.062) -.35(.001) 439.56 0.07 .79
Parental Support ? Deviance -.41(.000) 328.07 -.38(.000) -.35(.007) 329.94 1.87 .17 -.40(.000) -.44(.000) 328.59 0.52 .47
Parental Monitoring ? Deviance -.58(.000) 397.71 -.61(.000) -.50(.000) 397.86 0.15 .70 -.41(.024) -.33(.071) 400.13 2.41 .12
Parental Communication ? Deviance -.32(.000) 321.47 -.35(.000) -.16(.353) 325.61 4.14 .04 -.23(.242) -.23(.051) 321.48 0.01 .93
Parental Conflict ? Deviance .43(.000) 342.82 .46(.000) .49(.011) 343.99 1.17 .28 .41(.037) .46(.000) 343.75 0.93 .33
Parental Peer Approval ? Deviance -.40(.000) 453.09 -.42(.000) -.44(.000) 454.13 1.04 .31 -.15(.203) -.38(.000) 454.49 1.39 .24
Note. The first number in each cell is the standardized path coefficient while the second number (in parentheses) is the p value; the variant chi-square value
indicated for each path represents the chi-square of the fully variant model while the invariant chi-square value represents the chi-square of the fully invariant
model, i.e., when the path is held invariant across all 4 groups (low religious males, low religious females, high religious males, high religious females). The
change in chi-square value represents the difference between the variant chi-square and the invariant chi-square. The chi-square critical values at 1 df (fully
variant model df = 220; single path invariant model df = 221) are: *3.84 (p = .05), **6.64 (p = .01), and ***10.83 (p = .001).
168
adolescents (see Table 35). Among both younger and older adolescents, religiosity was
not found to moderate the relationship between family processes and deviant behavior
since, based on the chi-square difference statistics, none of the model parameters differed
between the low and high religiosity groups.
Table 35
Moderation of Family Processes by Age and Religiosity: Path Coefficients and Comparisons
Variant
P2
Younger Older
Path Total Low High
Invariant
P2 )P2 p Low High
Invariant
P2 )P2 p
Parental Closeness ? Deviance -.51(.000) 439.49 -.49(.003) -.36(.011) 442.07 2.58 .11 -.43(.000) -.53(.000) 434.68 0.19 .67
Parental Support ? Deviance -.41(.000) 328.07 -.44(.000) -.41(.000) 330.38 2.31 .13 -.32(.005) -.38(.002) 328.10 0.03 .86
Parental Monitoring ? Deviance -.58(.000) 397.71 -.68(.004) -.53(.001) 401.19 3.48 .06 -.54(.000) -.45(.000) 397.92 0.21 .65
Parental Communication ? Deviance -.32(.000) 321.47 -.40(.030) -.15(.450) 325.29 3.82 .05 -.26(.024) -.30(.030) 321.50 0.03 .87
Parental Conflict ? Deviance .43(.000) 342.82 .43(.004) .42(.005) 344.62 1.81 .18 .39(.011) .40(.014) 343.89 1.07 .30
Parental Peer Approval ? Deviance -.40(.000) 453.09 -.35(.001) -.35(.004) 454.32 1.23 .27 -.33(.001) -.48(.000) 453.48 0.39 .54
Note. The first number in each cell is the standardized path coefficient while the second number (in parentheses) is the p value; the variant chi-square value
indicated for each path represents the chi-square of the fully variant model while the invariant chi-square value represents the chi-square of the fully invariant
model, i.e., when the path is held invariant across all 4 groups (low religious younger, low religious older, high religious younger, high religious older). The
change in chi-square value represents the difference between the variant chi-square and the invariant chi-square. The chi-square critical values at 1 df (fully
variant model df = 220; single path invariant model df = 221) are: *3.84 (p = .05), **6.64 (p = .01), and ***10.83 (p = .001).
170
CHAPTER 5: DISCUSSION
By focusing on several distinct issues in the debate concerning whether religiosity
has a meaningful negative relationship to deviant behavior, the current study represents
an important next step for research in this area. It has clarified that there is a moderate yet
consistent bivariate relationship between religiosity and deviance and that this
relationship is not attenuated by other important control factors nor does this relationship
appear to differ by sex or age group. In addition, the importance of the development and
use of multi-item measures of religiosity has been highlighted. Furthermore, the current
findings reaffirm religiosity as an important protective factor against most, if not all,
types of deviant behavior (not just some, e.g., antiascetic). Finally, the current study
provides information regarding the interplay between family processes and religiosity in
the deterrence of deviant behavior among adolescents by testing models involving
mediation, moderation, and combination.
Does religiosity matter?
The fact that religiosity has a consistent, though moderate, negative relationship
with adolescent deviant behavior, as found in several recent studies (e.g., Baier & Wright,
2001; Johnson et al., 2000), was confirmed again in this study. When using multi-item
scales, the relationship between religiosity and deviance, after partialling out the effects
of race and SES, was r = -.31 (r = -.28 after also including sex and age as controls). Thus,
171
religiosity explains approximately 8%-9% of the variability in deviance. One of the
important debates in this area of the literature has been whether or not the deterrent effect
of religiosity on deviance is important and independent, or whether it can be accounted
for by other socially constructive or protective factors. Several contingency explanations
have been offered, such as Stark, Kent, and Doyle?s (1982; see also Stark & Bainbridge,
1996; Tittle & Welch, 1983) idea that religiosity is only important in the social context of
a moral community or Benda and Corwyn?s (1997; see also Cochran et al., 1994)
contention that the effect of religiosity is shown to be spurious when other social control
measures are involved. However, evidence continues to mount indicating that religiosity
should be viewed as having a consistently important negative impact on deviant behavior
among adolescents. In addition to the current study as well as meta-analyses such as Baier
and Wright (2001) and Johnson et al. (2000), a recent examination by Smith and Faris
(2002) of the nationally representative data from the Monitoring the Future study showed
that religious twelfth-grade students were less likely to participate in any number of
deviant activities, such as substance abuse, risky behavior, crime and violence, and school
problems (see also Johnson et al., 2001). Thus, the cumulative evidence suggests that
religiosity has a consistent negative relationship with adolescent deviant behavior that is
not accounted for by race, SES, sex, age, or family processes.
What is the role of sex?
The literature concerning religiosity and deviance has not been clear at all on the
role of sex. It is known that adolescent girls report themselves to be more religious than
boys. For instance, Smith, Faris, Denton, and Regnerus (2003), in a recent summary of
172
Monitoring the Future and AddHealth data, found that more adolescent girls than boys
say their faith is very important to them while more adolescent boys than girls say that
their faith is not important to them. Similarly, they found that more girls than boys pray
daily and more boys than girls never pray. The more important question is, because
adolescent girls report being more religious, does that religious faith somehow strengthen
their resistance to deviant behavior more so than for boys? As a matter of fact, in previous
literature, there has been only a very slight indication of the opposite, that perhaps the
deterrent effect of religiosity is stronger for male than female adolescents (e.g., Baier &
Wright, 2001). In this case, the logic may be that, since religious faith is not typically as
important to adolescent boys as it is to girls, when boys do have faith, it provides a
stronger deterrent to deviant behavior than for girls, who may take their faith for granted.
The current findings did not support either of these conclusions. Results provided
evidence for the idea that the manner in which religiosity was related to deviance was
invariant by sex. That is, measures of religiosity did not have a stronger relationship with
deviant behavior for boys than for girls or vice versa, but rather, seemed equally
important for both as a potential protective factor against deviant behavior.
Recently, Regnerus (2003) has made an interesting finding that increases in
parental religiosity are actually associated with significant increases in delinquency for
adolescent boys while, for girls, the opposite is true, namely that increases in parental
religiosity yield significant decreases in delinquent behavior. Based on these findings,
Regnerus concludes that, for adolescent girls, religiosity in one?s parents is ?largely
protective? while, for adolescent boys, ?parents? religiosity . . . appears to backfire?
173
(p. 200). A subsequent study by Pearce and Haynie (2004) clarifies this issue through
findings which indicate that, when parent and child report being ?religiously different?
(i.e., religious parent, non-religious child), deviant behavior is much more likely from the
adolescent child.
These findings may or may not have bearing on the current study. On the one
hand, the current study focuses on the simple relationship between an adolescent?s
individual declaration of religiosity and the consistency of their behavioral outcomes
associated with that declaration. In this sense, the religiosity of the parent has no bearing;
whether or not their parent is religious does not matter, but rather, what matters is their
own religiosity. On the other hand, the current study does make an attempt to expand the
literature in the area of the interaction between religiosity and family processes in the
deterrence of deviance. In this sense, this new line of research bears heavily, though it is
outside the framework of the current investigation. In essence, what Regnerus (2003) and
Pearce and Haynie (2004) have found is that the intergenerational religiosity dynamic
may be a viable dimension of family processes. Citing evidence from Rossi and Rossi
(1990) and Pearce and Axinn (1998), they make a case for the fact that family religiosity
is a process, like any other of the family processes, that plays a role in family functioning.
What is the role of age?
The role of age in the religiosity-deviance literature has been similarly unclear.
Though Benda and Toombs (2000) found some indication that older age may be related to
stronger deterrent effects for religiosity, this was only an interaction effect (more church
attendance coupled with older age) and only for those older than 24 years of age or older.
174
The current findings provide support for the idea that the effect of religiosity on deviance
does not vary by age group. Evidence from Smith et al.?s (2003) examination of
nationally representative data sets suggests that, though there are some trends indicating a
decline of religiosity with age (e.g., frequency of prayer, percentage of those calling
themselves ?born again?), these differences do not seem as important as the trend toward
stability. This is more in line with the current results that seem to indicate that the effect
of religiosity on deviance does not change with age, at least during the adolescent years.
Evidently, according to these findings, religious faith did not have a stronger relationship
with deviance for older adolescents than for younger adolescents or vice versa. Thus,
perhaps when a young person makes a personal commitment to religious faith and
practice, a lifestyle involving less deviant behavior may typically accompany such a
commitment, whether they are male or female, younger or older.
What is the role of measurement?
Recent work in the area of youth religiosity confirms that more emphasis is being
placed on the importance of methodology in this area of research (Denton & Smith,
2001). One particular area that has recently received significant attention is the use of
single-item versus multi-item measures (Johnson et al., 2000). Most of the earlier
research on religiosity and deviance relied on single-item measures, such as church
attendance (e.g., Stark, 1969). However, it is commonly understood that, whenever
possible, the use of a multi-item scale provides a much more reliable and robust measure
of religious behavior (Smith et al., 2003). The current findings support this idea. Three
multi-item scales were tested in the current study, utilizing various conceptually-guided
175
configurations of the same six religiosity items. Results indicated that the relational
practice scale had a stronger relationship with deviant behavior than most of the single-
item measures, including church attendance. However, attempting to prove the advantage
of one particular measure over another was not a central focus of the current study.
Rather, the limited analyses in this area were intended only to provide impetus for future
research in a direction away from single-item measures and toward multi-item measures
by providing a helpful illustration of a possible approach to such. Along that line, it
should be noted that the relational practice scale, including just 3 items focusing on
personal rather than formal manifestations of religious faith and practice, achieved a
consistently strong negative relationship with deviant behavior, the size of which was
significantly larger than that involving any of the single-item measures in the study. In
addition, the scale was found to be internally consistent. To achieve an alpha reliability of
.75 using only 3 items is notable; however, there also seems to be room for improvement
since an even better reliability would decrease attenuation, and consequently, increase the
strength of association between predictor and outcome measures. Thus, these findings
could be viewed as a ?step in the right direction.? The development of a more extensive
and/or comprehensive religiosity scale that would exhibit even greater reliability and that
may also demonstrate greater predictive strength in the area of adolescent deviance would
be a next logical next step for research in this area. One approach may be to add content-
valid items to the scale in order to more broadly assess the construct of relational
religiosity. Another approach could involve the formulation of a multidimensional
measure that includes known dimensions or aspects of religiosity that were not assessed
176
in the current study. For example, Smith has proposed that there may be 3 dimensions to
religiosity ? moral order, learned competencies, social and organizational ties ? and that
each of these dimensions may have 3 sub-dimensions. On the other hand, it is also
possible that the current study approached the limits of how much variability religiosity
can explain in adolescent deviance. Only future work which tests this question more
extensively with larger scales and measures that are more broadly constructed will be able
to address and answer this question.
For many years, researchers have struggled to determine and develop the best way
to measure religiosity. Part of the problem is that religiosity itself can be conceptualized
very differently, depending on one?s perspective or agenda. Because of this, there is no
lack of measures of ?religiosity? (see Hill & Hood, 1999, for a comprehensive
compilation of measures of religion and spirituality), but there is still no consensus on the
best way to measure religiosity when attempting to test its relationship to adolescent
deviant behavior. The current results suggest that the most effective means of capturing
this relationship may be to focus on specific behaviors that exhibit a sincere lifestyle
expression of faith (e.g., church status, religious involvement, prayer) rather than mere
formal or ritualistic behaviors (e.g., church attendance) or even claims of spirituality or
belief (e.g., religious salience).
Does religiosity deter all types of deviance or only some?
Some research on the effects of religiosity on deviance has suggested that
religiosity has a differential effect depending on the type of deviance seeking to be
explained. For instance, there has been some evidence to support the notion that
177
religiosity will be more strongly related to forms of deviance that do not involve victims
but rather are mostly self-destructive (e.g., substance abuse; Albrecht et al., 1977;
Chadwick & Top, 1993). Jensen and Erickson (1979) proposed that the reason for this
may be that certain deviant behaviors (e.g., violence) are considered harmful by society at
large while other more ?mild? forms of deviance (e.g., alcohol use) are more vehemently
condemned only by churches in their moral teaching. Thus, they labeled this distinction
antiascetic (referring to the more ?mild? forms of deviance condemned more loudly by
churches than society) versus secular (referring to the more ?severe? forms of deviance
condemned equally by churches and society).
Results from the current study did not provide support for this line of research. In
fact, results indicated that there were no differences in the magnitude of the relationship
between religiosity and various types of deviance (i.e., victim/nonvictim,
secular/antiascetic). Rather, the current findings suggested that religious faith was
negatively associated with all types of deviance, no matter how they were categorized. In
addition, the findings emphasized that this effect was true regardless of sex or age group.
So, for both male and female adolescents as well as for both younger and older
adolescents, more religiosity was found to be significantly related to less deviant behavior
of all types.
This finding is an important addition to a growing body of literature that indicates
that religiosity is a robust social protective factor that should not be compartmentalized in
its range of impact. Research has found that religiosity is a key factor in adolescent health
promotion since it is positively associated with health-enhancing behaviors (Jessor,
178
Turbin, & Costa, 1998) and negatively associated with health-compromising behaviors
(Wallace & Forman, 1998), including early initiation of sexual activity, number of sexual
partners, and teenage pregnancy (Lammers, Ireland, Resnick, & Blum, 2000; Murry,
1994; Thornton & Camburn, 1989; Whitehead, Wilcox, & Rostosky, 2001) as well as
with suicidal ideation and attempts (Donahue, 1995). Furthermore, research suggests that
religious youth and their families report greater overall satisfaction with their lives and
more involvement with their families (Varon & Riley, 1999), more effective coping
strategies (Shortz & Worthington, 1994), and lower levels of depression and hopelessness
(Wright, Frost, & Wisecarver, 1993) than those who are non-religious. Evidently, based
on this wide range of impact, religiosity should be viewed as an important protective
factor, to be considered along with other important protective factors in the endeavor to
identify strategies that work to decrease adolescent deviance. Thus, according to the
current study, religiosity is not only important for the deterrence of more mild forms of
deviance, but has relevance to deviance in all its forms.
Do family processes mediate the effect of religiosity on deviance?
The importance of family processes in the prediction of adolescent deviant
behavior is well established (Cernkovich & Giordano, 1987, 1992; Gorman-Smith, Tolan,
& Henry, 2000; Lipsey & Derzon, 1998; Zhang & Messner, 1995). In particular, parental
closeness and monitoring have been found to be consistent predictors of deviance. The
strength of family processes in this relationship has led some to propose that the effect of
religiosity on deviance is unimportant when family processes are considered
simultaneously. Benda and Toombs (2000) found that, when certain family processes
179
(e.g., attachment to mother and father) were added to a regression equation in a block, the
effect of their single-item measure of religiosity (church attendance) was attenuated. In
the current study, utilizing six family processes measures and eight single-item religiosity
measures, the findings indicated that the effect of religiosity on deviance was never
mediated, whether measured with single items or scales.
These results indicated that not all single-item measures of religiosity are equal. In
other words, it is important to note that, not only are multi-item measures of religiosity
more highly associated with deviance than single-item measures, but certain single-item
measures seem to be more highly associated with deviance than others. Consequently, the
validity of research findings that have declared negligible the impact of religiosity on
deviance by utilizing a single-item measure (e.g., church attendance) may be reevaluated
on the basis of this understanding. Unfortunately, such research, characterized by Smith
(2003b) as reductionistic thinking, has oversimplified the importance of religiosity in
adolescent development. Recently, Smith (2003b) has chastised social scientists who
practice such reductionism through
claims that what on the surface appears to be religious phenomena are in fact
revealed by serious analyses to really only be about other things quite unrelated to
religion. Thus, what appears to be divine or spiritual or transcendent or pious or
sacred are really only about social class, race, gender, ethnicity, nationalism,
solidarity, social control, and so on. . . . this general tendency toward such
reductionism and many (though not all) of the specific cases of analytical
reductionism are intellectually parochial and simplistic. (p. 19)
180
The current findings provide evidence that religiosity is an important social
control construct in the etiology of adolescent deviant behavior. They illustrate that the
effects of religiosity do not seem to be diminished by other important protective factors.
Thus, while some single-item measures of religiosity, such as church attendance, will
almost certainly be attenuated by other important social control measures, such as family
processes (though this was not the case in the current study), this type of evidence should
not be used to attempt to prove that religiosity is unimportant, since other single-item
measures (not to mention multi-item scales) will maintain an important relationship with
deviance, even in the presence of such variables. Consequently, researchers who intend to
study religiosity should carefully select single items with greater robustness (e.g., church
status) or utilize more reliable multi-item measures in order to assess the full complexity
of religiosity.
Indeed, results such as those in the current study give credence to the
conceptualization of religiosity as a multidimensional construct. In other words, there is
more to being ?religious? than just going to church. There are numerous young people
who attend church ?religiously,? yet they do not consistently live by the morals to which
they are so frequently exposed during those church gatherings. They may attend church
because their parents force them to do so (see Smith, 2003a, for evidence of parental
expectations on adolescents? church participation) or for the purpose of developing social
networks (i.e., all of their friends go and they use it as a time to socialize). There may
even be those who attend church because they are ?supposed to? based on
cultural/societal, familial, or theological expectations, yet they are disillusioned with what
181
they view as inauthentic institutional religion (Beaudoin, 2000; Zoba, 1999) or simply
bored by what they see as an adult-dominated societal structure that does not relate to the
needs of young people (Rabey, 2001). In these far from uncommon scenarios, the young
person?s presence at church may have very little impact on their life choices. Though they
hear a considerable amount of moral teaching that, if applied, would benefit their life by
curbing their participation in deviant behaviors, they choose to ignore the teaching and
participate in the deviance anyway. Thus, there are some who, when asked as part of a
research study, may indicate frequent church attendance, but their lifestyle patterns and
decisions may not reflect any morally different behavior. Consequently, research based on
church attendance as a measure of religiosity may, not surprisingly, find little relationship
between ritualistic attendance at church services and deviant behavior.
The multidimensional nature of religiosity has been recently advocated by Smith
(2003b) in his attempt to theorize the effect of religiosity on adolescents. In this article,
Smith attempts to identify the specific means by which religiosity may positively affect
adolescents. This attempt is grounded in sociological theory and presents a
multidimensional conceptualization of religiosity as a cluster of nine ?distinct but
connected and potentially mutually reinforcing factors? that ?cluster as groups of three
beneath three larger conceptual dimensions of social influence? (p. 19). This represents
an important advance in the study of religiosity. Identification of the multidimensional
nature of religiosity and utilization of these dimensions in its measurement is an essential
ingredient in future research. It is possible that doing so may enable researchers not only
to understand the multilevel importance of religiosity as an influence on adolescent
182
behavior but also to identify an even broader, more extensive range of impact, thus
reinforcing the notion that religiosity is important in the deterrence of all forms of
deviance, not just some.
Results regarding the mediation of the multi-item religiosity measures by family
processes revealed that relational practice, rather than religiosity, provided the most
robust measure of religious behavior (i.e., was attenuated the least often), while ritualistic
participation, as hypothesized, was not as robust. These findings provide great insight into
the measurement of religiosity. More specifically, they support the concept that there are
at least two very different ways to approach the measurement of religiosity in research
studies. The first focuses on behaviors or beliefs that indicate a formal commitment to a
set of ritualistic behaviors that have little or nothing to do with actual morality in life,
while the second focuses on behaviors or beliefs that result from a perceived relational
encounter with a supernatural being precipitating the development of a set of internally
motivated beliefs and practices. Measures utilizing the first approach (ritualistic
participation) are not as strongly related to deviance as those utilizing the second
(relational practice), though they are not unrelated. This may be an indication that the
presence of high levels of ritualism provides inconclusive information regarding a young
person?s participation in deviant behavior. That is, this approach to measurement may be
inherently incomplete and incapable of presenting a well-rounded portrait of a religious
person. More specifically, both ritualistically religious and relationally religious young
people will exhibit high levels of ritualistic religiosity (e.g., church attendance, reading
books of spiritual teachings, etc.); therefore, the variable is confounded by the presence of
183
both young people who apply their faith to daily life and those who do not. However, the
measurement approach that emphasizes the relational element of religiosity may offer a
more narrow focus, thus screening out any who may desire to appear religious without
ever altering their lifestyle.
Does religiosity moderate the effect of family processes on deviance?
In assessing the interplay of religiosity and family processes in their relationship
with deviance, a next logical question may be this: Do religious families somehow differ
in the way family processes affect adolescent deviant behavior for them? Results from the
current study indicate the answer to that question is evidently, no they do not. In other
words, these findings reveal that the manner in which family processes influence deviant
behavior does not differ for religious versus nonreligious families. This same pattern was
found to be mostly true when examined by sex and age group, thus indicating that family
processes deter deviance in a similar manner regardless of level of religiosity, sex, or age.
The only discovery in these findings was that, among male adolescents, parental
communication had a stronger relationship with deviance for those exhibiting low levels
of religiosity than among those exhibiting high levels of religiosity.
This may be explained by the fact that religious families may rely more heavily on
parental edict than persuasion (see, for instance, Bartkowski & Wilcox, 2000, who study
?parental yelling? among Protestant families) and may more frequently expect immediate
compliance to parental demands (Danso et al., 1997). Since religious families are known
to more frequently use corporal punishment (Ellison et al., 1996a, 1996b; Gershoff et al.,
1999; Xu, Tung, & Dunaway, 2000) and since use of corporal punishment is strongly
184
related to parental expectations for immediate child behavioral compliance (Gershoff,
2002; Roberts & Powers, 1990), it may be that, in religious families, children are more
frequently expected to obey without discussion or to do what the parents say without
questioning, even if what they are saying is inconsistent with their own behavior (Alain,
1989). If this is the case, then religious parents may grow lazy in the application of
communication skills to situations involving the adolescent?s opportunity to be involved
in deviant behavior, choosing to issue edicts with the expectation of immediate and
unquestioning obedience rather than to discuss with the adolescent the nature of the
dilemma and all possible solutions, resulting in a mutually acceptable decision for the
best possible course of action. This approach to problem-solving may make the parent-
adolescent communication process less important in the deterrence of deviant behavior.
In other words, having good parent-adolescent communication may make no difference in
the choice to be involved in deviant behavior for boys in these type of homes.
The idea that parents would unilaterally expect compliance to nondeviant
behavior from their adolescents is viewed ambivalently by those who value individuality.
In fact, parents who do so are often viewed as ?psychologically controlling? (e.g., Barber,
2002). However, it is important to remember that adolescents are not always
developmentally or cognitively capable of making the best choices autonomously. Parents
must sometimes direct behavior in an attempt, not just to achieve short-term behavioral
conformity, but long-term dispositional compliance (the internalization of general norms
of compliance to authority; Baumrind, 1983; Baumrind, Larzelere, & Cowan, 2002). For
instance, Bates, Pettit, Dodge, and Ridge (1998) have found that high levels of parental
185
control are associated with better behavior for children exhibiting a difficult
temperamental disposition (see also Kochanska, 1993, 1995, 1997). Thus, the current
findings may reflect a more balanced perspective which seeks to acknowledge that one of
the strengths of authoritative parenting is its insistence on behavioral conformity,
sometimes even without communication between parent and teen. In other words,
authoritative parents may utilize a balance of education and edict which includes making
every reasonable attempt to help the adolescent understand the importance of a certain
course of action or choice of behavior. However, when such attempts fail, such parents
still stand strong in their assertion that the adolescent will not be allowed to engage in
some activity or behavior deemed dangerous or inappropriate.
Weaknesses of the Current Study
Like so many other empirical studies, the current study utilized a convenience
sample and, consequently, suffers from lack of random selection and representativeness;
therefore, its findings may not be generalizable to the population of adolescents.
In addition, the data were collected using only one assessment methodology,
namely self-report. The inherent weakness of self-report data is that it represents a single
informant?s responses to a single survey instrument at a single point in time. This may
result in inflated intercorrelations among the measures since it is a basic human tendency
to want to be internally consistent when asked to respond to a number of very similar
items in one sitting and, unfortunately, survey-takers tend to self-correct or self-censor
their answers in order to meet their personal criterion of consistency. Thus, the data in the
186
current study were somewhat tainted by the fact that they were gleaned through self-
report and no other method (i.e., multi-method assessment) was used to corroborate them.
It may also be a matter of concern that these data were collected in the
southeastern portion of the United States, a region that is often called ?The Bible Belt.?
Research on religiosity and deviance has documented effects using data from this region
(e.g., Georgia; Higgins & Albrecht, 1977) when no effects were documented using the
same study methodology in other regions of the country (e.g., California; Burkett &
White, 1974; Hirschi & Stark, 1969). This has led some to hypothesize that religiosity
only deters deviance in regional areas where religiosity is more culturally prominent (the
?religious ecology? theory; see Stark, Kent, & Doyle, 1982). Thus, it is possible that the
current findings are accordingly geographically limited and can not be generalized outside
the southern region of the United States.
Unfortunately, very few studies in the area of religiosity and deviance have been
conducted longitudinally. This is a weakness in the literature as a whole. Consequently,
this lack of longitudinal data prevents researchers from obtaining thorough
documentation that religiosity has any true causal impact on adolescents? behavior. An
extensive probe of the impact of all types of religiosity on all types of behavioral
outcomes, while simultaneously observing family processes and numerous other social
control factors, over time would be an invaluable addition to the field.
The current investigation used race and SES as controls. Consequently, though the
influence of these variables was effectively partialled out, this study has not contributed
knowledge concerning the role these demographic variables may play in the interaction
187
between religiosity, family processes, and deviant behavior among adolescents. It may be
that the observed effects may differ by race or SES. Since variations in demographic
status are often the source of intense developmental scrutiny, the present study fell short
in this regard.
Future Directions for Research
Based on the aforementioned weaknesses, it is a given that future research should
seek to utilize large, randomly-selected, nationally-representative samples from all
geographic regions of the country and longitudinal designs to more effectively study this
area. In addition, future research could be improved in several specific ways. First,
considerable effort should be exerted in the development of a stronger multi-item
measure of religiosity. Such a measure is long overdue. Smith?s (2003b)
multidimensional theoretical conception of the sociological influence of religiosity on
adolescents? behavior may be a good starting point. A multi-item measure could be
developed that would thoroughly test this conceptualization, and doing so should be
viewed as an important next step, since further research in this area utilizing different
measures (i.e., more broadly-construed and/or comprehensive measures) may be able to
document potentially different effects than those found in this study and other previous
work.
Second, though the current findings did much to clarify the role of sex and age in
the relationship between religiosity and family processes in the deterrence of deviance,
further replication and expansion is required for more complete clarification. If such
future investigation uncovers, as the current investigation did, similarity rather than
188
difference in developmental process, this will be an important addition to a growing
literature with similar emphasis. If, however, developmental differences are uncovered, it
will be important to determine who differs in what way and how those differences should
be managed. It is important to note that developmental similarity is no less interesting
than developmental difference. Both are equally consequential for application in
prevention and intervention efforts.
Third, in the spirit of the asset-building approach advocated by the Search
Institute, future research should seek to more thoroughly examine the patterns of
interaction between religiosity and family processes when combined. If, as the current
findings suggest, religiosity and family processes combined can significantly impact the
levels of deviant behavior in adolescents, then not only should the effect size of such a
change be determined, but also the mechanisms that contribute to the most effective
combination of the two should be enumerated.
Fourth, Regnerus? (2003) and Pearce and Haynie?s (2004) intriguing idea that
religiosity may be viewed as a family process should be explored. If, as they propose,
religiosity may be viewed as a fluid behavioral characteristic of a family system that
affects family functioning, it may be interesting to study it as such. In other words, is it
feasible to view religiosity (or ?spirituality?) as a process that is universally important in
the functioning of every family? It would certainly be interesting and beneficial to pursue
the answer to this question.
Finally, future research in this area should extend beyond the borders of the
United States. Comparative social science research utilizing international data sets is
189
challenging and, consequently, less common. However, to explore whether the
relationship between religiosity, family processes, and deviance is invariant cross-
culturally would be a fascinating endeavor.
190
REFERENCES
Acock, A. C., & Bengston, V. L. (1978). On the relative influence of mothers and fathers:
A covariation analysis of political and religious socialization. Journal of Marriage
and the Family, 40, 519-530.
Adams, G. R., Gullotta, T., & Clancy, M. A. (1985). Homeless adolescents: A descriptive
study of similarities and differences between runaways and throwaways.
Adolescence, 20, 715-724.
Adler, F., Mueller, G. O. W., & Laufer, W. S. (1995). Criminology. New York: McGraw-
Hill.
Alain, M. (1989). Do what I say, not what I do: Children's reactions to parents' behavioral
inconsistencies. Perceptual & Motor Skills, 68(1), 99-102.
Albrecht, S. L., Chadwick, B. A., & Alcorn, D. S. (1977). Religiosity and deviance:
Application of an attitude-behavior contingent consistency model. Journal for the
Scientific Study of Religion, 16(3), 263-274.
Allen, D. E., & Sandhu, H. S. (1967). A comparative study of delinquents and non-
delinquents: Family affect, religion, and personal income. Social Forces, 46, 263-
269.
Allport, G. W., & Ross, J. M. (1967). Personal religious orientation and prejudice.
Journal of Personality and Social Psychology, 5, 432-443.
191
Alwin, D. F. (1986). Religion and parental child-rearing orientations: Evidence of a
Catholic-Protestant convergence. American Journal of Sociology, 92, 412-440.
Amato, P. R., & Fowler, F. (2002). Parenting practices, child adjustment, and family
diversity. Journal of Marriage and Family, 64, 703-716.
Anderson, C. A., Hinshaw, S. P., & Simmel, C. (1994). Mother child interactions in
ADHD and comparison boys: Relationships with overt and covert externalizing
behavior. Journal of Abnormal Child Psychology, 22, 247-265.
Argyle, M., & Beit-Hallahmi, B. (1975). The social psychology of religion. London:
Routledge & Kegan Paul.
Ary, D. V., Duncan, T. E., Duncan, S. C., & Hops, H. (1999). Adolescent problem
behavior: The influence of parents and peers. Behaviour Research and Therapy,
37, 217-230.
Bachman, J. G., O?Malley, P. M., Schulenberg, J. E., Johnston, L. D., Bryant, A. L., &
Merline, A. (2002). Religious attendance and importance. In J. G. Bachman, P. M.
O?Malley, J. E. Schulenberg, L. D. Johnston, A. L. Bryant, & A. Merline (Eds.),
The decline of substance use in young adulthood (pp. 40-70). Mahwah, NJ:
Laurence Erlbaum.
Bahr, H. M. (1981). Religious intermarriage and divorce in Utah and the mountain states.
Journal for the Scientific Study of Religion, 20, 251-261.
Bahr, H. M. (1982). Religious contrasts in family role definitions and performance: Utah
Mormons, Catholics, Protestants, and others. Journal for the Scientific Study of
Religion, 21, 200-217.
192
Bahr, H. M., & Chadwick, B. A. (1985). Religion and family in Middletown USA.
Journal of Marriage and the Family, 47, 407-414.
Bahr, S. J., Hawks, R. D., & Wang, G. (1993). Family and religious influences on
adolescent substance abuse. Youth & Society, 24(4), 443-465.
Bahr, S. J., Maughan, S. L., Marcos, A. C., & Li, B. (1998). Family, religiosity, and the
risk of adolescent drug use. Journal of Marriage and the Family, 60, 979-992.
Baier, C. J., & Wright, B. R. E. (2001). ?If you love me, keep my commandments:? A
meta-analysis of the effect of religion on crime. Journal of Research in Crime and
Delinquency, 38(1), 3-21.
Bailey, D. B., Skinner, D., Rodriguez, P., Gut, D., & Correa, V. (1999). Awareness, use,
and satisfaction with services for Latino parents of young children with
disabilities. Exceptional Children, 65, 367-381.
Baker, R. L. A., & Mednick, B. R. (1984). Influences on human development: A
longitudinal perspective. Boston: Kluwer-Nijhoff.
Balk, D. (1983). How teenagers cope with sibling death: Some implications for school
counselors. School Counselor, 31, 150-158.
Barbarin, O. A., & Chesler, M. (1984). Coping as an interpersonal strategy: Families with
childhood cancer. Family Systems Medicine, 2, 279-289.
Barber, B. K. (1992). Family, personality, and adolescent problem behaviors. Journal of
Marriage and the Family, 54, 69-79.
Barber, B. K. (1996). Parental psychological control: Revisiting a neglected construct.
Child Development, 67, 3296-3319.
193
Barber, B. K. (2002). Reintroducing parental psychological control. In B. K. Barber (Ed.),
Intrusive parenting: How psychological control affects children and adolescents
(pp. 3-13). Washington, D.C.: American Psychological Association.
Barber, B. K., Olsen, J. E., & Shagle, S. C. (1994). Associations between parental
psychological and behavioral control and youth internal and externalized
behaviors. Child Development, 65, 1120-1136.
Barber, B. K., & Rollins, B. C. (Eds.).(1990). Parent adolescent relationships. Lanham,
MD: University Press of America.
Barnes, G. M. (1984). Adolescent alcohol abuse and other problem behaviors: Their
relationships and common parental influences. Journal of Youth and Adolescence,
13(4), 329-348.
Barnes, H. E., & Teeters, N. K. (1951). New horizons in criminology. Englewood Cliffs,
NJ: Prentice-Hall.
Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in
social psychological research: Conceptual, strategic, and statistical considerations.
Journal of Personality and Social Psychology, 51(6), 1173-1182.
Barrera, M. Jr., Chassin, L., & Rogosch, F. (1993). Effects of social support and conflict
on adolescent children of alcoholic and nonalcoholic fathers. Journal of
Personality and Social Psychology, 64, 602-612.
Barron, M. R. (1954). The juvenile in delinquent society. New York: Alfred Knopf.
Bartkowski, J. P., & Wilcox, W. B. (2000). Conservative Protestant child discipline: The
case of parental yelling. Social Forces, 79(1), 263-290.
194
Bates, J. E., Bayles, K., Bennett, D., Ridge, B., & Brown, M. (1991). Origins of
externalizing behavior problems at eight years of age. In D. J. Pepler, & K. H.
Rubin (Eds.), The development and treatment of child aggression (pp. 93-120).
Hillsdale, NJ: Erlbaum.
Bates, J. E., Marvinney, D., Kelly, T., Dodge, K. A., Bennett, D. S, & Pettit, G. S. (1994).
Child-care history and kindergarten adjustment. Developmental Psychology,
30(5), 690-700.
Bates, J. E., Pettit, G. S., Dodge, K. A., & Ridge, B. (1998). Family and child factors in
stability and change in children?s aggressiveness in elementary school. In J.
McCord (Ed.), Coercion and punishment in long-term perspectives (pp. 124-138).
New York: Cambridge University Press.
Batson, C. D., Schoenrade, P., & Ventis, W. L. (1993). Religion and the individual: A
social-psychological perspective. New York: Oxford University Press.
Baumrind, D. (1965). Parental control and parental love. Children, 12, 230-234.
Baumrind, D. (1967). Child care practices anteceding 3 patterns of preschool behavior.
Genetic Psychology Monographs, 75, 43-88.
Baumrind, D. (1968). Authoritarian vs. authoritative parental control. Adolescence, 3(11),
255-272.
Baumrind, D. (1971). Current patterns of parental authority. Developmental Psychology
Monograph, 4(1), 1-103.
Baumrind, D. (1978). Parental disciplinary patterns and social competence in children.
Youth and Society, 9, 239-276.
195
Baumrind, D. (1983). Rejoinder to Lewis?s reinterpretation of parental firm control
effects: Are authoritative parents really harmonious? Psychological Bulletin, 94,
132-142.
Baumrind, D. (1987). A developmental perspective on adolescent risk-taking in
contemporary America. New Directions for Child Development, 37, 93-122.
Baumrind, D. (1989). Rearing competent children. In W. Damon (Ed), Child development
today and tomorrow (pp. 349-378). San Francisco, CA: Jossey-Bass.
Baumrind, D. (1991). The influence of parenting style on adolescent competence and
substance use. Journal of Early Adolescence, 11, 56-95.
Baumrind, D. (1996). Parenting: The discipline controversy revisited. Family Relations,
45, 405-414.
Baumrind, D., & Black, A. E. (1967). Socialization practices associated with dimensions
of competence in preschool boys and girls. Child Development, 38, 291-327.
Baumrind, D., Larzelere, R. E., & Cowan, P. A. (2002). Ordinary physical punishment: Is
it harmful? Comment on Gershoff (2002). Psychological Bulletin, 128(4), 580-
589.
Beaudoin, T. (2000). Virtual faith: The irreverent spiritual quest of Generation X. San
Francisco: Jossey-Bass.
Beavers, W. R. (1981). A systems model of family for family therapists. Journal of
Marital & Family Therapy, 7(3), 299-307.
Beavers, W. R. (1989). Beavers Systems Model. In C. N. Ramsey, Jr. (Ed), Family
systems in medicine (pp. 62-74). New York: The Guilford Press.
196
Beavers, W. R., & Hampson, R. B. (1993). Measuring family competence: The Beavers
Systems Model. In F. Walsh (Ed.), Normal family processes (2nd ed.)(pp.
73-103). New York: The Guilford Press.
Becker, W. C. (1964). Consequences of different kinds of parental discipline. In M. L.
Hoffman, & W. W. Hoffman (Eds.), Review of child development research: Vol.
1. (pp. 169-208). New York: Russell Sage Foundation.
Benda, B. B. (1995). The effect of religion on delinquency revisited. Journal of Research
in Crime and Delinquency, 32(4), 446-466.
Benda, B. B., & Corwyn, R. F. (1997). Religion and delinquency: The relationship after
considering family and peer influences. Journal for the Scientific Study of
Religion, 36(1), 81-92.
Benda, B. B., & Corwyn, R. F. (2000). A theoretical model of religiosity and drug use
with reciprocal relationships: A test using structural equation modeling. Journal
of Social Service Research, 26(4), 43-67.
Benda, B. B., & Corwyn, R. F. (2002). The effect of abuse in childhood and in
adolescence on violence among adolescents. Youth & Society, 33(3), 339-365.
Benda, B. B., & Toombs, N. J. (2000). Religiosity and violence: Are they related after
considering the strongest predictors? Journal of Criminal Justice, 28, 483-496.
Benda, B. B., & Whiteside, L. (1995). Testing an integrated model of delinquency using
LISREL. Journal of Social Service Research, 21(2), 1-32.
Benson, P. L., Donahue, M. J., & Erickson, J. A. (1989). Adolescence and religion: A
review of the literature from 1970 to 1986. In M. L. Lynn & D. O. Moberg (Eds.),
197
Research in the social scientific study of religion: A research annual: Vol. 1 (pp.
153-181).Greenwich, CT: JAI Press, Inc.
Benson, P. L., Yeager, R. J., Wood, P. K., Guerra, M. J., & Manno, B. V. (1986).
Catholic high schools: Their impact on low-income students. Washington, DC:
National Catholic Educational Association.
Bentler, P. M., & Bonett, D. G. (1980). Significance tests and goodness of fit in the
analysis of covariance structures. Psychological Bulletin, 88, 588-606.
Blumstein, A., & Cohen, J. (1987). Charactierizing criminal careers. Science, 237, 985-
991.
Blumstein, A., Cohen, J., Roth, J. A., & Visher, C. A. (Eds.) (1986). Criminal careers and
career criminals: Vol. 1. Washington, DC: National Academy Press.
Bock, E. W., & Radelet, M. L. (1988). The marital integration of religious independents:
A re-evaluation of its significance. Review of Religious Research, 29, 228-241.
Bonger, W. A. (1916). Criminality and economic conditions. Boston: Little Brown.
Booth, A., Johnson, D. R., Branaman, A., & Sica, A. (1995). Belief and behavior: Does
religion matter in today?s marriage? Journal of Marriage and the Family, 57, 661-
671.
Breault, K. D., & Kposowa, A. J. (1987). Explaining divorce in the United States: A
study of 3,111 counties, 1980. Journal of Marriage and the Family, 49, 549-558.
Brinkerhoff, M. B., Grandin, E., & Lupri, E. (1992). Religious involvement and spousal
violence ? the Canadian case. Journal for the Scientific Study of Religion, 31, 15-
31.
198
Brody, G. H., & Flor, D. L. (1998). Maternal resources, parenting practices, and child
competence in rural, single-parent African American families. Child
Development, 69(3), 803-816.
Brody, G. H., Stoneman, Z., & Flor, D. (1996). Parental religiosity, family processes, and
youth competence in rural, two-parent African American families. Developmental
Psychology, 32(4), 696-706.
Brody, G. H., Stoneman, Z., Flor, D., & McCrary, C. (1994). Religion?s role in
organizing family relationships: Family processes in rural, two-parent African
American families. Journal of Marriage and the Family, 56, 878-888.
Brook, J. S., Brook, D. W., Gordon, A. S., Whiteman, M., & Cohen, P. (1990). The
psychosocial etiology of adolescent drug use: A family interactional approach.
Psychology Monographs, 116(2), 111-267.
Brownfield, D., & Sorenson, A. M. (1991). Religion and drug use among adolescents: A
social support conceptualization and interpretation. Deviant Behavior, 12, 259-
276.
Burbach, D. J., & Bourdin, C. M. (1986). Parent-child relations and the etiology of
depression: A review of methods and findings. Clinical Psychology Review, 6,
133-153.
Burkett, S. R. (1993). Perceived parents? religiosity, friends? drinking, and hellfire: A
panel study of adolescent drinking. Review of Religious Research, 35, 136-154.
Burkett, S. R., & White, M. (1974). Hellfire and delinquency: Another look. Journal for
the Scientific Study of Religion, 13, 455-462.
199
Buysse, W. H. (1997). Behaviour problems and relationships with family and peers
during adolescence. Journal of Adolescence, 20, 645-659.
Byrne, B. (1994). Structural equation modeling with EQS and EQS/Windows. Thousand
Oaks, CA: Sage.
Call, V. R. A, & Heaton, T. B. (1997). Religious influence on marital stability. Journal
for the Scientific Study of Religion, 36(3), 382-392.
Capaldi, D. M., Stoolmiller, M., Clark, S., & Owen, L. D. (2002). Heterosexual risk
behaviors in at-risk young men from early adolescence to young adulthood:
Prevalence, prediction, and association with STD contraction. Developmental
Psychology, 38(3), 394-406.
Carlucci, K., Genova, J., Rubackin, F., Rubackin, R., & Kayson, T. (1993). Effects of
sex, religion, and amount of alcohol consumption on self-reported
drinking-related problem behaviors. Psychological Reports, 72(3), 983-987.
Carr-Saunders, A. M., Mannheim, H., & Rhodes, E. C. (1944). Young offenders. New
York: The Macmillan Company.
Cernkovich, S. A., & Giordano, P. C. (1987). Family relationships and delinquency.
Criminology, 25, 295-321.
Cernkovich, S. A., & Giordano, P. C. (1992). School bonding, race, and delinquency.
Criminology, 30(2), 261-290.
Chadwick, B. A., & Top, B. L. (1993). Religiosity and delinquency among LDS
adolescents. Journal for the Scientific Study of Religion, 32(1), 51-67.
Chamberlain, P., Ray, J., & Moore, K. (1996). Characteristics of residential care for
200
adolescent offenders: A comparison of assumptions and practices in two models.
Journal of Child and Family Studies, 5, 259-271.
Chan, L.Y., & Heaton, T. B. (1989). Demographic determinants of delayed divorce.
Journal of Divorce, 13, 97-112.
Clydesdale, T. T. (1997). Family behaviors among early US baby boomers: Exploring the
effects of religion and income change, 1965-1982. Social Forces, 76, 605-635.
Cochran, J. K. (1989). Another look at delinquency and religiosity. Sociological
Spectrum, 9, 147-162.
Cochran, J. K. (1993). The variable effects of religiosity and denomination on adolescent
self-reported alcohol use by beverage type. Journal of Drug Issues, 23(3), 479-
482.
Cochran, J. K., & Akers, R. L. (1989). Beyond hellfire: An exploration of the variable
effects of religiosity on adolescent marijuana and alcohol use. Journal of
Research in Crime and Delinquency, 26(3), 198-225.
Cochran, J. K., Wood, P. B., & Arneklev, B. J. (1994). Is the religiosity-delinquency
relationship spurious? A test of arousal and social control theories. Journal of
Research in Crime and Delinquency, 31(1), 92-123.
Collins, W. A. (1988). Research on the transition to adolescence: Continuity in the study
of developmental processes. In M. Gunnar (Ed.), Twenty-first Minnesota
symposium on child psychology (pp. 1-15). Hillsdale, NJ: Lawrence Erlbaum
Associates.
Collins, W. A. (1990). Parent-child relationships in the transition to adolescence:
201
Continuity and change in interaction, affect, and cognition. In R. Montemayor, G.
Adams, and T. Gullotta (Eds.), Advances in adolescent development: Vol. 2. The
transition from childhood to adolescence. Beverly Hills, CA: Sage.
Collins, W. A., & Laursen, B. (1992). Conflict and the transition to adolescence. In C. U.
Shantz & W. W. Hartup (Eds.), Conflict in child and adolescent development (pp.
216-241). Cambridge: Cambridge University Press.
Collins, W. A., Laursen, B., Mortensen, N., Luebker, C., & Ferreira, M. (1997). Conflict
processes and transitions in parent and peer relationships: Implications for
autonomy and regulation. Journal of Adolescent Research, 12, 178-198.
Coogan, J. E. (1952). The myth mind in an engineer?s world. Federal Probation, 16, 26-
30.
Coogan, J. E. (1954). Religion a preventive of delinquency. Federal Probation, 18, 29-
35.
Cooper, C. (1988). Commentary: The role of conflict in adolescent parent relationships.
In M. Gunnar (Ed.), Twenty-first Minnesota symposium on child psychology (pp.
181-187). Hillsdale, NJ: Lawrence Erlbaum Associates.
Coopersmith, S. (1967). The antecedents of self-esteem. San Francisco: W. H. Freeman.
Cortes, J. B. (1965). Juvenile delinquency: A biosocial approach. In A. D?Agostino (Ed.),
Family, church, and community (pp. 114-155). New York: P. J. Kenedy.
Corwyn, R. F., & Benda, B. B. (2000). Religiosity and church attendance: The effects on
use of ?hard drugs? controlling for sociodemographic and theoretical factors. The
International Journal for the Psychology of Religion, 10(4), 241-258.
202
Corwyn, R. F., & Benda, B. B. (2001). Violent youths in southern public schools in
America. International Journal of Adolescence and Youth, 10(1-2), 69-90.
Costanzo, P. R., & Woody, E. Z. (1985). Domain-specific parenting styles and their
impact on the child?s development of particular deviance: The example of obesity
proneness. Journal of Social and Clinical Psychology, 3, 425-445.
Coulthard, P. & Fitzgerald, M. (1999). In God we trust? Organized religion and personal
beliefs as resources and coping strategies and their implications for health in
parents with a child on the autistic spectrum. Mental Health, Religion, and
Culture, 2, 19-33.
Dadds, M. R., Schwartz, S., & Sanders, M. R. (1987). Marital discord and treatment
outcome in behavioral treatment of child conduct disorders. Journal of Consulting
and Clinical Psychology, 55, 396-403.
Danso, H., Hunsberger, B., & Pratt, M. (1997). The role of parental religious
fundamentalism and right-wing authoritarianism in child-rearing goals and
practices. Journal for the Scientific Study of Religion, 36, 496-511.
Davis, K. (1948). Human society. New York: Macmillan.
Davis, K., & Moore, W. E. (1945). Some principles of stratification. American
Sociological Review, 10, 242-249.
Deater-Deckard, K., & Dodge, K. A. (1997). Externalizing behavior problems and
discipline revisited: Nonlinear effects and variation by culture, context, and
gender. Psychological Inquiry, 8(3), 161-175.
Deater-Deckard, K., Dodge, K. A., Bates, J. E., & Pettit, G. S. (1996). Physical discipline
203
among African American and European American mothers: Links to children?s
externalizing behaviors. Developmental Psychology, 32(6), 1065-1072.
Deater-Deckard, K., Dodge, K. A., Bates, J. E., & Pettit, G. S. (1998). Multiple risk
factors in the development of externalizing behavior problems: Group and
individual differences. Development and Psychopathology, 10, 469-493.
Dekovic, M. (1999). Risk and protective factors in the development of problem behavior
during adolescence. Journal of Youth and Adolescence, 28(6), 667-685.
Dekovic, M., Noom, J. J., & Meeus, W. (1997). Expectations regarding development
during adolescence: Parental and adolescent perceptions. Journal of Youth and
Adolescence, 26, 253-272.
Denton, M. L., & Smith, C. (2001). Methodological issues and challenges in the study of
American youth and religion. Retrieved October 14, 2003, from the National
Study of Youth and Religion web site: http://www.youthandreligion.org
/publications/docs/methods.pdf
Dishion, T. J., Patterson, G. R., & Kavanagh, K. A. (1992). An experimental test of the
coercion model: Linking theory, measurement, and intervention. In J. McCord, &
R. Tremblay (Eds.), The interaction of theory and practice: Experimental studies
of intervention (pp. 253-282). New York: Guilford.
Donahue, M. (1995). Religion and the well-being of adolescents. Journal of Social
Issues, 51, 145-160.
Donahue, M. J., & Benson, P. L. (1995). Religion and the well-being of adolescents.
Journal of Social Issues, 51(2), 145-160.
204
Dornbusch, S. M., Carlsmith, J. M., Bushwall, S. J., Ritter, P. L., Leiderman, H., Hastorf,
A. H., & Gross, R. T. (1985). Single parents, extended households, and the
control of adolescents. Child Development, 56, 326-341.
Dornbusch, S. M., Erickson, K. G., Laird, J. & Wong, C. A. (2001). The relation of
family and school attachment to adolescent deviance in diverse groups and
communities. Journal of Adolescent Research, 16(4), 396-422.
Dudley, M. G., & Kosinski, F. A. (1990). Religiosity and marital satisfaction: A research
note. Review of Religious Research, 32, 78-86.
Dudley, R. L., & Dudley, M. G. (1986). Transmission of religious values from parents to
adolescents. Review of Religious Research, 28, 3-15.
Dudley, R. L., Mutch, P. B., & Cruise, R. J. (1987). Religious factors and drug usage
among Seventh-Day Adventist youth in North America. Journal for the Scientific
Study of Religion, 26(2), 218-233.
Duncan, S. C., Duncan, T. E., Biglan, A., & Ary, D. (1998). Contributions of the social
context to the development of adolescent substance use: A multivariate latent
growth modeling approach. Drug and Alcohol Dependence, 50, 57-71.
Durbin, D., Darling, N., Steinberg, L., & Brown, B. (1993). Parenting style and peer
group membership among European-American adolescents. Journal of Research
on Adolescence, 3, 87-100.
Durkheim, E. (1951). Suicide: A study in sociology. Glencoe, IL: Free Press.
Eckenrode, J., Laird, M., & Doris, J. (1993). School performance and disciplinary
problems among abused and neglected children. Developmental Psychology, 29,
205
53-62.
Egeland, B., & Sroufe, L. A. (1981). Developmental sequelae of maltreatment in infancy.
New Directions for Child Development, 11, 77-92.
Eisenstadt, T. H., Eyberg, S., McNeil, C. B., Newcomb, K., & Funderburk, B. (1993).
Parent-child interaction therapy with behavior problem children: Relative
effectiveness of two stages and overall treatment outcome. Journal of Clinical
Child Psychology, 22, 42-51.
Ellickson, P. L., & Morton, S. C. (1999). Identifying adolescents at risk for hard drug use:
Racial/ethnic variations. Journal of Adolescent Health, 25, 382-395.
Elifson, K. W., Petersen, D. M., & Hadaway, C. K. (1983). Religiosity and delinquency:
A contextual analysis. Criminology, 21(4), 505-527.
Elliott, D. S., Huizinga, D. S., & Ageton, S. S. (1985). Explaining delinquency and drug
use. Beverly Hills, CA: Sage.
Elliott, D. S., Huizinga, D. S., & Menard, S. (1989). Multiple problem youth:
Delinquency, substance use, and mental health problems. New York: Springer-
Verlag.
Ellis, H. (1910). The criminal. New York: Walter Scott.
Ellis, L., & Thompson, R. (1989). Relating religion, crime, arousal, and boredom.
Sociology and Social Research, 73(3), 132-139.
Ellison, C. G., Bartkowski, J. P., & Anderson, K. L. (1999). Are there religious variations
in domestic violence? Journal of Family Issues, 20, 87-113.
Ellison, C. G., Bartkowski, J. P., & Segal, M. L. (1996a). Conservative Protestantism and
206
the parental use of corporal punishment. Social Forces, 74, 1003-1028.
Ellison, C. G., Bartkowski, J. P., & Segal, M. L. (1996b). Do conservative Protestants
spank more often? Further evidence from the National Survey of Families and
Households. Social Science Quarterly, 77, 663-673.
Ellison, C. G., & Sherkat, D. E. (1993a). Conservative Protestantism and the corporal
punishment of children: Clarifying the issues. American Sociological Review, 58,
131-144.
Ellison, C. G., & Sherkat, D. E. (1993b). Obedience and autonomy: Religion and parental
values reconsidered. Journal for the Scientific Study of Religion, 32, 313-329.
Engs, R. C., & Hanson, D. J. (1990). Gender differences in drinking patterns and
problems among college students: A review of the literature. Journal of Alcohol &
Drug Education, 35(2), 36-47.
Erickson, J. A. (1992). Adolescent religious development and commitment: A structural
equation model of the role of family, peer group, and educational influences.
Journal for the Scientific Study of Religion, 31, 131-152.
Erickson, M. F., Sroufe, L. A., & Egeland, B. (1985). The relationship between quality of
attachment and behavior problems in preschool in a high-risk sample. In I.
Bretherton & E. Waters (Eds.), Growing points in attachment theory and research
(pp. 147-186). Monographs of the Society for Research in Child Development,
50(1-2, Serial No. 209).
Erikson, E. H. (1959). Identity and the life cycle: Selected papers. Psychological Issues,
1(1), 1-171.
207
Erikson, E. (1968). Identity: Youth and crisis. New York: Norton.
Evans, T. D., Cullen, F. T., Burton, V. S., Dunaway, R. G., Payne, G. L., & Kethineni, S.
R. (1996). Religion, social bonds, and delinquency. Deviant Behavior, 17, 43-70.
Evans, T. D., Cullen, F. T., Dunaway, R. G., & Burton, Jr., V. S. (1995). Religion and
crime reexamined: The impact of religion, secular controls, and social ecology on
adult criminality. Criminology, 33(2), 195-217.
Farrington, D. P. (1986). Age and crime. In M. Tonry & M. Morris (Eds.), Crime and
justice: An annual review of research (Vol. 7, pp. 189-250). Chicago: University
of Chicago Press.
Farrington, D. P. (1989). Early predictors of adolescent aggression and adult violence.
Violence and Victims, 4, 79-100.
Farrington, D. P. (1995). The challenge of teenage antisocial behavior. In M. Rutter (Ed.),
Psychosocial disturbances in young people (pp. 83-130). Cambridge: Cambridge
University Press.
Farrington, D. P., Gundy, G., & West, D. J. (1975). The familial transmission of
criminality. Medicine, Science and the Law, 15, 177-186.
Farrington, D. P., & Hawkins, J. D. (1991). Predicting participation, early onset and later
persistence in officially recorded offending. Criminal Behaviour and Mental
Health, 1, 1-33.
Fergusson, D. M., Horwood, L. J., Kershaw, K. L., & Shannon, F. T. (1986). Factors
associated with reports of wife assault in New Zealand. Journal of Marriage and
the Family, 48, 407-412.
208
Fernquist, R. M. (1995). A research note on the association between religion and
delinquency. Deviant Behavior, 16, 169-175.
Fiese, B. H., & Tomcho, T. J. (2001). Finding meaning in religious practices: The relation
between religious holiday rituals and marital satisfaction. Journal of Family
Psychology, 15(4), 597-609.
Flannery, D. J., Vazsonyi, A. T., Torquati, J. C., & Fridrich, A. (1994). Ethnic and gender
differences in risk for early adolescent substance use. Journal of Youth &
Adolescence, 23(2), 195-213.
Fletcher, A. C., & Jefferies, B. C. (1999). Parental mediators of associations between
perceived authoritative parenting and early adolescent substance use. Journal of
Early Adolescence, 19(4), 465-487.
Flor, D. L., & Knapp, N. F. (2001). Transmission and transaction: Predicting adolescents?
internalization of parental religious values. Journal of Family Psychology, 15(4),
627-645.
Florsheim, P., Tolan, P., & Gorman-Smith, D. (1998). Family relationships, parenting
practices, the availability of male family members, and the behavior of inner-city
boys in single-mother and two-parent families. Child Development, 69(5), 1437-
1447.
Forehand, R., Brody, G. H., Slotkin, J., Fauber, R., McCombs, A., & Long, N. (1988).
Young adolescent and maternal depression: Assessment, interrelations, and family
predictors. Journal of Consulting and Clinical Psychology, 56(3), 422-426.
Forehand, R., Long, N., Brody, G. H., & Fauber, R. (1986). Home predictors of young
209
adolescents? school behavior and academic performance. Child Development, 57,
1528-1533.
Forehand, R., Miller, K. S., Dutra, R., & Watts-Chance, M. (1997). Role of parenting in
adolescent deviant behavior: Replication across and within two ethnic groups.
Journal of Consulting and Clinical Psychology, 65(6), 1036-1041.
Forgatch, M. S. (1991). The clinical science vortex: A developing theory of antisocial
behavior. In D. J. Pepler, & K. H. Rubin (Eds.), The development and treatment of
childhood aggression (pp. 291-315). Hillsdale, NJ: Erlbaum.
Foshee, V. A., & Hollinger, B. R. (1996). Maternal religiosity, adolescent social bonding,
and adolescent alcohol use. Journal of Early Adolescence, 16(4), 451-468.
Foster, S. L., & Robin, A. L. (1988). Family conflict and communication in adolescence.
In E. J. Mash & L. G. Terdal (Eds.), Behavioral assessment of childhood
disorders (pp. 717-775). New York: The Guilford Press.
Francis, L. J. (1985). Personality and religion: Theory and measurement. In L. B. Brown
(Ed.), International series in experimental social psychology, Vol. 11: Advances
in the psychology of religion (pp. 171-185). Pergamon: Oxford.
Free, M. D. (1991). Clarifying the relationship between the broken home and juvenile
delinquency: A critique of the current literature. Deviant Behavior, 12, 109-167.
Free, M. D. (1994). Religiosity, religious conservatism, bonds to school, and juvenile
delinquency among three categories of drug users. Deviant Behavior, 15, 151-170.
Freud, S. (1962). Three essays on the theory of sexuality (J. Strachey, Trans.). New York:
Basic Books. (Original work published 1905)
210
Friedrich, W. N., Cohen, D. S., & Wilturner, L. T. (1988). Specific beliefs as moderator
variables in maternal coping with mental retardation. Journal of Children?s
Health Care, 17, 40-44.
Galambos, N. L., & Maggs, J. L. (1991). Out-of-school care of young adolescents and
self-reported behavior. Developmental Psychology, 27, 644-655.
Gannon, T. M. (1967a). Religious attitude and behavior changes of institutional
delinquents. Sociological Analysis, 28, 215-225.
Gannon, T. M. (1967b). Religious control and delinquent behavior. Sociology and Social
Research, 51, 418-431.
Geismar, L. L., & Wood, K. (1986). Family and delinquency: Resocializing the young
offender. New York: Human Sciences Press.
George, C., & Main, M. (1979). Social interaction of young abused children. Child
Development, 50, 306-318.
Gershoff, E. T. (2002). Corporal punishment by parents and associated child behaviors
and experiences: A meta-analytic and theoretical review. Psychological Bulletin,
128(4), 539-579.
Gershoff, E. T., Miller, P. C., & Holden, G. W. (1999). Parenting influences from the
pulpit: Religious affiliation as a determinant of corporal punishment. Journal of
Family Psychology, 13, 307-320.
Gilligan, C. (1987). Adolescent development reconsidered. In C. E. Irwin, Jr. (Ed.), New
directions for child development, Vol. 37 (pp. 63-92). San Francisco: Jossey-Bass.
Glueck, S., & Glueck, E. (1950). Unraveling juvenile delinquency. Cambridge, MA:
211
Harvard University Press.
Gorman-Smith, D., Tolan, P. H., Henry, D. B. (2000). A developmental-ecological model
of the relation between family functioning to patterns of delinquency. Journal of
Quantitative Criminology, 16(2), 169-198.
Gorsuch, R. L., & McFarland, S. G. (1972). Single vs. multiple-item scales for measuring
religious values. Journal for the Scientific Study of Religion, 11, 53-64.
Gove, W. R., & Crutchfield, R. D. (1982). The family and juvenile delinquency.
Sociological Quarterly, 23, 301-319.
Grasmick, H. G., Bursik, R., & Kimpel, M. (1991). Protestant fundamentalism and
attitudes toward the corporal punishment of children. Violence and Victims, 6,
283-297.
Grasmick, H. G., Morgan, C. S., & Kennedy, M. B. (1992). Support for corporal
punishment in schools: A comparison of the effects of socioeconomic status and
religion. Social Science Quarterly, 73, 177-187.
Gray, E. (1988). The link between child abuse and juvenile delinquency: What we know
and recommendations for policy and research. In G. Hotaling, D. Finkelhor, J. T.
Kirkpatrick, & M. A. Straus (Eds.), Family abuse and its consequences (pp. 109-
123). Beverly Hills, CA: Sage.
Gray, M. R., & Steinberg, L. (1999). Unpacking authoritative parenting: Reassessing a
multidimensional construct. Journal of Marriage and the Family, 61, 574-587.
Gray-Ray, P., & Ray, M. C. (1990). Juvenile delinquency in the Black community. Youth
& Society, 22(1), 67-84.
212
Greenberg, M. T., Speltz, M. L., & DeKlyen, M. (1993). The role of attachment in the
early development of disruptive behavior problems. Development and
Psychopathology, 5, 191-213.
Greenberg, M. T., Speltz, M. L., DeKlyen, M., & Endriga, M. C. (1991). Attachment
security in preschoolers with and without externalizing problems: A replication.
Development and Psychopathology, 3, 413-430.
Grusec, J., & Lytton, H. (1988). Social development: History, theory, and research. New
York: Springer Verlag.
Gunnoe, J. L., Hetherington, E. M., & Reiss, D. (1999). Parental religiosity, parenting
style, and adolescent social responsibility. Journal of Early Adolescence, 19, 199-
225.
Hall, G. S. (1904). Adolescence. New York: Appleton.
Hall, J. A. (1987). Parent-adolescent conflict: An empirical review. Adolescence, 22, 767-
789.
Hatch, R. C., James, D. E., & Schumm, W. R. (1986). Spiritual intimacy and marital
satisfaction. Family Relations, 35, 539-545.
Hauser, S. (1991). Families and their adolescents. New York: Free Press.
Hauser, S., Powers, S. I., Noam, G., Jacobson, A., Weiss, B., & Follansbee, D. (1984).
Familial contexts of adolescent ego development. Child Development, 55, 195-
213.
Haworth, A. M., Hill, A. E., & Glidden, L. M. (1996). Measuring religiousness of parents
of children with developmental disabilities. Mental Retardation, 34, 271-279.
213
Hayduk, L. A. (1987). Structural equation modeling with latent variables. New York:
John Wiley & Sons.
Hayes, B. C., & Pittlekow, Y. (1993). Religious belief, transmission, and the family: An
Australian study. Journal of Marriage and the Family, 55, 755-766.
Heath, A. C., Madden, P. A., Grant, J. D., McLaughlin, T. L., Todorov, A. A., &
Bucholz, K. K. (1999). Resiliency factors protecting against teenage alcohol use
and smoking: Influences of religion, religious involvement and values, and
ethnicity in the Missouri Adolescent Female Twin Study. Twin Research, 2, 145-
155.
Heaton, T. B., & Goodman, K. L. (1985). Religion and family formation. Review of
Religious Research, 26, 343-359.
Heaton, T. B., & Pratt, E. L. (1990). The effects of religious homogamy on marital
satisfaction and stability. Journal of Family Issues, 11, 191-207.
Henggeler, S. W. (1989). Delinquency in adolescence. Newbury Park, CA: Sage.
Hetherington, E. M., Cox, M., & Cox, R. (1979). Play and social interaction in children
following divorce. Journal of Social Issues, 35, 26-49.
Hickman, M., & Piquero, A. (2001). Exploring the relationships between gender, control
balance, and deviance. Deviant Behavior, 22, 323-351.
Higgins, P. C., & Albrecht, G. L. (1977). Hellfire and delinquency revisited. Social
Forces, 55, 952-958.
Hill, J. P. (1988). Adapting to menarche: Familial control and conflict. In M. R. Gunnar
& W. A. Collins (Eds.), The Minnesota Symposia on Child Psychology: Vol. 21.
214
Development during the transition to adolescence (pp. 43-77). Hillsdale, NJ:
Erlbaum.
Hill, P. C., & Hood, R. W., Jr. (Eds.). (1999). Measures of religiosity. Birmingham, AL:
Religious Education Press.
Hirschi, T. (1969). Causes of delinquency. Berkeley: University of California Press.
Hirschi, T., & Gottfredson, M. (1983). Age and the explanation of crime. American
Journal of Sociology, 89, 552-584.
Hirschi, T., & Stark, R. (1969). Hellfire and delinquency. Social Problems, 17, 202-213.
Hoffman-Plotkin, D., & Twentyman, C. (1984). A multimodal assessment of behavioral
and cognitive deficits in abused and neglected preschoolers. Child Development,
55, 784-802.
Hoge, D. R., Petrillo, G. H., & Smith, E. I. (1982). Transmission of religious and social
values from parents to teenage children. Journal of Marriage and the Family, 44,
569-580.
Hollingshead, A. B. (1975). Four-factor index of social status. Yale University
Department of Sociology, New Haven, CT.
Holmbeck, G. N. (1997). Toward terminological, conceptual, and statistical clarity in the
study of mediators and moderators: Examples from the child-clinical and pediatric
psychology literatures. Journal of Consulting and Clinical Psychology, 65(4),
599-610.
Howes, C. (1988). Abused and neglected children with their peers. In G. T. Hotaling, D.
Finkelhor, J. T. Kirkpatrick, & M. A. Straus (Eds.), Family abuse and its
215
consequences (pp. 99-108). Beverly Hills, CA: Sage.
Howes, C., & Eldredge, R. (1985). Responses of abused, neglected, and non-maltreated
children to the behaviors of peers. Journal of Applied Developmental Psychology,
6, 261-270.
Hoyle, R. H. (Ed.)(1995). Structural equation modeling: Concepts, issues, and
applications. Thousand Oaks, CA: Sage.
Hoyle, R. H., & Smith, G. T. (1994). Formulating clinical research hypotheses as
structural equation models: A conceptual overview. Journal of Consulting and
Clinical Psychology, 62(3), 429-440.
Huesmann, L. R., Eron, L. D., Lefkowitz, M. M., & Walder, L. O. (1984). Stability of
aggression over time and generations. Developmental Psychology, 20(6),
1120-1134.
Huizinga, D., Menard, S., & Elliott, D. (1989). Delinquency and drug use: Temporal and
developmental patterns. Justice Quarterly, 6, 419-455.
Hurrelman, J. (1990). Parents, peers, teachers, and other significant partners in
adolescence. International Journal of Adolescence and Youth, 2, 211-236.
Huselid, R. F., & Cooper, M. L. (1994). Gender roles as mediators of sex differences in
expressions of psychopathology. Journal of Abnormal Psychology, 103, 595-603.
Jensen, G. F., & Erickson, M. L. (1979). The religious factor and delinquency: Another
look at the hellfire hypotheses. In R. Wuthnow (Ed.), The religious dimension:
New directions in quantitative research (pp. 157-177). New York: Academic
Press.
216
Jessor, R., Turbin, M., & Costa, F. (1998). Risk and protection in successful outcomes
among disadvantaged adolescents. Applied Developmental Science, 2, 194-208.
Johnson, B. R., De Li, S., Larson, D. B., & McCullough, M. (2000). A systematic review
of the religiosity and delinquency literature. Journal of Contemporary Criminal
Justice, 16(1), 32-52.
Johnson, B. R., Jang, S. J., Larson, D. B., & De Li, S. (2001). Does adolescent religious
commitment matter? A reexamination of the effects of religiosity on delinquency.
Journal of Research in Crime and Delinquency, 38(1), 22-43.
Johnson, V., & Pandina, R. J. (1991). Effects of the family environment on adolescent
substance use, delinquency, and coping styles. American Journal on Alcohol
Abuse, 17, 71-88.
Judd, C. M., & Kenny, D. A. (1981). Process analysis: Estimating mediation in treatment
evaluations. Evaluation Review, 5, 602-619.
Justice, B., & Duncan, D. F. (1976). Running away: An epidemic problem of
adolescence. Adolescence, 22, 365-371.
Kalmer, L., & Weir, E. (1936). Crime and religion. Chicago: Franciscan Herald Press.
Kandel, D. B., Kessler, R., & Margulies, R. (1978). Antecedents of adolescents? initiation
into stages of drug use: A developmental analysis. In D. B. Kandel (Ed.),
Longitudinal research in drug use: Empirical findings and methodological issues
(pp. 73-79). Washington, D.C.: Hemisphere.
Kazdin, A. E. (1987). Conduct disorders in childhood and adolescence. Thousand Oaks,
CA: Sage.
217
Keiley, M. K., Howe, T. R., Dodge, K. A., Bates, J. E., & Pettit, G. S. (2001). The timing
of child physical maltreatment: A cross-domain growth analysis of impact on
adolescent externalizing and internalizing problems. Development &
Psychopathology, 13(4), 891-912.
Kellam, S. G., Brown, C. H., Rubin, R. R., & Ensminger, M. E. (1983). Paths leading to
teenage psychiatric symptoms and substance use: Developmental epidemiological
studies in Woodlawn. In S. B. Buze, F. J. Earls, & J. E. Barrett (Eds.), Childhood
psychopathology and development (pp. 17-51). New York: Raven Press.
Kenny, D. A., Kashy, D. A., & Bolger, N. (1998). Data analysis in social psychology. In
D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The handbook of social
psychology (pp. 233-265). Boston: McGraw-Hill.
Kerr, M., & Stattin, H. (2000). What parents know, how they know it, and several forms
of adolescent adjustment: Further support for a reinterpretation of monitoring.
Developmental Psychology, 36(3), 366-380.
Kim, J. E., Hetherington, E. M., & Reiss, D. (1999). Associations among family
relationships, antisocial peers, and adolescents? externalizing behaviors: Gender
and family-type differences. Child Development, 70(5), 1209-1230.
Knight, D. K., Broome, K. M., Cross, D. R., & Simpson, D. D. (1998). Antisocial
tendency among drug-addicted adults: Potential long-term effects of parental
absence, support, and conflict during childhood. American Journal of Drug and
Alcohol Abuse, 24(3), 361-375.
Knudten, R. D., & Knudten, M. S. (1971). Juvenile delinquency, crime, and religion.
218
Review of Religious Research, 17(3), 130-152.
Kochanska, G. (1993). Toward a synthesis of parental socialization and child
temperament in early development of conscience. Child Development, 64, 325-
347.
Kochanska, G. (1995). Children?s temperament, mothers? discipline, and security of
attachment: Multiple pathways to emerging internalization. Child Development,
66, 597-615.
Kochanska, G. (1997). Multiple pathways to conscience for children with different
temperaments: From toddlerhood to age 5. Developmental Psychology, 33, 228-
240.
Kohlberg, L. (1969). Stage and sequence: The cognitive-developmental approach to
socialization. In D. A. Goslin (Ed.), Handbook of socialization theory and
research (pp. 347-480). Chicago: Rand McNally.
Kvaraceus, W. C. (1944). Delinquent behavior and church attendance. Sociology and
Social Research, 28, 284-289.
Lamborn, S., Mounts, N., Steinberg, L., & Dornbusch, S. (1991). Patterns of competence
and adjustment among adolescents from authoritative, authoritarian, indulgent,
and neglectful families. Child Development, 62, 1049-1065.
Lammers, C., Ireland, M., Resnick, M., & Blum, R. (2000). Influences on adolescents?
decision to postpone onset of sexual intercourse: A survival analysis of virginity
among youths aged 13 to 18 years. Journal of Adolescent Health, 26, 42-48.
Larson, L. E., & Goltz, J. W. (1989). Religious participation and marital commitment.
219
Review of Religious Research, 30, 387-400.
Lauritsen, J. L. (1993). Sibling resemblance in juvenile delinquency: Findings from the
National Youth Survey. Criminology, 31, 387-409.
Laursen, B., & Collins, W. (1994). Interpersonal conflict during adolescence.
Psychological Bulletin, 115, 197-209.
Laursen, B., Coy, K. C., & Collins, W. A. (1998). Reconsidering changes in parent-child
conflict across adolescence: A meta-analysis. Child Development, 69(3), 817-832.
Leas, L., & Mellor, D. (2000). Prediction of delinquency: The role of depression, risk-
taking, and parental attachment. Behaviour Change, 17(3), 155-166.
Lehrer, E. L., & Chadwick, C. U. (1993). Religion as a determinant of marital stability.
Demography, 30, 385-403.
Lerner, J. V., Hertzog, C., Hooker, K. A., Hassibi, M., & Thomas, L. (1988). A
longitudinal study of negative emotional states and adjustment from early
childhood through adolescence. Child Development, 59(2), 356-366.
Linden, R., & Currie, R. (1977). Religiosity and drug use: A test of social control theory.
Canadian Journal of Criminology and Corrections, 19, 346-355.
Lipsey, M. W., & Derson, J. H. (1998). Predictors of violent or serious delinquency in
adolescence and early adulthood: A synthesis of longitudinal research. In R.
Loeber, & D. P. Farrington (Eds.). Serious & violent juvenile offenders: Risk
factors and successful interventions (pp. 86-105). Thousand Oaks, CA: Sage
Publications.
Litchfield, A. W., Thomas, D. L., & Li, B. D. (1997). Dimensions of religiosity as
220
mediators of the relations between parenting and adolescent deviant behavior.
Journal of Adolescent Research, 12(2), 199-226.
Loeber, R. (1982). The stability of antisocial and delinquent child behavior: A review.
Child Development, 53, 1431-1446.
Loeber, R., & Dishion, T. (1983). Early predictors of male delinquency: A review.
Psychological Bulletin, 94, 68-99.
Loeber, R., & Dishion, T. J. (1984). Boys who fight at home and school: Family
conditions influencing cross-setting consistency. Journal of Consulting and
Clinical Psychology, 42, 471-481.
Loeber, R., & Stouthamer-Loeber, M. (1986). Family factors as correlates and predictors
of juvenile conduct problems and delinquency. In M. Tonry, & N. Morris (Eds.),
Crime and justice: An annual review of research, Vol. 7 (pp. 219-339). Chicago:
University of Chicago Press.
Loehlin, J. (1992). Latent variable models: An introduction to factor, path and structural
analysis. Hillsdale, NJ: Lawrence Erlbaum.
Lombroso, C. (1918). Crime: Its causes and remedies. Boston: Little Brown.
Lunden, W. A. (1964). Statistics on delinquents and delinquency. Springfield, IL: Charles
C. Thomas.
Lynch, C. O. (2001). Risk and protective factors associated with adolescent sexual
activity. Adolescent and Family Health, 2(3), 99-107.
Maccoby, E. E. (1986). Social groupings in childhood: Their relationship to prosocial and
antisocial behavior in boys and girls. In D. Olweus, J. Block, & M. Radke-Yarrow
221
(Eds.), Development of antisocial and prosocial behavior: Research, theories, and
issues (pp. 263-284). New York: Academic Press.
Maccoby, E. E., & Martin, J. A. (1983). Socialization in the context of the family: Parent-
child interaction. In P. H. Mussen (Series Ed.) & E. M. Hetherington (Vol. Ed.),
Handbook of Child Psychology: Vol. 4. Socialization and Social Development (4th
ed., pp. 1-101). New York: John Wiley & Sons.
Maggs, J. L., & Galambos, N. L. (1993). Alternative structural models for understanding
adolescent problem behavior in two-earner families. Journal of Early
Adolescence, 13, 79-101.
Mahoney, A., Pargament, K. I., Jewell, T., Swank, A. B., Scott, E., Emery, E., & Rye, M.
(1999). Marriage and the spiritual realm: The role of proximal and distal religious
constructs in marital functioning. Journal of Family Psychology, 13, 1-18.
Mahoney, A., Pargament, K. I., Tarakeshwar, N., & Swank, A. B. (2001). Religion in the
home in the 1980s and 1990s: A meta-analytic review and conceptual analysis of
links between religion, marriage, and parenting. Journal of Family Psychology,
15(4), 559-596.
Marcia, J. E. (1980). Identity in adolescence. In J. Addison (Ed.), Handbook of adolescent
psychology (pp. 159-187). New York: Wiley.
Marcos, A. C., Bahr, S. J., & Johnson, R. C. (1986). Test of a bonding/association theory
of adolescent drug use. Social Forces, 65, 135-161.
Marsh, H. W., Balla, J. R., & McDonald, R. P. (1988). Goodness-of-fit indexes in
confirmatory factor analysis: The effect of sample size. Psychological Bulletin,
222
103(3), 391-410.
Marsh, H. W., & Hocevar, D. (1985). The application of confirmatory factor analysis to
the study of self-concept: First and higher order factor structures and their
invariance across age groups. Psychological Bulletin, 97, 562-582.
McCord, W., & McCord, J. (1959). Origins of crime: A new evaluation of the
Cambridge-Somerville study. New York: Columbia University Press.
McDonald, R. M., & Towberman, D. B. (1993). Psychosocial correlates of adolescent
drug involvement. Adolescence, 28, 925-936.
McIntosh, D., & Spilka, B. (1990). Religion and physical health: The role of personal
faith and control beliefs. In M. L. Lynn, & D. O. Moberg (Eds.), Research in the
social scientific study of religion, Vol. 2 (pp. 167-194). Greenwich, CT: JAI Press.
McLuckie, B. F., Zahn, M., & Wilson, R. A. (1975). Religious correlates of teenage drug
use. Journal of Drug Issues, 5, 129-139.
Middleton, W. C., & Fay, P. J. (1941). Attitudes of delinquent and non-delinquent girls
toward Sunday observance, the Bible, and war. Journal of Educational
Psychology, 32, 555-558.
Middleton, R., & Putney, S. (1962). Religion, normative standards, and behavior.
Sociometry, 25, 141-152.
Miller, L., Warner, V., Wickramaratne, P., & Weissman, M. (1997). Religiosity and
depression: Ten-year follow-up of depressed mothers and offspring. Journal of the
American Academy of Child and Adolescent Psychiatry, 36, 1416-1425.
Miller, M. E. (1965). The place of religion in the lives of juvenile offenders. Federal
223
Probation, 29, 50-54.
Miller, N. Cowan, P. A., Cowan, C. P., Hetherington, E. M., & Clingempeel, W. G.
(1993). Externalizing in preschoolers and early adolescents: A cross-study
replication of a family model. Developmental Psychology, 29, 3-18.
Moffitt, T. E. (1983). Adolescence-limited and life-course-persistent antisocial behavior:
A developmental taxonomy. Psychological Review, 100, 674-701.
Moffitt, T. E. (1987). Parental mental disorder and offspring criminal behavior: An
adoption study. Psychiatry, 50, 346-360.
Montemayor, R. (1983). Parents and adolescents in conflict: All families some of the time
and some families most of the time. Journal of Early Adolescence, 3, 83-103.
Montemayor, R., & Hanson, E. (1985). A naturalistic view of conflict between
adolescents and their parents and siblings. Journal of Early Adolescence, 5, 23-30.
Mounts, N. S. (2002). Parental management of adolescent peer relationships in context:
The role of parenting style. Journal of Family Psychology, 16(1), 58-69.
Murry, V. (1994). Black adolescent females: A comparison of early versus late coital
initiators. Family Relations, 43, 342-348.
Nagin, D. S., & Farrington, D. P. (1992). The onset and persistence of offending.
Criminology, 30, 501-523.
Nye, F. I. (1958). Family relationships and delinquent behavior. New York: John Wiley.
Olson, D. H. (1986). Circumplex Model: VII. Validation studies and FACES III. Family
Processes, 25(3), 337-351.
Olson, D. H. (2000). Circumplex Model of marital and family systems. Journal of Family
224
Therapy, 22(2), 144-167.
Olson, D. H., McCubbin, H. I., Barnes, H., Larsen, A., Muxen, M., & Wilson, M. (1983).
Families: What makes them work? Beverly Hills, CA: Sage.
Olson, D. H., Russell, C. S., & Sprenkle, D. H. (1983). Circumplex Model of marital and
family systems: VI. Theoretical update. Family Processes, 22(1), 69-83.
Olson, D. H., Sprenkle, D. H., & Russell, C. S. (1979). Circumplex Model of marital and
family systems: I. Cohesion and adaptability dimensions, family types, and
clinical applications. Family Processes, 18(1), 3-28.
Olweus, D. (1979). Stability and aggressive reaction patterns in males: A review.
Psychological Bulletin, 86, 852-875.
Olweus, D. (1980). Familial and temperamental determinants of aggressive behavior in
adolescent boys: A causal analysis. Developmental Psychology, 16(6), 644-660.
Osgood, D. W., Johnston, L. D., O'Malley, P. M., & Bachman, J. G. (1988). The
generality of deviance in late adolescence and early adulthood. American
Sociological Review, 53(1), 81-93.
Palmer, C. E., & Noble, D. N. (1986). Premature death: Dilemmas of infant mortality.
Social Casework, 67, 332-339.
Pargament, K. I. (1990). God help me: Towards a theoretical framework for the
psychology of religion. In M. L. Lynn, & D. O. Moberg (Eds.), Research in the
social scientific study of religion, Vol. 2 (pp. 195-224). Greenwich, CT: JAI Press.
Parker, C. P., Baltes, B. B., & Christiansen, N. D. (1997). Support for affirmative action,
justice perceptions, and work attitudes: A study of gender and racial-ethnic group
225
differences. Journal of Applied Psychology, 82(3), 376-389.
Paternoster, R., & Bachman, R. (2001). Explaining criminals and crime. Los Angeles:
Roxbury.
Patterson, G. R. (1982). Coercive family processes. Eugene, OR: Castalia.
Patterson, G. R. (1995). Coercion as a basis for early age of onset for arrest. In J. McCord
(Ed.), Coercion and punishment in long-term perspectives (pp. 81-105). New
York: Cambridge University Press.
Patterson, G. R., Chamberlain, P., & Reid, J. B. (1982). A comparative evaluation of
parent training procedures. Behavior Therapy, 13, 638-651.
Patterson, G. R., & Dishion, T. J. (1985). Contribution of families and peers to
delinquency. Criminology, 23, 63-79.
Patterson, G. R., Dishion, T. J., & Bank, L. (1984). Family interaction: A processes
model of deviancy training. Aggressive Behavior, 10(3), 253-267.
Patterson, G. R., Dishion, T. J., & Chamberlain, P. (1993). Outcomes and methodological
issues relating to treatment of antisocial children. In T. R. Giles (Ed.), Handbook
of effective psychotherapy (pp. 43-88). New York: Plenum.
Patterson, G. R., Forgatch, M. S., Yoerger, K. L., & Stoolmiller, M. (1998). Variables
that initiate and maintain an early-onset trajectory for juvenile offending.
Development and Psychopathology, 10, 531-547.
Patterson, G. R., Reid, J. B., & Dishion, T. (1992). Antisocial boys: A social interaction
approach. Eugene, OR: Castalia.
Patterson, G. R., & Stouthamer-Loeber, M. (1984). The correlation of family
226
management practices and delinquency. Child Development, 55, 1299-1307.
Pearce, L. D., & Axinn, W. G. (1998). The impact of family religious life on the quality
of mother-child relations. American Sociological Review, 63, 810-828.
Pearce, L. D., & Haynie, D. L. (2004). Intergenerational religious dynamics and
adolescent delinquency. Social Forces, 82(4), 1553-1572.
Pearlman, S., Philip, A. F., Robbins, L. C., Robbins, E. S., Robinson, E. E., & Schmitter,
B. (1972). Religious affiliations and patterns of drug usage in an urban university
population. In S. Einstein & S. Allen (Eds.), Student drug surveys (pp. 139-186).
Farmingdale: Baywood.
Penning, M., & Barnes, G. E. (1982). Adolescent marijuana use: A review. International
Journal of Addictions, 17, 749-791.
Perkins, H. W. (1987). Parental religion and alcohol use problems as intergenerational
predictors of problem drinking among college youth. Journal for the Scientific
Study of Religion, 26(3), 340-357.
Pettit, G. S. (1997). The developmental course of violence and aggression: Mechanisms
of family and peer influence. The Psychiatric Clinics of North America, 20(2),
283-299.
Pettit, G. S., & Bates, J. E. (1989). Family interaction patterns and children?s behavior
problems from infancy to 4 years. Developmental Psychology, 25, 413-420.
Pettit, G. S., Bates, J. E., & Dodge, K. (1993). Family interaction patterns and children?s
conduct problems at home and school: A longitudinal perspective. School
Psychology Review, 22, 401-418.
227
Pettit, G. S., Bates, J. E., Dodge, K. A., & Meece, D. (1999). The impact of after-school
peer contact on early adolescent externalizing problems is moderated by parental
monitoring, perceived neighborhood safety, and prior adjustment. Child
Development, 70, 768-778.
Pettit, G. S., & Dodge, K. A. (1993). Family interaction patterns and children?s conduct
problems at home and school: A longitudinal perspective. School Psychology
Review, 22(3), 403-420.
Pettit, G. S., Harrist, A. W., Bates, J. E., & Dodge, K. A. (1991). Family interaction,
social cognition, and children?s subsequent relations with peers at kindergarten.
Journal of Social and Personal Relationships, 8, 383-402.
Pettit, G. S., Laird, R. D., Bates, J. E., & Dodge, K. A. (1997). Patterns of after-school
care in middle childhood: Risk factors and developmental outcomes. Merrill-
Palmer Quarterly, 43, 515-538.
Pettit, G. S., Laird, R. D., Dodge, K. A., Bates, J. E., & Criss, M. M. (2001). Antecedents
and behavior-problem outcomes of parental monitoring and psychological control
in early adolescence. Child Development, 72(2), 583-598.
Piaget, J. (1965). Moral judgment of the child. New York: Free Press.
Pickering, L. E., & Vazsonyi, A. T. (2002). The impact of adolescent employment on
family relationships. Journal of Adolescent Research, 17(2), 196-218.
Pittman, L. D., & Chase-Lansdale, P. L. (2001). African American adolescent girls in
impoverished communities: Parenting style and adolescent outcomes. Journal of
Research on Adolescence, 11(2), 199-224.
228
Porter, B., & O?Leary, K. D. (1980). Marital discord and childhood problems. Journal of
Abnormal Child Psychology, 8, 287-295.
Posner, J. K., & Vandell, D. L. (1994). Low-income children?s after-school care: Are
there beneficial effects of after-school programs? Child Development, 65, 440-
456.
Potvin, R. H., Hoge, D. R., & Nelson, H. M. (1976). Religion and American youth: With
emphasis on Catholic adolescents and young adults. Washington, DC: United
States Catholic Conference.
Rabey, S. (2001). In search of authentic faith: How emerging generations are
transforming the church. Colorado Springs: Waterbrook Press.
Rankin, J. H. (1983). The family context of delinquency. Social Problems, 30, 466-479.
Rebellon, C. J. (2002). Reconsidering the broken homes/delinquency relationship and
exploring its mediating mechanisms. Criminology, 40(1), 103-136.
Reckless, W. C., & Smith, N. (1932). Juvenile delinquency. New York: McGraw-Hill.
Regnerus, M. D. (2003). Linked lives, faith, and behavior: Intergenerational religious
influence on adolescent delinquency. Journal for the Scientific Study of Religion,
42(2), 189-203.
Reich, W., Earls, F., & Powell, J. (1988). A comparison of the home and social
environments of children of alcoholic and non-alcoholic parents. British Journal
of Addiction, 83, 831-839.
Richardson, J. L., Dwyer, K., McGuigan, K., Hansen, W. B., Johnson, C. A., Sussman,
S., Brannon, B., & Flay, B. (1989). Substance use among eighth-grade students
229
who care for themselves after school. Pediatrics, 84, 556-566.
Riley, D., & Shaw, M. (1985). Parental supervision and juvenile delinquency. Home
Office Research Study No. 83, London: HMSO.
Roberts, M. W., & Powers, S. W. (1990). Adjusting chair timeout enforcement
procedures for oppositional children. Behavior Therapy, 21, 257-271.
Robins, L. N. (1978). Sturdy childhood predictors of adult antisocial behaviour:
Replications from longitudinal studies. Psychological Medicine, 8(4), 611-622.
Roehlkepartain, E. C., & Benson, P. L. (1993). Youth in Protestant churches: A special
Search Institute report. Minneapolis, MN: Search Institute.
Rogers-Dulan, J. (1998). Religious connectedness among urban African American
families who have a child with disabilities. Mental Retardation, 36, 91-103.
Rohrbaugh, J., & Jessor, R. (1975). Religiosity in youth: A personal control against
deviant behavior. Journal of Personality, 43, 136-155.
Ross, L. E. (1994). Religion and deviance: Exploring the impact of social control
elements. Sociological Spectrum, 14, 65-86.
Rossi, A. S., & Rossi, P. H. (1990). Of human bonding: Parent-child relations across the
life course. New York: Aldine de Gruyter.
Rothbaum, F., & Weisz, J. R. (1994). Parental caregiving and child externalizing
behavior in nonclinical samples: A meta-analysis. Psychological Bulletin, 116(1),
55-74.
Rubenstein, J. L., & Feldman, S. S. (1993). Conflict-resolution behavior in adolescent
boys: Antecedents and adaptational correlates. Journal of Research on
230
Adolescence, 3, 41-66.
Ruiz, S. Y., Roosa, M. W., & Gonzales, N. A. (2002). Predictors of self-esteem for
Mexican American and European American youths: A reexamination of the
influence of parenting. Journal of Family Psychology, 16(1), 70-80.
Rutter, M., & Giller, H. (1984). Juvenile delinquency: Trends and perspectives. New
York: Guilford Press.
Sampson, E. E. (1988). The debate on individualism: Indigenous psychologies of the
individual and their role in personal and societal functioning. American
Psychologist, 43, 15-22.
Sampson, R. J., & Laub, J. H. (1993). Crime in the making: Pathways and turning points
through life. Cambridge, MA: Harvard University Press.
Scanzoni, J., & Arnett, C. (1987). Enlarging the understanding of marital commitment via
religious devoutness, gender role preferences, and locus of marital control.
Journal of Family Issues, 8, 136-156.
Scaramella, L. V., Conger, R. D, & Simons, R. L. (1999). Parental protective influences
and gender-specific increases in adolescent internalizing and externalizing
problems. Journal of Research on Adolescence, 9(2), 111-141.
Schaefer, E. S. (1959). A circumplex model for maternal behavior. Journal of Abnormal
and Social Psychology, 59, 226-235.
Schaefer, E. S. (1965). Children?s reports of parental behavior: An inventory. Child
Development, 36, 413-424.
Schur, E. M. (1969). Our criminal society. Englewood Cliffs, NJ: Prentice-Hall.
231
Seligman, M. E. P. (1991). Learned optimism. New York: Simon & Schuster.
Selnow, G. W. (1987). Parent-child relationships and single- and two-parent families:
Implications for substance usage. Journal of Drug Education, 17, 315-326.
Shehan, C. L., Bock, E. W., & Lee, G. R. (1990). Religious heterogamy, religiosity, and
marital happiness ? the case of Catholics. Journal of Marriage and the Family,
52, 73-79.
Shek, D. T. L., & Ma, H. K. (2001). Parent-adolescent conflict and adolescent antisocial
and prosocial behavior: A longitudinal study in a Chinese context. Adolescence,
36(143), 545-555.
Shortz, J. L., & Worthington, E. L., Jr. (1994). Young adults? recall of religiosity,
attributions, and coping in parental divorce. Journal for the Scientific Study of
Religion, 33, 172-179.
Simons, R. L., Johnson, C., & Conger, R. D. (1994). Harsh corporal punishment versus
quality of parental involvement as an explanation of adolescent maladjustment.
Journal of Marriage and the Family, 56, 591-607.
Simons, R. L., Whitbeck, L. B., Conger, R. D., & Conger, K. (1991). Parenting factors,
social skills, and value commitment as precursors to school failure, involvement
with deviant peers, and delinquent behavior. Journal of Youth and Adolescence,
20, 645-663.
Skinner, D., Bailey, D. B., Correa, V., & Rodriguez, P. (1999). Narrating self and
disability: Latino mothers? construction of identities vis-a-vis their child with
special needs. Exceptional Children, 65, 481-495.
232
Slater, E. J., & Haber, J. D. (1984). Adolescent adjustment following divorce as a
function of familial conflict. Journal of Consulting and Clinical Psychology, 52,
920-921.
Slicker, E. K. (1998). Relationship of parenting style to behavioral adjustment in
graduating high school seniors. Journal of Youth and Adolescence, 27(3), 345-
372.
Sloane, D. M., & Potvin, R. H. (1986). Religion and delinquency: Cutting through the
maze. Social Forces, 65(1), 87-105.
Smith, C. (2003a). Religious participation and parental moral expectations and
supervision of American youth. Reviews of Religious Research, 44(4), 414-424.
Smith, C. (2003b). Theorizing religious effects among American adolescents. Journal for
the Scientific Study of Religion, 42(1), 17-30.
Smith, C., Denton, M. L., Faris, R., & Regnerus, M. (2002). Mapping American
adolescent religious participation. Journal for the Scientific Study of Religion,
41(4), 597-612.
Smith, C., & Faris, R. (2002). Religion and American adolescent delinquency, risk
behaviors, and constructive social activities. A research report of the National
Study of Youth and Religion, No. 1, Chapel Hill, NC: NSYR.
Smith, C., Faris, R., Denton, M. L., & Regnerus, M. (2003). Mapping American
adolescent subjective religiosity and attitudes of alienation toward religion: A
research report. Sociology of Religion, 64(1), 111-113.
Smith, C., & Kim, P. (2002). Religious youth are more likely to have positive
233
relationships with their fathers. Retrieved July 3, 2002, from the National Study of
Youth and Religion web site: http://www.youthandreligion.org/news/7-2-
2002.html
Smith, C., & Krohn, M. D. (1995). Delinquency and family life among male adolescents:
The role of ethnicity. Journal of Youth and Adolescence, 24, 69-93.
Snarey, J. R., & Dollahite, D. C. (2001). Varieties of religion-family linkages. Journal of
Family Psychology, 15(4), 646-651.
Snyder, J., & Patterson, G. R. (1987). Family interaction and delinquent behavior. In H.
C. Quay (Ed.), Handbook of delinquent behavior (pp. 216-243). New York: John
Wiley.
Sokol-Katz, J, Dunham, R., & Zimmerman, R. (1997). Family structure versus parental
attachment in controlling adolescent deviant behavior: A social control model.
Adolescence, 32(125), 199-215.
Speltz, M. L., Greenberg, M. T., & DeKlyen, M. (1990). Attachment in preschoolers with
disruptive behavior: A comparison of clinic-referred and non-problem children.
Development and Psychopathology, 2, 31-46.
Sroufe, L. A. (1983). Infant caregiver attachment and patterns of adaptation in preschool:
The roots of maladaptation and competence. In M. Perlmutter (Ed.), Minnesota
Symposium on Child Psychology: Vol. 16 (pp. 41-81). Hillsdale, NJ: Erlbaum.
Sroufe, L. A., Egeland, B., & Kreutzer, T. (1990). The fate of early experience following
developmental change: Longitudinal approaches to individual adaptation in
childhood. Child Development, 61, 1363-1373.
234
Stark, R. (1996). Religion as context: Hellfire and delinquency one more time. Sociology
of Religion, 57(2), 163-173.
Stark R., & Bainbridge, W. S. (1996). Religion, deviance, and social control. New York:
Routledge.
Stark, R., Kent, L., & Doyle, D. P. (1982). Religion and delinquency: The ecology of a
?lost? relationship. Journal of Research in Crime and Delinquency, 19, 4-24.
Stattin, H., & Kerr, M. (2000). Parental monitoring: A reinterpretation. Child
Development, 71(4), 1072-1085.
Steinberg, L. (1986). Latchkey children and susceptibility to peer pressure.
Developmental Psychology, 22, 433-439.
Steinberg, L. (1987). Single parents, stepparents, and the susceptibility of adolescents to
antisocial peer pressure. Child Development, 58, 269-275.
Steinberg, L. (1990). Interdependency in the family: Autonomy, conflict, and harmony. In
S. Feldman & G. Elliot (Eds.), At the threshold: The developing adolescent (pp.
255-276). Cambridge, MA: Harvard University Press.
Steinberg, L. (2001). We know some things: Parent-adolescent relationships in retrospect
and prospect. Journal of Research on Adolescence, 11(1), 1-19.
Steinberg, L., Darling, N. E., & Fletcher, A. C. (1995). Authoritative parenting and
adolescent adjustment: An ecological journey. In P. Moen, G. H. Elder, & K.
L?scher (Eds.), Examining lives in context: Perspectives on the ecology of human
development (pp. 423-466). Washington, DC: American Psychological
Association.
235
Steinberg, L., Elmen, J. D., & Mounts, N. S. (1989). Authoritative parenting,
psychosocial maturity, and academic success among adolescents. Child
Development, 60, 1424-1436.
Steinberg, L., Lamborn, S., Darling, N., Mounts, N., & Dornbusch, S. (1994). Over-time
changes in adjustment and competence among adolescents from authoritative,
authoritarian, indulgent, and neglectful families. Child Development, 65, 754-770.
Steinberg, L., Lamborn, S., Dornbusch, S., & Darling, N. (1992). Impact of parenting
practices on adolescent achievement: Authoritative parenting, school
involvement, and encouragement to succeed. Child Development, 63, 1266-1281.
Steinberg, L., Mounts, N., Lamborn, S., & Dornbusch, S. (1991). Authoritative parenting
and adolescent adjustment across various ecological niches. Journal of Research
on Adolescence, 1, 19-36.
Steinberg, L., & Silk, J. S. (2002). Parenting adolescents. In M. H. Bornstein (Ed.),
Handbook of parenting: Vol. 1. Children and parenting (2nd ed., pp. 103-133).
Mahwah, NJ: Lawrence Erlbaum Associates.
Steiner, F. (1924). Religion and roguery. New York: Truth Seekers.
Sternberg, K. J., Lamb, M. E., Greenbaum, C., Cicchetti, D., Dawud, S., Cortes, R. M.,
Krispin, O., & Lorey, F. (1993). Effects of domestic violence on children?s
behavior problems and depression. Developmental Psychology, 29, 44-52.
Stierlin, H. (1974). Separating parents and adolescents. New York: Quadrangle.
Stoolmiller, M., Patterson, G. R., & Snyder, J. (1997). Parental discipline and child
antisocial behavior: A contingency-based theory and some methodological
236
refinements. Psychological Inquiry, 8, 223-229.
Straus, M. A. (1991). Discipline and deviance: Physical punishment of children and
violence and other crime in adulthood. Social Problems, 38, 133-154.
Sullivan, K. T. (2001). Understanding the relationship between religiosity and marriage:
An investigation of the immediate and longitudinal effects of religiosity on
newlywed couples. Journal of Family Psychology, 15(4), 610-626.
Sutton, S. E., Cowen, E. L., Crean, H. F., & Wyman, P. A. (1999). Pathways to
aggression in young, highly stressed urban children. Child Study Journal, 29(1),
49-67.
Taft, D. R. (1956). Criminology. New York: Macmillan.
Tappan, P. W. (1949). Juvenile delinquency. New York: McGraw-Hill.
Thomas, D. L. (1988). Future prospects for religion and family studies: The Mormon
case. In D. L. Thomas (Ed.), The religion and family connection: Social science
perspectives (pp. 357-382). Provo, UT: Religious Studies Center, Brigham Young
University.
Thornton, A., & Camburn, D. (1989). Religious participation and adolescent sexual
behavior. Journal of Marriage and the Family, 51, 641-653.
Tittle, C. R., & Welch, M. R. (1983). Religiosity and deviance: Toward a contingency
theory of constraining effects. Social Forces, 61, 653-682.
Tolan, P. H., Gorman-Smith, D., Huesmann, L. R., & Zelli, A. (1997). Assessment of
family relationship characteristics: A measure to explain risk for antisocial
behavior and depression among urban youth. Psychological Assessment, 9(3),
237
212-223.
Tolan, P. H., & Lorion, R. P. (1988). Multivariate approaches to the identification of
delinquency-proneness in males. American Journal of Community Psychology,
16, 547-561.
Tomlinson, R. K. (1991).Unacceptable adolescent behavior and parent-adolescent
conflict. Child and Adolescent Social Work Journal, 8, 33-51.
Travers, J. F., & Davis, R. G. (1961). A study of religious motivation and delinquency.
Journal of Educational Sociology, 34, 205-220.
Turner, C. J., & Willis, R. J. (1979). The relationship between self-reported religiosity
and drug use by college students. Journal of Drug Education, 9(1), 67-78.
Vandewater, E. A., & Lansford, J. E. (1998). Influences of family structure and parental
conflict on children?s well-being. Family Relations, 47(4), 323-330.
Varon, S. R., & Riley, A. W. (1999). Relationship between maternal church attendance
and adolescent mental health and social functioning. Psychiatric Services, 50,
799-805.
Vazsonyi, A. T., & Flannery, D. J. (1997). Early adolescent delinquent behaviors:
Associations with family and school domains. Journal of Early Adolescence,
17(3), 271-293.
Vazsonyi, A. T., Hibbert, J. R., & Snider, J. B. (2003). Exotic enterprise no more:
Adolescent reports of family and parenting processes from youth in four countries.
Journal of Research on Adolescence, 13(2), 129-160.
Vazsonyi, A. T, & Killias, M. (2001). Immigration and crime among youth in
238
Switzerland. Criminal Justice & Behavior, 28(3), 329-366.
Vazsonyi, A. T., & Pickering, L. E. (2000). Family processes and deviance: A
comparison of apprentices and non-apprentices. Journal of Adolescent Research,
15(3), 368-391.
Vazsonyi, A. T., & Pickering, L. E. (2003). The importance of family and school domains
in adolescent deviance: African-American and Caucasian youth. Journal of Youth
and Adolescence, 32(2), 115-128.
Vazsonyi, A. T., Pickering, L. E., Belliston, L., Hessing, D., & Junger, M. (2002).
Routine activities and deviant behaviors: American, Dutch, Hungarian and Swiss
youth. Journal of Quantitative Criminology, 18(4), 397-422.
Vazsonyi, A. T., Pickering, L. E., Junger, M., & Hessing, D. (2001). An empirical test of
A General Theory of Crime: A four-nation comparative study of self-control and
the prediction of deviance. Journal of Research in Crime and Delinquency, 38(2),
91-131.
Vener, A. M., Zaenglein, M. M, & Stewart, C. (1977). Traditional religious orthodoxy,
respect for authority and nonconformity in adolescence. Adolescence, 12, 43-56.
Volk, R. J., Edwards, D. W., Lewis, R. A., & Sprenkle, D. H. (1989). The assessment of
adaptability and cohesion in families of adolescent drug abusers and non-abusers.
Family Relations, 38, 266-272.
von Hentig, H. (1948). The criminal and his victim. New Haven, CT: Yale University
Press.
Vuchinich, S., Bank, L., & Patterson, G. R. (1992). Parenting, peers, and the stability of
239
antisocial behavior in preadolescent boys. Developmental Psychology, 28, 510-
521.
Wahler, R. G., & Dumas, J. E. (1986). Maintenance factors in coercive mother-child
interactions: The compliance and predictability hypotheses. Journal of Applied
Behavior Analysis, 19(1), 13-22.
Wallace, J. M., Jr., and Forman, T. A. (1998). Religion?s role in promoting health and
reducing risk among American youth. Health Education and Behavior, 25, 721-
741.
Wallerstein, J. S., & Kelly, J. B. (1980). Effects of divorce on the visiting father-child
relationship. American Journal of Psychiatry, 137, 1534-1539.
Wattenberg, W. W. (1950). Church attendance and juvenile misconduct. Sociology and
Social Research, 34, 195-202.
Webster-Stratton, C. (1984). Randomized trial of two parent training programs for
families with conduct-disordered children. Journal of Consulting and Clinical
Psychology, 52, 666-678.
Weintraub, K. J., & Gold, M. (1991). Monitoring and delinquency. Criminal Behaviour
and Mental Health, 1, 268-281.
Weisner, T. S., Beizer, L., & Stolze, L. (1991). Religion and families of children with
development delays. American Journal on Mental Retardation, 95, 647-662.
Weiss, B., Dodge, K., Bates, J., & Pettit, G. (1992). Some consequences of early harsh
discipline: Child aggression and a maladaptive social information processesing
style. Child Development, 63, 1321-1335.
240
Wells, L. E., & Rankin, J. H. (1985). Broken homes and juvenile delinquency: An
empirical review. Criminal Justice Abstracts, 17, 249-272.
West, D. J., & Farrington, D. P. (1973). Who becomes delinquent? Second report of the
Cambridge Study in Delinquent Development. London: Heinemann.
West, D., & Farrington, D. (1977). Who becomes delinquent? London: Heinemann.
White, H. R., Johnson, V., & Buyske, S. (2000). Parental modeling and parenting
behavior effects on offspring alcohol and cigarette use: A growth curve analysis.
Journal of Substance Abuse, 12(3), 287-310.
Whitehead, B. D., Wilcox, B. & Rostosky, S. (2001). Keeping the faith: The role of
religious and faith communities in preventing teen pregnancy. Washington, D.C.:
National Campaign to Prevent Teen Pregnancy.
Whitehead, P. C. (1970). Religious affiliation and use of drugs among adolescent
students. Journal for the Scientific Study of Religion, 9, 152-154.
Widom, C. S. (1989). Child abuse, neglect, and violent criminal behavior. Criminology,
27, 251-271.
Wiehe, V. (1990). Religious influence on parental attitudes toward the use of corporal
punishment. Journal of Family Violence, 5, 173-186.
Wilcox, W. B. (1998). Conservative Protestant childrearing: Authoritarian or
authoritative? American Sociological Review, 63, 796-809.
Wills, T. A., & Vaughn, R. (1989). Social support and substance use in early adolescence.
Journal of Behavioral Medicine, 12, 321-339.
Wilson, J., & Musick, M. (1996). Religion and marital dependency. Journal for the
241
Scientific Study of Religion, 35, 30-40.
Wilson, J. Q., & Herrnstein, R. J. (1985). Crime and human nature. New York: Simon &
Schuster.
Windle, M. & Miller-Tutzauer, C. (1992). Confirmatory factor analysis and concurrent
validity of the Perceived Social Support-Family Measure among adolescents.
Journal of Marriage and the Family, 54, 777-787.
Wright, L., Frost, C., & Wisecarver, S. (1993). Church attendance, meaningfulness of
religion, and depressive symptomatology among adolescents. Journal of Youth
and Adolescence, 22, 559-568.
Xu, X., Tung, Y., & Dunaway, R. G. (2000). Cultural, human, and social capital as
determinants of corporal punishment: Toward an integrated theoretical model.
Journal of Interpersonal Violence, 15, 603-630.
Young, M., Denny, G., Luquis, R., & Young, T. (1998). Correlates of sexual satisfaction
in marriage. The Canadian Journal of Human Sexuality, 7, 115-127.
Youniss, J., & Smollar, J. (1985). Adolescent relations with mothers, fathers, and friends.
Chicago: University of Chicago Press.
Zahn-Waxler, C. (1993). Warriors and worriers. Gender and psychopathology.
Development and Psychopathology, 5(1-2), 79-89.
Zaslow, M. J. (1989). Sex differences in children?s response to parental divorce: Samples,
variables, ages and sources. American Journal of Orthopsychiatry, 59, 118-141.
Zhang, L., & Messner, S. F. (1995). Family deviance and delinquency in China.
Criminology, 33(3), 359-387.
242
Zhang, L., Wieczorek, W. F., & Welte, J. W. (1997). The impact of age of onset of
substance use on delinquency. Journal of Research in Crime and Delinquency,
34(2), 253-268.
Zoba, W. (1999). Generation 2K: What parents need to know about the millennials.
Downers Grove, IL: InterVarsity Press.
243
APPENDICES
244
APPENDIX A: ADOLESCENT FAMILY PROCESSES SCALE (AFP)
Scale Item
Closenessa
(Maternal,
" = .84;
Paternal,
" = .89)
1. My mother/father often asks about what I am doing in school.
2. My mother/father gives me the right amount of affection.
3. One of the worst things that could happen to me would be to find
out that I let my mother/father down.
4. My mother/father is usually proud of me when I finish something
at which I?ve worked hard.
5. My mother/father trusts me.
6. I am closer to my mother/father than are a lot of kids my age.
Supporta
(Maternal,
" = .84;
Paternal,
" = .82)
7. My mother/father sometimes puts me down in front of other
people.
8. Sometimes my mother/father won?t listen to me or my opinions.
9. My mother/father sometimes gives me the feeling that I?m not
living up to her/his expectations.
10. My mother/father seems to wish I were a different type of
person.
Monitoringa
(Maternal,
" = .79;
Paternal,
" = .89)
11. My mother/father wants to know who I am with when I go out
with friends or on a date.
12. In my free time away from home, my mother/father knows who
I?m with and where I am.
13. My mother/father wants me to tell her/him where I am if I don?t
come home right after school.
14. When I am not at home, my mother/father knows my
whereabouts.
Communicationb
(Maternal,
" = .85;
Paternal,
" = .89)
15. How often do you talk to your mother/father about things that
are important to you?
16. How often do you talk to your mother/father about major
personal decisions?
17. How often do you talk with your mother/father about problems
you have at school?
18. How often do you talk with your mother/father about your job
plans for the future?
19. How often do you talk with your mother/father about how well
you get along with your teachers?
245
Conflictb
(Maternal,
" = .85;
Paternal,
" = .86)
20. How often do you have disagreements or arguments with your
mother/father?
21. How often do you purposely not talk to your mother/father
because you are mad at her/him?
22. How often do you get angry at your mother/father?
Peer Approvalb
(Maternal,
" = .83;
Paternal,
" = .87)
23. How often does your mother/father approve of your friends?
24. How often does your mother/father approve of your
boyfriend/girlfriend?
25. How often does your mother/father like when you go out with
your friends?
Note. a Possible responses: 1 = strongly disagree, 2 = disagree, 3 = neither disagree nor
agree, 4 = agree, 5 = strongly agree; b Possible responses: 1 = never, 2 = occasionally,
3 = sometimes, 4 = often, 5 = very often.
246
APPENDIX B: RELIGIOSITY ITEMS AND SCALES
Item Responses
Church Attendance: ?How often do you
attend the regularly scheduled services of
a church (not including weddings,
funerals, baptisms, or any other special
occasions of a religious nature)??
1 = practically never
2 = only on holidays or holy days
3 = only occasionally
4 = once a week
5 = more than once a week
Church Status: ?What is your current
status in relation to church??
1 = no association with church
2 = attending non-member
3 = non-attending member
4 = attending member
Religious Involvement: ?What is your
current level of involvement with a
church or religious organization??
1 = I never participate
2 = I rarely participate
3 = I occasionally participate
4 = I frequently participate
5 = I am a paid staff member
Bible Reading: ?How often do you read
the Bible or any other book of religious
faith??
1 = never
2 = infrequently
3 = sometimes
4 = quite a bit
5 = every day
Praying: ?How often do you pray?? 1 = never
2 = infrequently
3 = sometimes
4 = quite a bit
5 = every day
Music Listening: ?How often do you
listen to Christian music (e.g., Steven
Curtis Chapman, Michael W. Smith, DC
Talk, Petra, Glad, Twila Paris, etc.) for
personal pleasure??
1 = never
2 = infrequently
3 = sometimes
4 = quite a bit
5 = every day
247
Religious Conviction: ?Occasionally I
find it necessary to compromise my
religious beliefs in order to protect my
social and economic well-being.?
1 = strongly disagree
2 = disagree
3 = agree
4 = strongly agree
Religious Salience: ?I try hard to carry my
religion over into all my other dealings in
life.?
1 = strongly disagree
2 = disagree
3 = agree
4 = strongly agree
Ritualistic Participation
(" = .71)
Church Attendance
Bible Reading
Religious Salience
Relational Practice
(" = .75)
Church Status
Religious Involvement
Praying
Religiosity
(" = .86)
Church Attendance
Church Status
Religious Involvement
Bible Reading
Praying
Music Listening
Religious Salience
248
APPENDIX C: NORMATIVE DEVIANCE SCALE (NDS)
Scale Items
Vandalism
(" = .89)
1. Smashed bottles on the street, school grounds, or other areas?
2. Intentionally damaged or destroyed property belonging to your
parents or other family members (brothers or sisters)?
3. Intentionally damaged or destroyed property belonging to a
school, college, or university?
4. Intentionally damaged or destroyed other property (signs,
windows, mailboxes, parking meter, etc.) that did not
belong to you?
5. Intentionally damaged or destroyed property belonging to your
employer or at your workplace?
6. Slashed or in any way damaged seats on a bus, in a movie
theater, or something at another public place?
7. Written graffiti on a bus, on school walls, on rest room walls, or
on anything else in a public place?
8. Committed acts of vandalism when coming or going to a
football game or other sports event?
Alcohol Use
(" = .89)
9. Consumed hard liquor (e.g. tequila, whiskey, vodka, or gin)
before you were 21?
10. Consumed alcoholic beverages (e.g. beer, wine, or wine
coolers) before you were 21?
11. Got drunk (intentionally) just for the fun of it (at any age)?
12. Got drunk just to fit in and be part of the crowd (at any age)?
13. Lied about your age to buy alcohol before you turned 21?
14. Had an older brother/sister or friend buy alcohol for you?
15. Bought alcohol for a brother/sister or friend?
249
Drug Use
(" = .92)
16. Used tobacco products regularly (e.g., cigarettes, chew, snuff,
etc.)?
17. Used "soft" drugs such as marijuana (grass, pot)?
18. Used "hard" drugs such as crack, cocaine, or heroin?
19. Gone to school when you were drunk or high on drugs?
20. Gone to work when you were drunk or high on drugs?
21. Gone to a concert when you were drunk or high on drugs?
22. Gone to a club/dance/party when you were drunk or high on
drugs?
23. Gone to a club/dance/party to get drunk or high on drugs?
24. Sold any drugs such as marijuana (grass, pot), cocaine, or
heroin?
School
misconduct
(" = .86)
25. Cheated on school/college/university tests (e.g., cheat sheet,
copy from neighbor, etc.)?
26. Been sent out of a classroom because of "bad" behavior (e.g.
inappropriate behaviors, cheating etc.)?
27. Been suspended or expelled from school/college/university?
28. Stayed away from school/classes when your parent(s) thought
you were there?
29. Intentionally missed classes over a number of days for "no
reason," just for fun (e.g., there was no family emergency)?
30. Been in trouble at school so that your parents received a phone
call about it?
31. Skipped school/work (pretending you are ill)?
General Deviance
(" = .88)
32. Intentionally disobeyed a stop sign or a red traffic light while
driving a vehicle?
33. Been on someone else's property when you knew you were not
supposed to be there?
34. Failed to return extra change that you knew a cashier gave you
by mistake?
35. Tried to deceive a cashier to your advantage (e.g. flash a larger
bill and give a smaller one)?
36. Let the air out of the tires of a car or bike?
37. Lied about your age to get into a nightclub/bar?
38. Made nuisance/obscene telephone calls?
39. Avoided paying for something (e.g. movies, bus or subway
rides, food, etc.)?
40. Used fake money or other things in a candy, coke, or stamp
machine?
41. Shaken/hit a parked car just to turn on the car's alarm?
250
42. Stayed out all night without informing your parents about your
whereabouts?
Theft
(" = .87)
43. Stolen, taken, or tried to take something from a family member
or relative (e.g. personal items, money, etc.)?
44. Stolen, taken, or tried to take something worth $10 or less (e.g.
newspaper, pack of gum, mail, money, etc.)?
45. Stolen, taken, or tried to take something worth between $10
and $100 (e.g. shirt, watch, cologne, video game cartridge,
shoes, money, etc.)?
46. Stolen, taken, or tried to take something worth more than $100
(e.g. leather jacket, car stereo, bike, money, etc.)?
47. Stolen, taken, or tried to take something that belonged to "the
public" (e.g. street signs, construction signs, etc.)?
48. Stolen or tried to steal a motor vehicle (e.g., car or
motorcycle)?
49. Bought, sold, or held stolen goods or tried to do any of these
things?
Assault
(" = .82)
50. Hit or threatened to hit a person?
51. Hit or threatened to hit your parent(s)?
52. Hit or threatened to hit other students/peers or people?
53. Used force or threatened to beat someone up if they didn't give
you money or something else you wanted?
54. Been involved in gang fights or other gang activities?
55. Beaten someone up so badly they required medical attention?
Note. Items were prefaced with the leader, ?Have you ever . . .? and had 5 possible
responses: 1 = no/never, 2 = once/one time, 3 = 2-3 times, 4 = 4-6 times, 5 = 6 or more
times.
APPENDIX D: RELIGIOSITY CORRELATION MATRIX
A B C D E F G H I J
(B) Church Status .782
(C) Church Involvement .642 .633
(D) Bible Reading .463 .377 .432
(E) Praying .489 .458 .443 .579
(F) Christian Music Listening .402 .334 .406 .492 .448
(G) Religious Beliefs .023ns .079* .026ns -.087* .003ns -.062ns
(H) Religious Salience .465 .456 .477 .453 .503 .383 .110**
(I) Ritualistic Participation .840 .690 .650 .797 .647 .526 .011ns .760
(J) Relational Practice .769 .833 .825 .572 .803 .486 .039ns .576 .801
(K) Total Religiosity .823 .775 .767 .728 .763 .665 .010ns .695 .938 .934
Note. All correlations were significant at p < .001, unless otherwise specified as follows: ns = non-significant, * p < .05, ** p < .01;
using pairwise deletion, n?s ranged from 761-790.
APPENDIX E: FAMILY PROCESSES CORRELATION MATRIX
A B C D E F G H I J K
Maternal
   (B) Support .506
   (C) Monitoring .407 .106**
   (D) Communication .616 .371 .335
   (E) Conflict -.454 -.512 -.103** -.266
   (F) Peer Approval .546 .413 .254 .490 -.319
Paternal
   (G) Closeness .411 .216 .280 .341 -.160 .317
   (H) Support .312 .467 .179 .258 -.315 .284 .426
   (I) Monitoring .196 .068ns .355 .210 -.102** .173 .478 .057ns
   (J) Communication .295 .186 .158 .525 -.139 .290 .642 .307 .397
   (K) Conflict -.155 -.182 -.118** -.082* .321 -.136 -.329 -.528 -.001ns -.161
   (L) Peer Approval .306 .235 .202 .291 -.136 .535 .559 .353 .252 .487 -.172
Note. All correlations were significant at p < .001, unless otherwise specified as follows: ns = non-significant, * p < .05, ** p < .01;
using pairwise deletion, n?s ranged from 795-833.
APPENDIX F: DEVIANCE CORRELATION MATRIX
A B C D E F G H I J K
(B) Alcohol Use .540
(C) Drug Use .627 .792
(D) School Misconduct .682 .662 .706
(E) General Deviance .775 .634 .701 .763
(F) Theft .718 .525 .635 .677 .776
(G) Assault .637 .433 .518 .601 .649 .708
(H) Total Deviance .841 .809 .875 .868 .905 .831 .732
(I) Victim Deviance .731 .520 .619 .689 .771 .927 .922 .842
(J) Nonvictim Deviance .730 .889 .912 .877 .856 .725 .610 .972 .721
(K) Antiascetic Deviance .615 .951 .942 .721 .704 .611 .499 .888 .599 .951
(L) Secular Deviance .876 .650 .734 .870 .912 .883 .816 .961 .917 .880 .729
Note. All correlations were significant at p < .001, unless otherwise specified as follows: ns = non-significant, * p < .05, ** p < .01;
using pairwise deletion, n?s ranged from 819-836.
APPENDIX G: FAMILY PROCESSES AND RELIGIOSITY CORRELATION MATRIX
Church
Attendance
Church
Status
Church
Involvement
Bible
Reading Praying
Music
Listening
Religious
Conviction
Religious
Salience
Ritualistic
Participation
Relational
Practice Religiosity
Maternal
   Closeness .203 .223 .179 .135 .207 .133 -.013ns .187 .244 .227 .252
   Support .119** .153 .116** .087* .136 .025ns -.043ns .087* .171 .127 .147
   Monitoring .198 .207 .209 .164 .173 .206 -.063ns .177 .236 .229 .258
   Communication .190 .181 .213 .209 .251 .179 -.061ns .233 .268 .260 .282
   Conflict -.090* -.105** -.072* -.075* -.063ns -.039ns -.008ns -.081* -.101** -.112** -.109**
   Peer Approval .174 .155 .144 .120** .191 .140 -.072* .121** .202 .177 .207
Paternal
   Closeness .167 .164 .160 .158 .232 .162 -.040ns .154 .224 .203 .230
   Support .105** .155 .146 .123** .166 .065ns -.051ns .121** .187 .146 .168
   Monitoring .107** .083* .123** .100** .109** .138 -.076* .137 .123 .141** .151
   Communication .189 .144 .181 .172 .211 .188 -.055ns .199 .218 .236 .243
   Conflict -.102** -.132 -.132 -.083* -.069ns -.037ns .041ns -.080* -.130** -.114 -.121**
   Peer Approval .195 .195 .196 .148 .237 .174 -.024ns .164 .251 .211 .248
Note. All correlations were significant at p < .001, unless otherwise specified as follows: ns = non-significant, * p < .05, ** p < .01; using pairwise deletion, n?s ranged from 743-
781.
APPENDIX H: FAMILY PROCESSES AND DEVIANCE CORRELATION MATRIX
Vandalism
Alcohol
Use
Drug
Use
School
Misconduct
General
Deviance Theft Assault
Total
Deviance
Victim
Deviance
Non-Victim
Deviance
Secular
Deviance
Antiascetic
Deviance
Maternal
   Closeness -.356 -.197 -.250 -.286 -.307 -.319 -.311 -.329 -.338 -.288 -.235 -.354
   Support -.251 -.172 -.200 -.208 -.243 -.251 -.225 -.256 -.257 -.229 -.196 -.265
   Monitoring -.374 -.346 -.384 -.333 -.379 -.350 -.323 -.413 -.361 -.406 -.384 -.393
   Communication -.260 -.211 -.239 -.260 -.259 -.249 -.223 -.277 -.256 -.272 -.237 -.275
   Conflict .263 .200 .190 .229 .231 .229 .225 .246 .248 .238 .206 .255
   Peer Approval -.309 -.130 -.239 -.239 -.253 -.238 -.244 -.264 -.261 -.238 -.192 -.282
Paternal
   Closeness -.276 -.213 -.236 -.260 -.225 -.273 -.260 -.295 -.284 -.263 -.236 -.301
   Support -.254 -.189 -.254 -.239 -.237 -.254 -.244 -.273 -.266 -.258 -.232 -.274
   Monitoring -.240 -.245 -.218 -.236 -.265 -.232 -.168 -.277 -.210 -.271 -.245 -.264
   Communication -.164 -.165 -.157 -.210 -.151 -.165 -.100** -.194 -.140 -.193 -.170 -.187
   Conflict .217 .153 .150 .179 .199 .166 .196 .217 .193 .191 .160 .227
   Peer Approval -.270 -.206 -.264 -.237 -.224 -.251 -.210 -.289 -.247 -.262 -.247 -.281
Note. All correlations were significant at p < .001, unless otherwise specified as follows: ns = non-significant, * p < .05, ** p < .01; using pairwise deletion, n?s ranged from 791-
827.