! ! Restrictive Eating and Risk for Suicidal Behavior: An Aplication of the Interpersonal- Psychological Theory of Suicide by Kely L. Zuromski A thesis submited to the Graduate Faculty of Auburn University in partial fulfilment of the requirements for the Degree of Master of Science Auburn, Alabama May 4, 2014 Copyright 2014 by Kely L. Zuromski Approved by Tracy K. Wite, Chair, Asistant Profesor of Psychology Chris J. Correia, Profesor of Psychology Frank W. Weathers, Profesor of Psychology ! ii Abstract Though the relationship betwen methods of dieting and suicidal behavior has been established (e.g., Crow et al., 2008), existing studies have not examined mechanisms underlying this relationship. This study sought to enhance understanding of this relationship using the interpersonal-psychological theory of suicide (IPTS; Joiner, 2005). We compared a sample of selected undergraduates who endorsed recently engaging in one dieting method (i.e., restrictive eating; n = 99) to control undergraduates endorsing no lifetime use of dietary methods (n = 94). We hypothesized that, statisticaly controlling for presence of other non-restrictive dietary methods, the restrictive eating group would have great acquired capability for suicide (ACS) and higher likelihood of suicide atempt history, and that no group diferences would emerge on other IPTS constructs or suicidal ideation (SI). Contrary to hypotheses, no diferences emerged betwen groups on ACS, and frequency of restrictive eating within the dietary restrictor group was not significantly asociated with ACS. Consistent with hypotheses, dietary restrictors were more likely to have suicide atempt history, and groups had comparable levels of other IPTS constructs and SI. Though results were not entirely consistent with hypotheses, the current study represents the first atempt at isolating and examining a specific dietary method using the IPTS. Results suggest that, in isolation, restrictive eating may not be directly contributing to risk for suicide. ! iii Table of Contents Abstract.........................................................................................................ii List of Tables..................................................................................................iv Introduction....................................................................................................1 Method........................................................................................................14 Participants and Procedure........................................................................14 Measures.............................................................................................16 Results.........................................................................................................24 Preliminary Analyses...............................................................................24 Hypothesis 1: Group Diferences in Psychosocial Variables.................................25 Hypothesis 2: Group Diferences in Thwarted Belongingnes and Perceived Burdensomenes....................................................................................26 Hypothesis 3: Group Diferences in Suicidal Ideation and Atempts........................26 Hypothesis 4: Group Diferences in ACS.......................................................27 Hypothesis 5: ACS and Use of Restrictive Dietary Methods.................................27 Post-hoc Analyses..................................................................................28 Discussion....................................................................................................29 References....................................................................................................38 Appendix 1: Tables..........................................................................................51 ! iv List of Tables Table 1: Demographics for Dietary Restrictor and Control Groups..................................52 Table 2: Descriptive Statistics for Use of Lifetime Restrictive and Non-Restrictive Methods Within Dietary Restrictor Group................................................................53 Table 3: Items Asesing Frequency of Restrictive Dietary Methods Over Lifetime and Currently............................................................................................54 Table 4: Items Asesing Frequency of Non-Restrictive Methods Over Lifetime..................55 Table 5: Intercorrelations Betwen al Study Variables for Dietary Restrictor Group.............56 Table 6: Intercorrelations Betwen al Study Variables for Control Group.........................58 Table 7: Results of Hypotheses 1-4.......................................................................59 RESTRICTIVE EATING RISK FOR SUICIDE ! 1 Dieting, which refers to intentional and sustained restriction of food intake for the purposes of weight loss or maintenance (Stice, Fisher, & Lowe, 2004; Wadden, Brownel, & Foster, 2002) is endorsed by as many as 70% of men and women in the United States (French, Jefery, & Murray, 1999; Jefery, Adlis, & Forster, 1991). Dieting is especialy common among college-aged women, often acompanying weight gain that occurs during the transition from high school to college (Anderson, Shapiro, & Lundgren, 2003; Lowe et al., 2006). Although les common than among women, some research suggests that college-aged men also engage in dietary practices to control weight (Lavender, De Young, & Anderson, 2010; O?Dea & Abraham, 2002). Though dieting is usualy thought to mean restriction of food intake for the purposes of weight loss or maintenance (Stice et al., 2004), the distinction betwen normal dietary practices and eating pathology are unclear, as the term dieting is also commonly used to refer to more extreme weight control methods such as purging and excesive exercise (Ackard, Croll, & Kearney-Cooke, 2002; Brownel & Rodin, 1994; Mintz & Betz, 1988; Whitehouse & Button, 1988). Although these practices may be characteristic of eating disorders such as bulimia nervosa (BN) and anorexia nervosa (AN), some college students without clinical eating disorder diagnoses also report using methods such as purging, taking diet pils, or using laxatives to influence weight (Brownel & Rodin, 1994; Bunnel, Shenker, Nussbaum, Jacobson, & Cooper, 1990; Chamay-Weber, Narring, & Michaud, 2005). However, most frequently used by college students are practices that restrict food intake to influence weight (i.e., restrictive eating), such as eating fewer calories, eliminating certain types of food, skipping meals, and in extreme cases, RESTRICTIVE EATING RISK FOR SUICIDE ! 2 fasting (Lavender, De Young, & Anderson, 2010; Mintz & Betz, 1988; O?Dea & Abraham, 2002; Striegel-Moore et al., 2009; Tylka & Subich, 2002). Changes in diet and reduction of caloric intake are often asociated with health benefits for overweight or obese individuals, including decreased risk of cardiovascular problems (Goodpaster et al., 2010) and decreased risk of diabetes (Van Dam, Wilet, Rim, Stampfer, & Hu, 2002). Additionaly, dieting may be related to promotion of healthy behavior such as increased physical exercise and increased fruit and vegetable intake (French, Perry, Leon, & Fulkerson, 1995; Rafiroui, Sargent, Parra-Medina, Drane, & Valois, 2003). However, presence of more extreme, unhealthy weight loss behavior has received atention in the research literature because of its potential efects on long-term health outcomes such as osteoporosis and its asociation with a wide range of unhealthy behavior, such as problematic use of alcohol, marijuana, and cigaretes (Brownel & Rodin, 1994; Chamay-Weber et al., 2005; Heidelberg & Correia, 2009; Krahn, Kurth, Gomberg, & Drewnowski, 2005; Rafiroui et al., 2003). Though many negative health outcomes have been identified and studied in relation to dieting, one area that warrants further research is its relationship with suicidal behavior. Existing literature suggests a disconcerting correlation betwen dieting and suicidal ideation and atempts (Crow, Eisenberg, Story, & Neumark-Sztainer, 2008; French, Story, Downes, Resnick, & Blum, 1995; Neumark-Sztainer, Story, Dixon, & Murray, 1998; Rafiroui et al., 2003). Additionaly, dieting is correlated with depresive symptoms, low self-esteem, and problems in interpersonal relationships, al factors that may also be related to the presence of suicidal behavior (Ackard et al., 2002; Bunnel et al., 1990; Crow, Eisenberg, Story, & Neumark-Sztainer, 2006; Crow et al., 2008; French et al., 1996). RESTRICTIVE EATING RISK FOR SUICIDE ! 3 Relatively few studies have investigated the relationship betwen suicidal behavior and dieting, and those that have are hampered by limitations. Researchers have often asesed for suicidal ideation and atempts in a dichotomous manner (e.g., Crow et al., 2008; Neumark- Sztainer et al., 1998; Rafiroui et al., 2003), which limits ability to explore the relationship betwen dieting and the intensity and frequency of suicidal behavior. Furthermore, past studies reporting information on dieting and suicidal behavior have often not utilized a specific operational definition of dieting, which makes it dificult to discern to which behavior the term dieting refers (Rafiroui et al., 2003; Timko, Perone, & Crossfield, 2006). For example, dieting may be asesed with the question: How often have you gone on a diet during the last year? By diet, we mean changing the way you eat so you can lose weight (French et al., 1995b). This question is vague and could be interpreted by respondents to include a variety of disparate types of behavior (e.g., fasting for long stretches of time, eliminating sugary foods from one?s diet). Even when researchers collect information about specific dietary practices, practices are often grouped together for statistical analyses (e.g., Crow et al., 2008; Neumark-Sztainer et al., 1998; Rafiroui et al., 2003). Moreover, an overarching limitation is the atheoretical nature of the existing literature connecting dieting and suicidal behavior, which focuses only on the correlation betwen dieting and suicidal behavior, rather than on atempting to identify explanatory mechanisms. The interpersonal-psychological theory of suicide (IPTS; Joiner, 2005; Van Orden et al., 2010) proposes a comprehensive set of causal proceses for suicide that may elucidate the heretofore- unexplained link betwen dieting and suicidal behavior. Although risk factors such as previous suicide atempts (Christiansen & Jensen, 2007), mental disorders (Nock, Hwang, Sampson, & Kesler, 2010), social isolation (Duberstein et al., 2004), family conflict (Joiner et al., 2002), and RESTRICTIVE EATING RISK FOR SUICIDE ! 4 unemployment (Brown, Beck, Ster, & Grisham, 2000) have been asociated with suicidal behavior, each factor in isolation is limited as a predictor. The IPTS seks to iluminate how these distal risk factors, and others, are related proximaly to suicidal behavior through a precise and empiricaly-based causal pathway (Van Orden et al., 2010). The current study seks to integrate the literature on dieting and suicidal behavior with the theoretical propositions of the IPTS in order to isolate and examine the relationship betwen specific restrictive dieting methods and suicidal behavior. The Interpersonal-Psychological Theory of Suicide The IPTS suggests that three separate constructs must be in place before a lethal suicide atempt can occur: thwarted belongingnes, perceived burdensomenes, and an acquired capability for suicide (Joiner, 2005; Van Orden et al., 2010). Thwarted belongingnes refers to a feling of social disconnectednes, wherein an individual lacks support and does not provide support to others. Perceived burdensomenes refers to an individual?s sense that he or she is a liability to others, acompanied by felings of self-hate. The presence of either thwarted belongingnes or perceived burdensomenes in isolation is proposed to create pasive suicidal ideation (e.g., I wish I were dead). The joint presence of these constructs in combination with hopelesnes about these conditions ever improving wil produce active suicidal desire (e.g., I want to kil myself). However, acording to the theory, suicidal desire alone is not sufficient to lead to lethal (or near-lethal) suicidal behavior. In addition, an individual must also posses a lowered fear of death, as wel as elevated physical pain tolerance in order to engage in the fearsome and painful act of lethal self-injury. Lowered fear of death and elevated physical pain tolerance together comprise the third construct of the theory, known as the acquired capability for suicide (ACS; Van Orden et al., 2010). RESTRICTIVE EATING RISK FOR SUICIDE ! 5 The ACS is proposed to develop over time through repeated exposure to a variety of life experiences that are painful and provocative (i.e., fear-inducing). In this sense, the capability to inflict lethal self-harm is acquired over time, suggesting that it is not an innate characteristic. Instead, Joiner (2005) has adapted principles from opponent proces theory (OPT; Solomon & Corbit, 1974) to explain the proces by which the capability to withstand the fear and pain involved in self-injury can be acquired. OPT suggests that observable responses are the result of two underlying proceses: the primary proces and the opponent proces. The primary proces is the initial response to a stimulus (e.g., fear), and the opponent proces is the opposite response (e.g., relief). Over time, repeated exposure to a stimulus results in an increase in the magnitude of the opponent proces, whereas the magnitude of the primary proces response is reduced. Once the magnitude of the opponent proces surpases that of the primary proces, the observable result is habituation. In the context of ACS, instances of repeated self-harm are proposed to result in reduced fear of death and pain response (i.e., the primary proceses) and an increase in the opponent proceses of fearlesnes about death and analgesia (Joiner, 2005; Van Orden et al., 2010). Subsequently, acording to the IPTS, individuals with a history of nonfatal self-harm should have greater acquired capability for lethal self-injury than those without such a history. The relationship betwen both pain tolerance and fearlesnes about death and suicidal behavior has been demonstrated and supported in past research. Regarding pain tolerance, individuals with a history of suicide atempts and non-suicidal self-injury exhibit greater pain tolerance than control participants (Kemperman, Russ, & Shearin, 1997; Orbach, Mikulincer, King, Cohen, & Stein, 1997; Orbach et al., 1996). Similarly, other research indicates that there is a positive asociation betwen frequency of prior self-harm behavior and positive emotional RESTRICTIVE EATING RISK FOR SUICIDE ! 6 experiences following a recent episode of self-injury (Gordon et al., 2010) and following imagined self-injury (Brain, Haines, & Wiliams, 1998). Research also suggests that individuals with more previous suicide atempts self-report greater fearlesnes about death, compared with non-atempters (Van Orden, Wite, Gordon, Bender, & Joiner, 2008). This increased fearlesnes about death among suicide atempters can even distinguish atempters from suicide ideators (Smith, Cukrowicz, Poindexter, Hobson, & Cohen, 2010). Although much research on acquired capability has been gathered from studies of suicide atempters, the majority of individuals who atempt suicide wil not eventualy die by suicide, and many individuals who die by suicide do so on their first atempt (Rudd, Joiner, & Rajab, 1996). These data suggest that other factors besides previous suicide atempts must also contribute to acquiring a capability for suicide (Van Orden et al., 2010). Indeed, a growing empirical literature suggests that combat exposure (Kang & Bullman, 2008), childhood maltreatment (Beautrais, 2001), exposure to suicidality in others (Insel & Gould, 2008), and impulsive behavior (McGirr et al., 2008) are al linked to suicidal behavior, and Joiner (2005) and Van Orden et al. (2010) suggest that these experiences function as painful and/or provocative events that, after repeated exposure over a period of time, could habituate an individual to physical pain and/or fear of death. However, there are likely other painful and provocative life events that increase ACS. Identification of additional painful and provocative events would serve to clarify the complex pathway from suicidal ideation to fatal atempt in the context of the IPTS. Dieting, Perceived Burdensomenes, and Thwarted Belongingnes Although past research has established a link betwen dieting and suicidal behavior (Crow et al., 2008; French et al., 1995b; Neumark-Sztainer et al., 1998; Rafiroui et al., 2003), no RESTRICTIVE EATING RISK FOR SUICIDE ! 7 studies have specificaly examined the constructs of the IPTS in relation to dieting. However, several psychological correlates and predictors of dieting sem relevant to the first two constructs of the theory (i.e., perceived burdensomenes and thwarted belongingnes), and therefore, may partialy explain the link betwen dieting and suicidal behavior. For example, past research has supported the relationship betwen dieting practices and depresion, low self- estem, and problems in interpersonal relationships, al of which could be related to feling alone and/or like a burden (Ackard, Croll et al., 2002; Crow et al., 2006; Crow et al., 2008; French & Jefery, 1994; Van Orden et al., 2010). Another notable correlate of eating pathology that may be asociated with both perceived burdensomenes and thwarted belongingnes is perfectionism. Much research has established the link betwen perfectionism and eating pathology (e.g., Downey & Chang, 2007; Fairburn, Cooper, Doll, & Welch, 1999; Stice, 2002). Especialy pertinent to the thwarted belongingnes construct are findings that self-reported dieting is highest among individuals with high levels of perfectionism who are also experiencing interpersonal distres (Cain, Bardone-Cone, Abramson, Vohs, & Joiner, 2008). Among those with high levels of perfectionism, maladaptive perfectionists (i.e., those who perceive a high discrepancy betwen their high standards and ability to achieve those standards) may experience more interpersonal problems than adaptive perfectionists (i.e., those who strive for high standards but who are not self-critical about achieving these standards; Slaney, Pincus, Uliaszek, & Wang, 2006). Therefore, maladaptive perfectionism in particular may be asociated with thwarted belongingnes. Moreover, due to a discrepancy betwen their impossibly high standards and their ability to actualy met those standards (Slaney, Rice, Mobley, Trippi, & Ashby, 2001), maladaptive perfectionists may experience more intense felings of perceived burdensomenes, compared to adaptive RESTRICTIVE EATING RISK FOR SUICIDE ! 8 perfectionists, insofar as they fel that they are a liability on those around them and/or fel self- hatred because of these failings. Indeed, research by Rasmussen and colleagues (2012) suggests that perceived burdensomenes mediates the relationship betwen maladaptive perfectionism and suicidal ideation in a sample of unselected undergraduates, proposing that the perception of burdensomenes experienced by some maladaptive perfectionists may cause such psychological distres that it leads to suicidal ideation. Although dieting may be related to both perceived burdensomenes and thwarted belongingnes, the nature of the relationship betwen dieting and these constructs is unclear. Some longitudinal research has suggested that depresive symptoms, perfectionism, problems in parent-child relationships, and low self-esteem precede the development of dieting in adolescents (Hil & Palin, 1998; Stice, 2002; Von Soest & Wichstrom, 2009). That these psychosocial risk factors, which are presumably asociated with thwarted belongingnes and perceived burdensomenes, precede the development of dieting suggests that dieting, thwarted belongingnes, and perceived burdensomenes may be beter conceptualized as multifinal outcomes of these psychosocial risk factors. For example, maladaptive perfectionism may lead to both perceived burdensomenes and dieting. However, it is perceived burdensomenes, not dieting, that acounts for the relationship betwen maladaptive perfectionism and suicidal ideation, as found by Rasmussen et al. (2012). Thus, the relationship betwen dieting and suicidal ideation may be spurious and acounted for by psychosocial variables that are related to dieting and thwarted belongingnes and/or perceived burdensomenes. Nevertheles, existing research has not examined the relationship betwen thwarted belongingnes or perceived burdensomenes in relation to dieting practices and suicidal ideation, so the relationship among the variables remains unknown. The current study wil addres this limitation. RESTRICTIVE EATING RISK FOR SUICIDE ! 9 Dieting and the Acquired Capability for Suicide The range of behavior often asociated with dieting (e.g., fasting, skipping meals, purging) may be considered painful experiences that could facilitate the habituation to physical pain, and possibly, fear of death. Indeed, previous research has found that methods of weight control damage tisues, exert pain, and have adverse health consequences (Baker & Sandle, 1996; Helstrom, 2007; Sidiropoulos, 2007). Therefore, the repetition of these practices necesary to lose weight, acompanied by enduring the pain asociated with the behavior, may contribute to an increased tolerance of physical pain. This increased pain tolerance may make it possible for an individual who desires suicide to enact a medicaly lethal suicide atempt. In support of this notion that there is a relationship betwen methods of weight control and ACS, individuals diagnosed with anorexia nervosa (AN) ? a disorder characterized by restrictive eating ? have high rates of suicide (e.g., Preti et al., 2011). The suggestion that restrictive eating increases ACS, and therefore can partialy explain the link betwen AN and suicide, was first posited by Joiner (2005). However, the existing literature on this topic is limited and relies on imperfect measures of restrictive eating, ACS, or both. For example, there is evidence that individuals with AN have higher pain tolerance compared to control subjects (de Zwan, Biener, Bach, Wiesnagrotzki, & Stacher, 1996; Lautenbacher, Pauls, Strian, Pirke, & Krieg, 1991; Raymond et al., 1999); furthermore, among those with AN, there is a positive asociation betwen self-injury and pain tolerance (Claes, Vandereycken & Vertommen, 2006). However, no study has demonstrated this relationship betwen people who only engage in restrictive eating (as opposed to having a diagnosis of AN) and increased physical pain tolerance, which leaves open the possibility that other asociated features of AN (e.g., genetic risk factors) explain the asociation betwen this disorder and pain tolerance. Moreover, no study has RESTRICTIVE EATING RISK FOR SUICIDE ! 10 explored the possible relationship betwen restrictive eating and the other component of acquired capability ? fearlesnes about death. Three recent studies provide additional, albeit, indirect support for the relationship betwen restrictive eating and ACS. Holm-Denoma et al. (2008) demonstrated that individuals with AN use highly lethal suicide methods that would likely result in death even in individuals of normal weight status. On the basis of these results, the authors posited that people with AN are capable of using lethal suicide methods because they have habituated to pain after repeated engagement in self-starvation. Selby et al. (2010) found an asociation betwen the restrictive subtype of AN and the frequency and lethality of suicidal behavior, even after controlling for endorsement of other types of painful and/or provocative behavior (e.g., self-induced vomiting, non-suicidal self-injury). On the basis of these results, Selby et al. (2010) reasoned that the habitual starvation experienced by individuals with the restrictive subtype of AN results in habituation to fear of death and physical pain. In addition, Wite et al. (2012) investigated the relationship betwen various painful and provocative experiences and suicidal behavior in a sample of patients diagnosed with body dysmorphic disorder (BDD). One finding of this study was that BDD-related restrictive eating (i.e., dieting or restricting eating to improve appearance of body areas of concern) was asociated with suicide atempts, but not with suicidal ideation. The authors argued that this finding provides evidence of a specific relationship betwen restrictive food intake and ACS, as opposed to desire for suicide. Although these studies provide preliminary support for the relationship betwen restrictive eating and ACS, no direct measures of ACS were utilized. Additionaly, measures of restrictive eating were limited?no measure was utilized by Holm-Denoma et al. (2008), whereas Selby et al. (2010) only specified AN, restrictive subtype, rather than using a direct measure of dietary restriction, and Wite et al. RESTRICTIVE EATING RISK FOR SUICIDE ! 11 (2012) used a dichotomous measure of restrictive eating. Thus, studies of the relationship betwen restrictive eating and ACS in clinical samples are limited; furthermore, to our knowledge, there have been no investigations of the link betwen acquired capability and restrictive eating in non-clinical populations. Curent Study We aimed to integrate and improve upon two existing lines of research: the first, examining the relationship betwen dieting and suicidal behavior, and the second, investigating the relationship betwen restrictive eating and ACS. Existing literature on dieting and suicidal behavior is limited by its atheoretical nature, inconsistent operational definitions of dieting, and use of dichotomous measures of suicidal behavior. The current study addresed these limitations by using the IPTS as a framework and focusing specificaly on one form of restrictive eating (i.e., fasting for prolonged periods of time) within the broader category of dieting. The existing literature on restrictive eating and ACS is primarily limited by imprecise measurement of key constructs; furthermore, because al existing studies have been conducted with samples of individuals diagnosed with either AN or BDD, it was not possible to demonstrate a relationship betwen restrictive eating specificaly, as opposed to other asociated features of these disorders (e.g., genetics, clinical distres), and ACS. The current study aimed to addres these limitations by using psychometricaly sound instruments and by focusing on presence of restrictive eating in a non-clinical sample. Moreover, the relationship betwen restrictive eating and the other constructs of the theory (i.e., perceived burdensomenes and thwarted belongingnes) was examined for the first time. Because existing research has demonstrated high rates of dieting among college students (Anderson et al., 2003; Lowe et al., 2006, Mintz & Betz, 1988; O?Dea & Abraham, 2002; RESTRICTIVE EATING RISK FOR SUICIDE ! 12 Striegel-Moore et al., 2009), we recruited college students who endorsed engaging in restrictive eating (i.e., fasting for eight hours or more while awake in order to influence shape or weight). In order to study the unique relationship betwen restrictive eating and the theory constructs, we also collected information on use of non-restrictive dieting methods (e.g., purging, excesive exercise), so that we could use these as control variables. Additionaly, we collected data on a range of psychosocial variables that have a known asociation with dieting (e.g., depresion, self- estem, perfectionism, and problems in interpersonal relationships; Ackard et al., 2002; Crow et al., 2006; Crow et al., 2008; French & Jefery, 1994) in order to characterize the psychosocial correlates of dietary restrictors, and to control for these variables. Based on previous research on psychosocial correlates and suicidal behavior asociated with dieting, and on preliminary evidence using the IPTS to study restrictive eating, we hypothesized that: 1. Consistent with prior studies on psychosocial correlates of individual who diet, colege students who endorsed restrictive eating over the past month (i.e., dietary restrictors) would have lower self-estem, fewer social supports and les satisfaction with this social support, more depresive symptoms, higher maladaptive perfectionism, higher levels of concern for weight, body shape, and eating, higher dietary restraint, and lower body mas index (BMI) than control participants not engaging in any form of dieting. 2. Dietary restrictors would have increased levels of perceived burdensomenes and thwarted belongingnes compared to controls. However, we predicted that upon controlling for other psychosocial correlates, the levels of perceived burdensomenes and thwarted belongingnes would not difer betwen the groups. RESTRICTIVE EATING RISK FOR SUICIDE ! 13 3. Dietary restrictors would have higher scores on a measure of suicidal ideation than controls. However, after controlling for other psychosocial variables (including thwarted belongingnes and perceived burdensomenes), we predicted that the diference betwen the two groups on suicidal ideation would no longer be significant. Additionaly, dietary restrictors would have higher likelihood of past suicide atempts compared to controls. 4. Dietary restrictors would have higher acquired capability for suicide (i.e., pain tolerance and fearlesnes about death) than controls. However, we expected a stronger relationship betwen dietary restriction and pain tolerance than betwen dietary restriction and fearlesnes about death, given that restrictive eating is more likely to expose (and therefore, habituate) an individual to physical pain than to the concept of death. We predicted that these diferences would persist after controlling for psychosocial correlates. 5. Within the group of dietary restrictors, the frequency and number of lifetime restrictive methods used would have a positive correlation with both physical pain tolerance and fearlesnes about death. RESTRICTIVE EATING RISK FOR SUICIDE ! 14 Method Participants and Procedure Participants were Auburn University undergraduates recruited through the SONA Human Subject Pool Software. Undergraduates were alowed to participate if over 18 years of age, and participants under age 19 had to provide parental consent to participate. Participants were able to sign up for the study online through SONA if they met inclusion criterion in a brief online screener by endorsing the item within the past month, have you gone for long periods of time (8 hours or more) while awake without eating anything in order to influence your shape or weight? Along with this dietary restrictors group, we recruited a second group of control participants with similar demographics (i.e., age, gender, race, and ethnicity). In an alternate SONA study signup, control individuals were alowed to participate if they did not endorse engaging in any eating pathology (i.e., fasting, self-induced vomiting, excesive exercise, laxative/diuretic use, binge eating) over their lifetime. Additionaly, al participants (both dietary restrictors and controls) were subject to certain exclusion criteria, in order to ensure acurate and consistent measures of pain tolerance as wel as to protect participants from possible negative outcomes from the pain tolerance asesment. We excluded participants currently taking any over-the-counter or prescription medications that increase the chance of bleding (e.g., blood thinners such as aspirin, Coumadin, or Plavix; National Institutes of Health, 2012) as wel as individuals with bleding disorders (e.g., hemophilia). Additionaly, to reduce errors due to pain insensitivity, participants were excluded if they had consumed alcohol within one hour of participation or taken any type of analgesic within eight hours of participation. Similarly, because of the RESTRICTIVE EATING RISK FOR SUICIDE ! 15 relationship betwen smoking and reduced pain sensitivity (Pomerleau, Turk, & Fertig, 1984), non-smoking participants were selected. Following several incidents that occurred when using the algometer in the current study and in another study being conducted in our research laboratory, we also began to exclude participants who had a history of seizures and fainting midway through data collection, in order to reduce risk of potential harm to participants who may have these preexisting conditions. Prior to collecting data, we utilized G*Power version 3.1.2 (Erdfelder, Faul, & Buchner, 1996) to conduct a power analysis to determine the number of participants to recruit based on a priori hypotheses. For Hypothesis 5, which only includes the dietary restrictors, in order to have .80 power to detect a correlation of r = .30, we determined that a minimum of 84 dietary restrictors were needed. This analysis brought the total number of participants, combining both the dietary restrictor and control groups, needed to 168, which would give us .89 power to detect a medium efect size (f 2 = .25) for a MANCOVA with up to 10 covariates. Based on this analysis, the final sample used in the current study included 193 (n = 99 dietary restrictors; n = 94 controls) mostly female (83.90%) participants, whose mean age was 19.69 (SD = 1.97; range from 17 to 35). Most participants were non-Latino, White/European origin (90.20%); 7.80% were Black/African American; 2.10% were Latino/Hispanic; 1.00% were Asian American/Pacific Islander; 1.00% identified as multi-racial or other. Demographics for both groups (i.e., dietary restrictors and controls) were highly similar (se Table 1 for demographic information on both groups). Upon determining eligibility, al participants (dietary restrictors and controls) were invited to participate in the study in our research laboratory. In the laboratory, they completed two sets of tasks. The first task was a batery of measures on a laboratory computer including RESTRICTIVE EATING RISK FOR SUICIDE ! 16 use, frequency, and history of dietary behavior, current and lifetime suicidal behavior, and other psychosocial variables. The second task included measurement of physical pain tolerance, height, and weight, which were conducted by a trained undergraduate research asistant. Participants were counterbalanced for the order of these sets of tasks. Al participants were compensated for their participation through extra credit points for their respective psychology courses through the SONA Human Subject Pool Software. Measures Dietary behavior and attitudes. The Eating Disorder Examination?Questionnaire Version 6 (EDE-Q; Fairburn & Beglin 1994; Fairburn & Beglin, 2008) is a 28-item, self-report version of the EDE-Interview (Fairburn & Cooper, 1993). This questionnaire measures core atitudes about and symptoms of disordered eating over the past 28 days on four subscales: Restraint, Weight Concern, Shape Concern, and Eating Concern. The questionnaire also assesses for the occurrence and frequency of specific behavior such as binge eating, purging, using laxatives, using diuretics, and engaging in hard exercise with open-ended response questions (e.g., Over the past 28 days, how many times have you made yourself sick (vomit) as a means of controlling your shape or weight?). A systematic review of past research found aceptable internal consistency, test-retest reliability, and construct validity for the EDE-Q (Berg, Peterson, Frazier, & Crow, 2011). In the current study, al subscales scores of the EDE-Q were used to compare the atitudes and concerns of dietary restrictors compared to controls. In addition, endorsement of item 2 (i.e., During the past 28 days, have you gone for long periods of time [8 waking hours or more] without eating to influence your shape or weight?) was used to include participants in the dietary restrictor group. Because the Restraint subscale also includes the item asesing for fasting, which was used to group participants, the Restraint subscale total scores RESTRICTIVE EATING RISK FOR SUICIDE ! 17 were calculated without this item in analyses. Internal consistencies for the four subscales were al in the aceptable range for both groups in the current sample (? values ranged from .80-.88 for dietary restrictor group; ? values ranged from .68-.89 for control group), with the exception of the Eating Concern subscale, which had lower internal consistency (? = .34) within the control group. Upon inspection of item responses for this subscale, it became clear that a majority of participants (67% of control sample) endorsed the lowest score possible for al items of this subscale. Because restriction of range atenuates measures of asociation (e.g., Pearson?s correlation; Cronbach?s alpha; Sacket & Yang, 2000), it is likely that the lower alpha value reflects the relatively low degree of eating pathology within the control group, which was the goal of recruitment. Curent restrictive dietary methods. In addition to the item from the EDE-Q asesing for current (i.e., past 28 days) fasting, additional items generated for this study were used to ases current use of other restrictive methods. Descriptive statistics on the endorsement of these current restrictive method items within the dietary restrictor group can be found in Table 2. Participants were asked to indicate the number of times over the past 28 days that they have skipped meals, or used other dieting methods found by Mintz and Betz (1988) to be restrictive dieting practices. These items clarified the extent to which participants in the dietary restrictor group have recently utilized other forms of restrictive eating, in addition to fasting for eight hours or more. These additional restrictive methods included: eat low calorie food, eat on a special diet, count calories, and stop eating certain foods (se Table 3). For current endorsement, participants were provided open-ended responses to items 2-7 in Table 3 to indicate the number of days over the past 28 days the restrictive method was used. RESTRICTIVE EATING RISK FOR SUICIDE ! 18 Curent non-restrictive dietary methods. Other items from the EDE-Q asesing for current (i.e., past 28 days) endorsement of other non-restrictive dietary methods (i.e., self- induced vomiting, laxative use, excesive exercise, and binge eating) were also included for descriptive purposes. Descriptive statistics on the endorsement of current non-restrictive methods within the dietary restrictor group can be found in Table 2. Lifetime restrictive dietary methods. Lifetime use of restrictive methods was asesed using items generated for this study. Participants were asked to indicate the number of times over their lifetime that they have fasted, skipped meals, eaten low calorie food, eaten on a special diet, reduced calorie intake, counted calories, and stopped eating certain foods (se Table 3). These items were rated on a 10-point scale, with higher scores indicating higher frequency of behavior. The total number of restrictive methods a participant has ever used was calculated using data from the items in Table 3. If the participant endorsed using one of these methods (e.g., skipping meals, counting calories, etc.) over his/her lifetime, the item was coded 1; if he/she did not use a certain method over the lifetime, the item was coded 0. Then, these coded scores were summed to create a total score for number of lifetime restrictive eating practices. Total score of lifetime restrictive dietary methods and specific frequency of each method (based on 10-point scale) over the lifetime were used to test Hypothesis 5. Furthermore, the item asesing for lifetime fasting was used as one of the exclusion criteria for the control group. Descriptive statistics on the endorsement of these lifetime restrictive method items within the dietary restrictor group can be found in Table 2. Lifetime non-restrictive dietary methods. Additionaly, items created for this study were used to determine lifetime occurrence and frequency of non-restrictive dietary methods (se Table 4). Using items from this table (excluding item 5, which aseses for binge eating), a total RESTRICTIVE EATING RISK FOR SUICIDE ! 19 score for use of non-restrictive dietary methods over the lifetime was derived from items on this questionnaire using the same procedure described earlier in calculating number of restrictive eating methods used, in order to control for lifetime use of non-restrictive dietary methods in analyses. These items were also rated on a 10-point scale, with higher scores indicating higher frequency of behavior. The item asesing for frequency of binge eating episodes was used for descriptive purposes. Descriptive statistics on the endorsement of these additional non-restrictive dietary methods within the dietary restrictor group can be found in Table 2. As described earlier (se Participants and Procedure), in order to be included in the control group, participants must not have endorsed lifetime engagement in self-induced vomiting, excesive exercise, laxative/diuretic use, or binge eating; thus, these created items were used as exclusion criteria for the control group. Suicidal behavior. The Beck Suicide Scale (BSS; Beck & Ster, 1991) is a frequently used, 21-item self-report measure of suicide ideation and desire. Items are rated on a 0-2 scale, and higher scores on the first 19 items indicate increased suicidal ideation and possible intent. Total score from the first 19 items were used as a measure of suicidal ideation in the current study. Items 20 and 21 ases for the frequency and lethality of an individual?s past suicide atempts (9.10% of dietary restrictor group reported a past suicide atempt; 0.00% of control group). Item 20 was used to test group diferences in suicide atempt history for Hypothesis 3. Numerous studies have demonstrated the sound psychometric properties of the BSS, including its internal consistency and concurrent validity (e.g., Beck & Ster, 1991; Beck, Ster, & Ranieri, 1988). Internal consistencies for items 1-19 were in the aceptable range for both groups in the current sample (? = .88 dietary restrictor group; ? = .79 control group). RESTRICTIVE EATING RISK FOR SUICIDE ! 20 Fearlesnes about death. The Acquired Capability for Suicide Scale (ACSS; Van Orden et al., 2008) was originaly developed as a 20-item self-report measure to ases an individual?s fearlesnes about death (e.g., I am not at all afraid to die) and pain insensitivity (e.g., I can tolerate a lot more pain than most people). Items are rated on a five-point Likert scale, and higher scores indicate a greater acquired capability for suicide. For the purposes of this study, a shorter seven-item version of the ACSS was utilized that specificaly aseses fearlesnes about death. Support for this shorter version of the ACSS was generated from a recent factor analytic study, which found that these seven items demonstrate aceptable convergent and discriminant validity (Ribeiro et al., 2014). Internal consistency for this subscale was in the aceptable range for both groups in the current sample (? = .88 dietary restrictor group; ? = .77 control group). Physical pain tolerance. Physical pain tolerance was measured using a Medoc Algomed presure algometer. A trained undergraduate research asistant applied the algometer to the participant?s hand betwen the second and third finger. The experimenter gradualy increased the amount of presure applied at a rate of five kilopascals per second, and participants were instructed to depres an indicator button at the point at which the pain became too uncomfortable to continue. If participants reached the maximum presure safety cutoff (i.e., 2000 kilopascals) before indicating that the pain was too uncomfortable to continue, then the experimenter manualy discontinued the trial. The amount of presure that was applied at the point when the participant stopped the trial, or when the maximum presure cutoff was reached, was recorded as an indicator of pain tolerance. Participants completed five separate trials, and the values from each trial were averaged to calculate pain tolerance. In order to prevent habituation, there was a 60 second interval betwen each pain tolerance measurement (Orbach et al., 1997). Internal RESTRICTIVE EATING RISK FOR SUICIDE ! 21 consistency of the pain tolerance trials was adequate (? = .98 dietary restrictor group; ? = .98 control group). Additionaly, because of past research demonstrating asymmetry in pain tolerance due to handednes (Gobel & Westphal, 1987; Murray & Hagan, 1973; Pauli, Wiedemann, & Nicola, 1999), we collected information on participants? handednes to control for this variable, and pain tolerance measurements were conducted on al participants? right hands. This technique has been used as a reliable measure of physical pain tolerance in studies of the IPTS (Bender et al., 2011; Wite, Gordon, Smith, & Van Orden, 2012). Perceived burdensomenes and thwarted belongingnes. The Interpersonal Neds Questionnaire (INQ-R; Van Orden et al., 2008; Van Orden, Cukrowicz, Wite, & Joiner, 2012) is a 15-item, self-report scale that is used to measure a respondent?s current beliefs about feling connected to others (i.e., thwarted belongingnes) and feling like a burden (i.e., perceived burdensomenes). Thwarted belongingnes is measured with a nine-item subscale, and perceived burdensomenes is measured with a six-item subscale. Items are rated on a seven-point Likert scale, with higher ratings indicating higher levels of the constructs. Previous studies have found aceptable psychometric properties for both subscales (Bryan, Cukrowicz, West, & Morrow, 2010; Bryan, Clemans, & Hernandez, 2012; Van Orden et al., 2012). In the current study, internal consistency for the belongingnes subscale (? = .90 dietary restrictor group; ? = .90 control group) and burdensomenes subscales (? = .88 dietary restrictor group; ? = .92 control group) were found to be in the aceptable range for both groups. Perfectionism. The Almost-Perfect Scale Revised (APS-R; Slaney et al., 2001) consists of 23 items, which are rated on a seven-point Likert scale measuring perfectionism on three subscales: Standards, Order, and Discrepancy. Consistent with recent studies (e.g., Rice, Choi, Zhang, Morero, & Anderson, 2012; Rice, Richardson, & Clark, 2012), we used only the RESTRICTIVE EATING RISK FOR SUICIDE ! 22 Discrepancy subscale. The Discrepancy subscale measures the self-critical, maladaptive aspect of perfectionism and is comprised of 12 items (e.g., Doing my best never sems to be good enough). Slaney et al. (2001) found aceptable levels of internal consistency, reliability, and construct validity for all APS-R subscales. Internal consistency for this subscale was in the aceptable range for both groups in the current sample (? = .94 dietary restrictor group; ? = .95 control group). Depresive symptoms. The Center for Epidemiological Studies?Depresion Scale (CES-D; Radloff, 1977) is a 20-item scale measuring depresive symptoms experienced within the past wek (e.g., I felt sad). Items are rated on a three-point Likert scale; higher ratings indicate increased depresive symptomatology. The CES-D has good internal consistency, test- retest reliability, and has been shown to correlate with clinical ratings of depresion severity, atesting to its construct validity (Radloff, 1977). In the current study, internal consistency was found to be in the aceptable range for both groups (? = .90 dietary restrictor group; ? = .89 control group). Self-esteem. The State Self-Estem Scale (SSES; Heatherton & Polivy, 1991) is a 20- item, self-report scale that asks respondents to rate themselves based on how they have felt within the past wek. Items are rated on a five-point Likert scale and include items such as I fel confident about my abilities. Research has shown that the SSES divides into three subscales: Appearance, Social, and Performance (Heatherton & Polivy, 1991). The SSES has demonstrated aceptable internal consistency and adequate test-retest reliability for each subscale and the total score of the SSES (Heatherton & Polivy, 1991). For the purposes of this study, only the total score for self-esteem was utilized, with higher total scores indicating higher self-esteem. Internal RESTRICTIVE EATING RISK FOR SUICIDE ! 23 consistency was found to be in the aceptable range for both groups (? = .91 dietary restrictor group; ? = .91 control group). Social suport. The Social Support Questionnaire (SSQ-6; Sarason, Sarason, Shearin, & Pierce, 1987) is a self-report measure with six items asking about respondents? perceived number of social supports, and their satisfaction with their current level of social support. Respondents are asked to list people in their lives who fit certain descriptions (e.g., Who can you really count on to be dependable when you need help?) and to rate their overal level of social support on a six-point Likert scale for each description. The SSQ-6 yields two separate scores: a Number of Availability score, based on the number of people listed per item, and a Satisfaction score, based on ratings of overal social support for each item, both of which were used in our analyses. The SSQ-6 has demonstrated good reliability and validity (Sarason, Shearin, Pierce, & Sarason, 1987). In the current sample, internal consistency was adequate for both the Number of Availability scale (? = .92 dietary restrictor group; ? = .93 control group) and Satisfaction scale (? = .98 dietary restrictor group; ? = .96 control group). Height and weight. Height and weight were asesed in the laboratory using a digital scale and stadiometer (i.e., height measuring rod). This information was used to calculate the body mas index (BMI) of al participants using the formula provided by the Center for Disease Control and Prevention (2011). Demographics. Additional information such as gender, ethnicity, race, and age was collected for the purposes of collecting demographicaly similar dietary restrictor and control groups. RESTRICTIVE EATING RISK FOR SUICIDE ! 24 Results Al analyses were performed using IBM SPS version 21. Mising data were minimal, and mising value patern analysis in SPS Version 21 revealed that almost al variables had less than 5% mising data. Mising data were handled using multiple imputation in SPS. Based on the amount of mising data, we followed recommendations by Bodner (2008) and imputed six datasets, running the analyses in each imputed dataset. Whereas pooled results are available for some analytic techniques (e.g., regresion) when using multiple imputation, no pooled results are produced by SPS for analysis of covariance (ANCOVA) or multivariate analysis of covariance (MANCOVA) models, which were the main focus of our analyses. Thus, we instead inspected the output from the six imputed datasets to determine consistency in results for our MANCOVA analyses and reported only the results from the original (i.e., non-imputed) dataset. Output from the imputed datasets indicated that after acounting for minimal mising data, results were highly similar to results from the original dataset for al analyses. Preliminary Analyses Intercorrelations betwen al study variables can be found in Tables 5 and 6. As indicated previously (se Measures section), descriptive statistics on current and lifetime endorsement of restrictive and non-restrictive methods within the dietary restrictor group can be found in Table 2. Within the dietary restrictor group, use of a variety of restrictive methods currently and over the lifetime was typical (i.e., over lifetime, the least frequently used method of eating on a special diet was stil endorsed by 85% of the sample), and on average, dietary restrictors RESTRICTIVE EATING RISK FOR SUICIDE ! 25 endorsed fasting relatively frequently over the past month (M = 1.80; based off Likert scale 1 = 1-5 days; 2 = 6-12; 3 = 13-15; 4 = 16-22; 5 = 23-27; 6 = every day). Additionaly, endorsement of non-restrictive methods was relatively common over the past month and over the lifetime within the dietary restrictor group. Over the past month, the most common method endorsed was excesive exercise (69% of sample), whereas fewer participants endorsed other non-restrictive methods (i.e., self-induced vomiting, laxative use; 7- 8%), and 60% of the sample endorsed binge eating. Over the lifetime, the most common method endorsed also was engaging in excesive exercise (86% of sample). Fewer participants endorsed other non-restrictive methods like self-induced vomiting, diuretic, and laxative use (22-30% of sample), and a majority of the sample (69%) endorsed engaging in binge eating over the lifetime. Hypothesis 1: Group Differences in Psychosocial Variables To test Hypothesis 1, a MANCOVA was performed to compare group means for the dietary restrictor and control groups on a collection of psychosocial measures. The total scores for self-estem, number of social supports, satisfaction with social supports, depresive symptoms, maladaptive perfectionism, weight, shape, and eating concern, dietary restraint, and BMI were entered into a MANCOVA, with total score for lifetime use of non-restrictive dietary practices as a covariate. Results of this analysis revealed a significant efect of dietary group on psychosocial variables, Wilk?s ? = 0.62, F(8, 179) = 13.72, p < .01, partial ? 2 = .60. As noted previously, due to the low internal reliability (? = .34 control group) of the Eating Concern subscale, results on this scale should be viewed with caution. We followed up this MANCOVA by examining the univariate results for each dependent variable (se Table 7). In order to reduce the risk of Type I error, we used Benjamini and Hochberg?s (1995) method for multiple comparison correction. The univariate results revealed RESTRICTIVE EATING RISK FOR SUICIDE ! 26 significant diferences betwen the dietary restrictor and control groups on most psychosocial variables, with dietary restrictors having higher depresive symptoms, higher maladaptive perfectionism, higher weight/shape/eating concerns, higher dietary restraint, higher BMI, and lower self-estem compared to controls. Efect sizes on scores of weight, shape, and eating concern, and dietary restraint were large (partial ? 2 = .37-.56), whereas medium efect sizes were observed for self-estem and maladaptive perfectionism (partial ? 2 = .22 and .13, respectively), and smal efect sizes were observed for group diferences on depresion and BMI (partial ? 2 = .08). The groups did not difer on number of social supports or satisfaction with these supports (total supports and satisfaction; partial ? 2 = .01-.02). Hypothesis 2: Group Differences in Thwarted Belongingnes and Perceived Burdensomenes To test Hypothesis 2, a MANCOVA framework was utilized to compare dietary restrictors with controls on measures of perceived burdensomenes and thwarted belongingnes with total score for lifetime use of non-restrictive dietary practices as a covariate. Results of this analysis (se Table 7) indicated no significant efect of dietary group on thwarted belongingnes or perceived burdensomenes, Wilk?s ? = 0.99, F(2, 188) = .84, p = .44, partial ? 2 = .01. Because no group diferences emerged in this MANCOVA, we did not pursue follow-up analyses that were intended to test changes in group diferences after controlling for psychosocial variables. Hypothesis 3: Group Differences in Suicidal Ideation and Attempts In order to investigate the relationship betwen dietary restriction and suicidal ideation, we conducted an ANCOVA, with the total score for suicidal ideation as the dependent variable and lifetime non-restrictive dieting practices as a covariate. Results of this analysis (se Table 7) revealed non-significant efects of dietary group on levels of suicidal ideation, F(1, 190) = 2.46, RESTRICTIVE EATING RISK FOR SUICIDE ! 27 p = .12, partial ? 2 = .01. Because no group diferences emerged in this ANCOVA, we did not pursue follow-up analyses that were intended to test changes in group diferences after controlling for psychosocial variables. To test group diferences in suicide atempt history, we conducted a chi-square analysis, using a dichotomous variable for presence or absence of suicide atempt history. Results of this analysis indicated that there is a significant relationship betwen suicide atempt history and dietary group, ? 2 (1, N = 193) = 7.92, p = .00. Dietary restrictors were more likely to have a suicide atempt history than controls (i.e., 9.10% versus 0.00%). Hypothesis 4: Group Differences in ACS For Hypothesis 4, total scores for fearlesnes about death and pain tolerance were entered into a MANCOVA, with total score for lifetime use of non-restrictive dietary practices as a covariate. Results of this analysis (se Table 7) indicated no significant efect of dietary group on either facet of ACS, Wilk?s ? = 0.99, F(2, 175) = .56, p = .57, partial ? 2 = .01. Because no group diferences emerged in this MANCOVA, we did not pursue follow-up analyses that were intended to test changes in group diferences after controlling for psychosocial variables. In order to examine any efect of participant handednes in the relationship betwen dietary restriction and pain tolerance, we conducted a separate ANCOVA, including handednes and lifetime use of non-restrictive dietary practices as covariates. Results of this analysis indicated that even after controlling for handednes, no significant efect of dietary group emerged on levels of pain tolerance, F(1, 175) = .10, p = .75. Hypothesis 5: ACS and Use of Restrictive Dietary Methods To investigate Hypothesis 5, we used data from the dietary restrictors group only (n = 99) to examine whether the total score for restrictive methods (i.e., summed total of number of RESTRICTIVE EATING RISK FOR SUICIDE ! 28 restrictive methods used over the lifetime) and/or reported frequencies of each of these methods over the lifetime (i.e., responses to lifetime items from Table 3) were correlated with physical pain tolerance and fearlesnes about death scores. Pooled correlations, using the imputed datasets, can be found in Table 5, and demonstrate that neither total number of restrictive methods, nor frequency of those methods, were significantly correlated with either facet of ACS. Additionaly, we examined the correlation betwen current restrictive methods (i.e., past 28 days) and both facets of ACS, which did not reveal any significant correlations (se Table 5). We also conducted exploratory analyses to investigate correlations betwen current and lifetime non-restrictive dietary methods (e.g., purging, laxative use, etc.), total non-restrictive methods, and ACS, which are also found in Table 5. Results of these correlations indicate that most current or lifetime non-restrictive methods, and total score for use of lifetime non- restrictive methods were not significantly correlated with either facet of ACS. However, current endorsement of excesive exercise and binge eating were significantly correlated with pain tolerance (r?s = .22 and .24, respectively). Post-hoc Analyses Due to high endorsement of other non-restrictive dietary methods within the dietary restrictor group, we were concerned that controlling for the lifetime non-restrictive methods variable may have removed important variance from our MANCOVA analyses. Thus, we conducted post-hoc analyses to examine group diferences without including total score for lifetime non-restrictive methods as a covariate. Results of these post-hoc analyses revealed that for al study hypotheses, the patern of results was the same. That is, no new group diferences emerged after removing this covariate, and existing group diferences were consistent after removing this covariate. RESTRICTIVE EATING RISK FOR SUICIDE ! 29 Discusion The current study integrated and expanded upon two existing lines of research: findings establishing the relationship betwen dieting and suicidal behavior, and research on the link betwen restrictive eating and ACS. We sought to addres limitations in previous research by focusing on one restrictive eating method (i.e., fasting) within a non-clinical sample, in order to isolate and examine this behavior within the framework of the IPTS. Secondarily, we sought to characterize psychosocial diferences betwen dietary restrictors and individuals endorsing no eating pathology over the lifetime. In line with hypotheses and consistent with previous research (e.g., Ackard et al., 2002; Crow et al., 2006; Crow et al., 2008; French & Jefery, 1994), significant group diferences emerged betwen dietary restrictors and controls on multiple psychosocial variables, including depresive symptoms, self-estem, maladaptive perfectionism, and concern for shape/weight/eating, and dietary restraint. In terms of characterizing the dietary restrictor group in the current study, these findings demonstrate the increased pathology within this group compared to individuals with similar demographics who endorse neither restrictive nor non- restrictive dietary methods over the lifetime. Moreover, these findings lend additional support to research indicating that even within a non-clinical sample, symptoms of disordered eating are asociated with adverse social and emotional consequences, such as elevated depresive symptoms and lower self-esteem (e.g., Ackard, Fulkerson, & Neumark-Sztainer, 2011; Swanson, Crow, LeGrange, Swendsen, & Merikangas, 2011). Although not in the predicted direction, the groups also difered on body mas index (BMI), with dietary restrictors having higher BMI, on RESTRICTIVE EATING RISK FOR SUICIDE ! 30 average, compared to controls in the current sample. This finding is in line with past research suggesting that dieting may actualy be related to weight gain, especialy among college students (Lowe et al., 2006). Results testing group diferences on thwarted belongingnes and perceived burdensomenes provided mixed support for hypotheses. That is, although we predicted that there would be no significant diferences betwen the dietary restrictor and control groups on these constructs after acounting for other psychosocial variables (i.e., depresive symptoms; self-estem; perfectionism; social support), we anticipated that diferences would emerge prior to controlling for these variables. Contrary to this prediction, no diferences betwen groups emerged, even without controlling for other psychosocial variables. Although many psychosocial variables measured in the current study were elevated within the dietary restrictor group, it may be that psychosocial diferences observed reflect distinct pathology not necesarily contributing to perceived burdensomenes and thwarted belongingnes. Notably, no diferences were evident in levels of social support (i.e., number of social supports or satisfaction with this support) betwen the two groups. Because level of stability in interpersonal relationships contributes both to thwarted belongingnes (i.e., level of social connectednes) and perceived burdensomenes (i.e., presence of reciprocal care in relationships), it may be that the similar levels of social support, despite other diferences in psychosocial functioning, contributed to similarities on these IPTS constructs. Indeed, within both groups, we found smal to moderate negative bivariate correlations betwen social support variables (i.e., number of social support and satisfaction with this support) and IPTS constructs. Thus, if groups had comparable levels of social support, which sems to most clearly tap both IPTS constructs compared to other psychosocial variables, this finding may help to explain the observed similarities on levels of thwarted belongingnes and RESTRICTIVE EATING RISK FOR SUICIDE ! 31 perceived burdensomenes. Moreover, because social isolation is also one of the strongest predictors of suicidal ideation (Joiner & Van Orden, 2008; Van Orden et al., 2010), similarities betwen the groups on the measure of social support may also be related to similarities on suicidal ideation. In fact, review of past research has indicated that a greater number of reported family and friends is asociated with decreased risk for lethal suicidal behavior (Van Orden et al., 2010). In support of this notion, results were also inconsistent with predictions on group diferences in suicidal ideation, as the groups did not difer on this variable. Although past research has established increased risk for suicidal behavior among individuals who endorse disordered eating or, more broadly, dieting practices, this research has been limited by imprecise measurements of dieting, which may acount for diferences in the results of the current study. That is, past studies have been limited by lack of operational definitions of what constitutes dieting (e.g., Rafiroui et al., 2003; Timko et al., 2006), and have not sought to investigate specific dietary behavior (e.g., restrictive eating) and suicidal behavior, as was the goal of the current study. However, as indicated by post-hoc analyses, even after removing the covariate that controlled for lifetime use of non-restrictive methods, no group diferences emerged on suicidal ideation. Moreover, diferences in the samples utilized in previous studies may have contributed to discrepant findings. For example, most prior research on suicidal ideation and dieting has focused on adolescent populations (e.g., Crow et al., 2008; Neumark-Sztainer et al., 1998; Rafiroui et al., 2003), whereas our sample was recruited from a college student population. Compared to college students, adolescents who engage in dieting may represent a more pathological group. Additionaly, the prevalence of suicidal ideation endorsed by samples in previous studies was higher than in the current study. For example, among participants endorsing RESTRICTIVE EATING RISK FOR SUICIDE ! 32 more severe disordered eating (e.g., self-induced vomiting, laxative/diuretic use), Crow and colleagues (2008) found that betwen 38-55% of these participants endorsed suicidal ideation. Similarly, in a group of individuals endorsing self-induced vomiting and diet pil usage, Neumark-Sztainer and colleagues (1998) found that 34-62% of participants endorsed suicidal ideation. In our sample of dietary restrictors, only 18.18% endorsed suicidal ideation. Thus, diferences in the current study results may also reflect a restriction of range in our measure of suicidal ideation. Additionaly, it is likely that with the larger samples utilized in past studies (e.g., N = 16,296; Neumark-Sztainer et al., 1998) significant diferences on suicidal ideation betwen dieting individuals versus non-dieting would be more easily detected, though they actualy represent small efect sizes similar to those of the current study. Consistent with hypotheses, dietary restrictors were more likely to have a suicide atempt history than controls. This finding is in line with past research indicating the higher risk for suicide atempt among individuals engaging in dietary methods (Crow et al., 2008; Neumark- Sztainer et al., 1998). This finding is also consistent with prior research establishing the unique relationship betwen restrictive eating and suicide atempts, but not suicidal ideation (Wite et al., 2012). Based on past research (e.g., Holm-Denoma et al., 2008; Selby et al., 2010; Wite et al., 2012), we sought to investigate the postulation that engaging in certain types of disordered eating (i.e., restrictive eating) may be influencing an individual?s capability to enact lethal self- harm, which was the first test of the relationship betwen restrictive eating and measures of ACS. However, when considering group diferences on facets of ACS (i.e., fearlesnes about death and physical pain tolerance), which is proposed to be the mechanism by which lethal suicide atempts occur (Van Orden et al., 2010), no group diferences emerged. As such, the findings that no diferences exist betwen dietary restrictors and controls and that frequency of RESTRICTIVE EATING RISK FOR SUICIDE ! 33 current or lifetime restrictive eating within the dietary restrictor group was not related to either facet of ACS, suggest that engaging in repeated fasting is likely not functioning as a painful and/or fear-inducing event contributing to heightened ACS. By collecting information on and controlling for the presence of other non-restrictive dietary methods, the current study represents the first atempt at isolating and examining the relationship betwen one specific behavior (i.e., restrictive eating) and ACS. However, although it was our goal to isolate the behavior of restrictive eating in order to investigate its relationship with ACS, there was a high frequency of other types of non-restrictive dietary methods within the restrictor group. That is, although the target sample of the study was to recruit participants who only engage in restrictive eating, it does not appear that this behavior occurs in isolation, and, based on the frequencies of other non-restrictive dietary methods in our sample, it sems that other methods of disordered eating are prominent over the lifetime and currently as wel. For this reason, it is dificult to completely parse out the relationship betwen restrictive eating on risk for suicide. As such, group diferences on likelihood of suicide atempt do not necesarily reflect diferences in restrictive eating betwen the two groups. Notably, even with the unanticipated presence of non-restrictive methods in our dietary restrictor sample, many of which have also been proposed as painful and provocative experiences that could contribute to ACS, we stil did not find diferences betwen this group and controls on ACS. Moreover, the dietary restrictor group had fairly extensive histories using these other methods, and in considering the bivariate correlations betwen current and lifetime non-restrictive methods and ACS, no significant asociations emerged for lifetime methods, despite sufficient power in the study to detect medium-sized correlations. However, in considering current non-restrictive methods, significant asociations emerged betwen excesive RESTRICTIVE EATING RISK FOR SUICIDE ! 34 exercise and binge eating and the pain tolerance facet of ACS. This finding is consistent with other recent research within the framework of the IPTS which found that engaging in over- exercise over the past month was asociated with higher levels of ACS (Smith et al., 2012). Because there was no asociation betwen lifetime binge eating and excesive exercise and pain tolerance, it may be that recency of this behavior contributes diferentialy to ACS. That is, because presence of painful and/or fear-inducing events need to be particularly salient and severe in order to contribute to increases in ACS, it may that time since occurrence of these events would afect the degree to which they influence ACS. Although the IPTS proposes that ACS is a relatively stable construct once it has developed (Van Orden et al., 2010), there may diferential contributions to ACS based on time since event, which represents an important avenue for future research. However, related to the hypotheses of the study, it appears that for fasting, both current and lifetime endorsement of this behavior were not significantly related to ACS, even when this behavior was occurring recently and frequently. Based on non-significant relationships betwen both current and lifetime restrictive eating and ACS, one possibility is that engaging in restrictive eating may not pose the same level of potency and/or threat as other events proposed to contribute to ACS (e.g., combat exposure; childhood abuse; Van Orden et al., 2010), and that engaging in more frequent, but les severe, behavior afects ACS diferently. Indeed, the diferences in body mas index betwen the groups in our study (i.e., control group having lower BMI than the dietary restrictors) suggest that the restrictive eating endorsed by dietary restrictors in our sample was not severe enough as to result in the significant weight loss and/or unhealthy low weight that is characteristic of restrictive eating in AN, which may have afected the influence of this behavior on ACS. It sems that the severe restrictive eating and asociated low weight sen in clinical AN samples may confer RESTRICTIVE EATING RISK FOR SUICIDE ! 35 specific risk for suicide, beyond the efects of restrictive eating in a non-clinical, mostly healthy weight sample of individuals who endorse restrictive eating. To beter parse out the efects of types of dietary methods on ACS, future studies should also investigate non-clinical individuals engaging in only non-restrictive dietary methods (e.g., purging, excesive exercise) to compare their levels of ACS to individuals who engage in restrictive dietary methods. Investigation of this behavior within a non-clinical sample would help to control for asociated factors of clinical disorders that may confound understanding of how specific behavior can contribute to risk for suicide. Additionaly, in considering the facet of physical pain tolerance, dietary restrictors may not have demonstrated higher pain responses based on the measurement of physical pain tolerance used in the current study (i.e., presure algometer). Although some past research examining pain sensitivity among individuals with diagnoses of AN (i.e., those who engage in severe restrictive eating) has used measures of presure pain similar to the one utilized in the current study (e.g., Raymond et al., 1999) and has found diferences betwen AN and control subjects, other research has found these diferences utilizing a thermal pain measure (e.g., de Zwan et al., 1996; Lautenbacher et al., 1991). The use of diferent devices to measure pain may be especialy important to consider given that pain tolerance is proposed by the IPTS to likely be method-specific (Van Orden et al., 2010). Thus, as described by Van Orden and colleagues (2010), an individual may acquire the necesary pain tolerance to engage in one type of behavior (e.g., cutting), but may not have necesarily display the same tolerance for other types of behavior (e.g., taking a lethal number of pils). Although the behavior of interest in the current study, restrictive eating, has been demonstrated to inflict bodily harm, especialy in severe forms sen in the disorder anorexia nervosa (e.g., Crisp, Calender, Halek, & Hsu, 1992), it may be that RESTRICTIVE EATING RISK FOR SUICIDE ! 36 the type of pain tolerance to which this behavior contributes is not fully tapped by a measure of presure pain tolerance. It sems that the pain induced by engaging in restrictive eating (e.g., hunger pains) may reflect higher ability to sustain internaly-based pain, rather than pain involved in surface-level injury, which sems to be the type of pain that presure and thermal pain measurements inflict. As such, use of a method that might measure this type of internal pain may be beter suited for examining pain tolerance within disordered eating samples. For example, ischemic pain measurements, which involve using a blood-presure cuff or similar tourniquet on participants? arms and instructing participants to perform tasks with this arm (e.g., clenching fist; holding onto a handgrip) while sustaining this pain, have also been used in other studies patients with bulimia nervosa (e.g., Stein et al., 2003). Because this ischemic pain procedure involves application of more internally-based pain (similar to the type of pain sustained by engaging in restrictive eating), rather than skin/surface pain, this method may be beter able to measure the type of pain tolerance developed by those who engage in restrictive eating. Future studies should conduct a more nuanced exploration of these method-specific types of pain tolerance by incorporating other measurements of pain tolerance, which may help to further elucidate the relationship betwen restrictive eating and pain tolerance. Despite improvements and expansions upon past literature, this study had a number of limitations. First, the data are cross-sectional, which limits our ability to speak to causality and fully understand the specific efects of restrictive eating on IPTS constructs. Additionaly, the study sample was low in diversity (i.e., predominantly young, Non-Hispanic/Latino, White, female undergraduates). Therefore, results may not generalize to more diverse samples. As mentioned previously, although our measure of physical pain tolerance contributed to a multi- modal investigation of IPTS constructs, it also may be that this particular measure of pain RESTRICTIVE EATING RISK FOR SUICIDE ! 37 tolerance taps a type of pain tolerance that is not particularly relevant in examining restrictive eating. It may also be that measures of physical pain tolerance measure a diferent pain construct than do self-report measures of pain tolerance, which may reflect more of the cognitive appraisal aspect of pain. Although the IPTS proposes that pain tolerance is a key component of ACS, there is a dearth of literature examining which type of pain tolerance is a more potent predictor of lethal suicidal behavior, which warrants atention in future research, and may have influenced the observed similarities betwen groups in our study. Despite noted limitations, the application of the IPTS in the current study helped to integrate and improve upon two bodies of literature (i.e., dieting and suicidal behavior, and the IPTS and restrictive eating). Our design, including recruitment of two demographicaly similar samples of undergraduates, alowed for a detailed comparison betwen the groups and provides a characterization of the psychosocial functioning of non-clinical individuals who engage in restrictive eating. Additionaly, our focus on the specific behavior of restrictive eating, as opposed to clinical diagnostic status or a combined measure of disordered eating, addresed limitations in existing research, and alowed for a stringent test of some fundamental hypotheses of the IPTS. As such, our findings contribute to an enhanced understanding of the theoretical mechanisms by which disordered eating may confer increased risk for suicidal behavior. RESTRICTIVE EATING RISK FOR SUICIDE ! 38 References Ackard, D. M., Croll, J. K., & Kearney-Cooke, A. (2002). Dieting frequency among college females: Asociation with disordered eating, body image, and related psychological problems. Journal of Psychosomatic Research, 52(3), 129-136. Ackard, D. M., Fulkerson, J. A., & Neumark-Sztainer, D. (2011). Psychological and behavioral risk profiles as they relate to eating disorder diagnoses and symptomatology among a school-based sample of youth. International Journal of Eating Disorders, 44(5), 440-446. Anderson, D. A., Shapiro, J. R., & Lundgren, J. D. (2003). The freshman year of college as a critical period for weight gain: An initial evaluation. Eating Behaviors, 4(4), 363-367. Baker, E. H., & Sandle, G. I. (1996). Complications of laxative abuse. Annual Review of Medicine, 47, 127-134. Beautrais, A. L. (2001). Child and young adolescent suicide in New Zealand. Australian and New Zealand Journal of Psychiatry, 35, 647-653. Beck, A. T., & Ster, R. A. (1991). Manual for Beck Scale for Suicidal Ideation. San Antonio, TX: Psychological Corporation. Beck, A. T., Ster, R. A., & Ranieri, W. F. (1988). Scale for suicide ideation: Psychometric properties of a self-report version. Journal of Clinical Psychology, 44(4), 499-505. Bender, T. W., Gordon, K. H., Bresin, K., & Joiner, T. E. (2011). Impulsivity and suicidality: The mediating role of painful and provocative experiences. Journal of Afective Disorders, 129, 301-307. RESTRICTIVE EATING RISK FOR SUICIDE ! 39 Benjamini, Y., & Hochberg, Y. (1995). Controlling the false discovery rate: A practical and powerful approach to multiple testing. Journal of Royal Statistical Society, 57, 289-300. Berg, K. C., Peterson, C. B., Frazier, P., & Crow, S. J. (2011). Psychometric evaluation of the eating disorder examination and eating disorder examination-questionnaire: A systematic review of the literature. International Journal of Eating Disorder, 45(3), 428-438. Bodner, T. E. (2008). What improves with increased mising data imputations? Structural Equation Modeling, 15(4), 651-675. Brain, K. L., Haines, J., & Wiliams, C. L. (1998). The psychophysiology of self-mutilation: Evidence of tension-reduction. Archives of Suicide Research, 4, 227-242. Brown, G. K., Beck, A. T., Ster, R. A., & Grisham, J. R. (2000). Risk factors for suicide in psychiatric outpatients: A 20-year prospective study. Journal of Consulting and Clinical Psychology, 68(3), 371-377. Brownel, K. D., & Rodin, J. (1994). The dieting maelstrom: Is it possible and advisable to lose weight? The American Psychologist, 49(9), 781-791. Bryan, C. J., Clemans, T. A., & Hernandez, A. M. (2012). Perceived burdensomenes, fearlesnes of death, and suicidality among deployed military personnel. Personality and Individual Diferences, 52(3), 374-379. Bryan, C. J., Cukrowicz, K. C., West, C. L., & Morrow, C. E. (2010). Combat experience and the acquired capability for suicide. Journal of Clinical Psychology, 66(10), 1044-1056. Bunnel, D. W., Shenker, I. R., Nussbaum, M. P., Jacobson, M. S., & Cooper, P. (1990). Subclinical versus formal eating disorders: Diferentiating psychological features. International Journal of Eating Disorders, 9(3), 357-362. Cain, A. S., Bardone-Cone, A. M., Abramson, L. Y., Vohs, K. D., & Joiner, T. E. (2008). RESTRICTIVE EATING RISK FOR SUICIDE ! 40 Refining the relationships of perfectionism, self-eficacy, and stres to dieting and binge eating: Examining the appearance, interpersonal, and academic domains. International Journal of Eating Disorders, 41, 713-721. Center for Disease Control and Prevention. (2011). Adult BMI calculator. Retrieved from http:/ww.cdc.gov/healthyweight/asesing/bmi/adult_bmi/english_bmi_calculator/bmi _calculator.html Chamay-Weber, C., Narring, F., & Michaud, P. (2005). Partial eating disorders among adolescents: A review. Journal of Adolescent Health, 37(5), 417-427. Christiansen, E., & Jensen, B. F. (2007). Repetition of suicide atempt, suicide or al deaths after an episode of atempted suicide: A register-based survival analysis. Australian and New Zealand Journal of Psychiatry, 41(3), 257-265. Claes, L., Vandereycken, W., & Vertommen, H. (2006). Pain experience related to self-injury in eating disorder patients. Eating Behaviors, 7(3), 204-213. Crisp, A., Calender, J., Halek, C., & Hsu, L. (1992). Long-term mortality in anorexia nervosa. British Journal of Psychiatry, 161, 104-107. Crow, S., Eisenberg, M. E., Story, M., & Neumark-Sztainer, D. (2006). Psychosocial and behavioral correlates of dieting among overweight and non-overweight adolescents. Journal of Adolescent Health, 38(5), 569-574. Crow, S., Eisenberg, M. E., Story, M., & Neumark-Sztainer, D. (2008). Suicidal behavior in adolescents: Relationship to weight status, weight control behaviors, and body disatisfaction. International Journal of Eating Disorders, 41(1), 82-87. De Zwan, M., Biener, D., Bach, M., Wiesnagrotzki, S., & Stacher, G. (1996). Pain sensitivity, RESTRICTIVE EATING RISK FOR SUICIDE ! 41 alexithymia, and depresion in patients with eating disorders: Are they related? Journal of Psychosomatic Research, 41, 65-70. Downey, C. A., & Chang, E. C. (2007). Perfectionism and symptoms of eating disturbances in female college students: Considering the role of negative afect and body disatisfaction. Eating Behaviors, 8(4), 497-503. Duberstein, P. R., Conwel, Y., Conner, K. R., Eberly, S., Evinger, J.S., & Caine, E. D. (2004). Poor social integration and suicide: Fact or artifact? A case-control study. Psychological Medicine: A Journal of Research in Psychiatry and the Alied Sciences, 34(7), 1331- 1337. Erdfelder, E., Faul, F., & Buchner, A. (1996). GPOWER: A general power analysis program. Behavior Research Methods, Instruments, & Computers, 28, 1-11. Fairburn, C. G., & Beglin, S. J. (1994). Asesment of eating disorders: Interview or self-report questionnaire? International Journal of Eating Disorders, 16, 363-370. Fairburn, C. G., & Beglin, S. J. (2008). Eating disorder examination questionnaire. In C. G. Fairburn (Ed.), Cognitive Behavior Therapy and Eating Disorders (pp. 309-313). New York: Guilford Pres. Fairburn, C. G., & Cooper, Z. (1993). The eating disorder examination. In C. G. Fairburn & G. T. Wilson (Eds.), Binge eating: Nature, assesment, and treatment (12 ed.) (pp. 317-360). New York, NY: Guilford Pres. Fairburn, C. G., Cooper, Z., Doll, H. A., & Welch, S. L. (1999). Risk factors for anorexia nervosa: Three integrated case-control comparisons. Archives of General Psychiatry, 56(5), 468-476. French, S. A., & Jefery, R. W. (1994). Consequences of dieting to lose weight: Efects on RESTRICTIVE EATING RISK FOR SUICIDE ! 42 physical and mental health. Health Psychology, 13(3), 195-212. French, S. A., Jefery, R. W., & Murray, D. (1999). Is dieting good for you?: Prevalence, duration and asociated weight and behaviour changes for specific weight loss strategies over four years in US adults. International Journal of Obesity, 23, 320-327. French, S. A., Perry, C. L., Leon, G. R., & Fulkerson, J. A. (1995a). Dieting behaviors and weight change history in female adolescents. Health Psychology, 14(6), 548-555. French, S. A., Story, M., Downes, B., Resnick, M. D., & Blum, R. W. (1995b). Frequent dieting among adolescents: Psychosocial and health behavior correlates. American Journal of Public Health, 85, 695-710. French, S. A., Story, M., Neumark-Sztainer, D., Downes, B., Resnick, M., & Blum, R. (1996). Ethnic diferences in psychosocial and health behavior correlates of dieting, purging, and binge eating in a population-based sample of adolescent females. International Journal of Eating Disorders, 22, 315-322. Gobel, H., & Westphal, W. (1987). Die laterale asymmetrie der menschlichen schmerzempfindlichkeit. Der Schmerz, 1, 14-121. Goodpaster, B. H., DeLany, J. P., Oto, A. D., Kuller, L., Vockley, J., South-Paul, J. E.,...Jakicic, J. M. (2010). Efects of diet and physical activity interventions on weight loss and cardiometabolic risk factors in severely obese adults. Journal of the American Medical Asociation, 304(16), 1795-1802. Gordon, K. H., Selby, E. A., Anestis, M. D., Bender, T. W., Wite, T. K., Braithwaite, S.,?Joiner, T. E. (2010). The reinforcing properties of repeated deliberate self-harm. Archives of Suicide Research, 14(4), 329-341. Heatherton, T. F., & Polivy, J. (1991). Development and validation of a scale for measuring state RESTRICTIVE EATING RISK FOR SUICIDE ! 43 self-estem. Journal of Personality and Social Psychology, 60, 895-910. Heidelberg, N. F., & Correia, C. J. (2009). Dieting behavior and alcohol use behaviors among national eating disorders screening program participants. Journal of Alcohol & Drug Education, 53(3), 53-64. Helstrom, I. (2007). Oral complications in anorexia nervosa. European Journal of Oral Sciences, 85, 71-86. Hil, A. J., & Palin, V. (1998). Dieting awarenes and low self-worth: Related isues in 8-year- old girls. International Journal of Eating Disorders, 24(4), 405-413. Holm-Denoma, J. M., Wite, T. K., Gordon, K. H., Herzog, D. B., Franko, D. L., Fichter, M.,?Joiner, T. E. (2008). Deaths by suicide among individuals with anorexia as arbiters betwen competing explanations of the anorexia?suicide link. Journal of Afective Disorders, 107, 231-236. Insel, B. J., & Gould, M. S. (2008). Impact of modeling on adolescent suicidal behavior. Psychiatric Clinics of North America, 31, 293-316. Jefery, R. W., Adlis, S. A., & Forster, J. L. (1991). Prevalence of dieting among working men and women: The healthy worker project. Health Psychology, 10(4), 274-281. Joiner, T. (2005). Why people die by suicide. Cambridge, MA: Harvard University Pres. Joiner, T. E., Petit, J. W., Walker, R. L., Voelz, Z. R., Cruz, J., Rudd, M. D., & Lester, D. (2002). Perceived burdensomenes and suicidality: Two studies on the suicide notes of those atempting and those completing suicide. Journal of Social and Clinical Psychology, 21(5), 531-545. Joiner, T. E., & Van Orden, K. (2008). The interpersonal psychological theory of suicidal RESTRICTIVE EATING RISK FOR SUICIDE ! 44 behavior indicates specific and crucial psychotherapeutic targets. International Journal of Cognitive Therapy, 1, 80-89. Kang, H. K., & Bullman, T. A. (2008). Risk of suicide among US veterans after returning from the Iraq or Afghanistan war zones. Journal of the American Medical Asociation, 300, 652-653. Kemperman, I., Russ, M. J., & Shearin, E. (1997). Self-injurious behavior and mood regulation in borderline patients. Journal of Personality Disorder, 11, 146-157. Krahn, D. D., Kurth, C. L., Gomberg, E., & Drewnowski, A., (2005). Pathological dieting and alcohol use in college women?a continuum of behaviors. Eating Behaviors, 6, 43-52. Lautenbacher, S., Pauls, A. M., Strian, F., Pirke, K., & Krieg, J. (1991). Pain sensitivity in anorexia nervosa and bulimia nervosa. Biological Psychiatry, 29, 1073-1078. Lavender, J. M., De Young, K. P., & Anderson, D. A. (2010). Eating disorder examination questionnaire (EDE-Q): Norms for undergraduate men. Eating Behaviors, 11(2), 119- 121. Lowe, M. R., Annunziato, R. A., Markowitz, J. T., Didie, E., Belace, D. L., Riddel, L.,?Stice, E. (2006). Multiple types of dieting prospectively predict weight gain during the freshman year of college. Appetite, 47, 83-90. McGirr, A., Renaud, J., Bureau, A., Seguin, M., Lesage, A., & Turecki, G. (2008). Impulsive- aggresive behaviors and completed suicide across the life cycle: A predisposition for younger age of suicide. Psychological Medicine, 38, 407-417. Mintz, L. B., & Betz, N. E. (1988). Prevalence and correlates of eating disordered behaviors among undergraduate women. Journal of Counseling Psychology, 35, 463-471. Murray, F. S., & Hagan, B. C. (1973). Pain threshold and tolerance of hands and fet. Journal of RESTRICTIVE EATING RISK FOR SUICIDE ! 45 Comparative and Physiological Psychology, 84, 639-643. National Institutes of Health. (2012). Blood thinners. Retrieved from http:/ww.nlm.nih.gov/medlineplus/bloodthinners.html Neumark-Sztainer, D., Story, M., Dixon, L. B., & Murray, D. M. (1998). Adolescents engaging in unhealthy weight control behaviors: Are they at risk for other health-compromising behaviors? American Journal of Public Health, 88(6), 952-955. Nock, M. K., Hwang, I., Sampson, N. A., & Kesler, R. C. (2010). Mental disorders, comorbidity, and suicidal behavior: Results from the national comorbidity survey replication. Molecular Psychiatry, 15(8), 868-876. O?Dea, J. A., & Abraham, S. (2002). Eating and exercise disorders in young college men. Journal of American College Health, 50(6), 272-278. Orbach, I., Palgi, Y., Stein, D., HarEven, D., LotemPeleg, M.,?Elizur, A. (1996). Tolerance for physical pain in suicidal subjects. Death Studies, 20(4), 327-341. Orbach, I., Mikulincer, M., King, R., Cohen, D., & Stein, D. (1997). Thresholds and tolerance of physical pain in suicidal and nonsuicidal adolescents. Journal of Consulting and Clinical Psychology, 65, 646-652. Pauli, P., Wiedemann, G., & Nicola, M. (1999). Presure pain thresholds as asymmetry in left- and right-handers: Asociations with behavioural measures of cerebral laterality. European Journal of Pain, 3, 151-156. Pomerleau, O. F., Turk, D. C., & Fertig, J. B. (1984). The efects of cigarete smoking on pain and anxiety. Addictive Behaviors, 9, 265-271. Preti, A., Rocchi, M. B., Sisti, D., Camboni, M. V., & Miotto, P. (2011). A comprehensive meta- RESTRICTIVE EATING RISK FOR SUICIDE ! 46 analysis of the risk of suicide in eating disorders. Acta Psychiatrica Scandinavica, 124, 6- 17. Radloff, L. (1977). The CES-D scale: A self-report depresion scale for research in the general population. Applied Psychological Measurement, 1, 385-401. Rafiroui, A. C., Sargent, R. G., Parra-Medina, D., Drane, W. J., & Valois, R. F. (2003). Covariations of adolescent weight-control, health-risk and health-promoting behaviors. American Journal of Health Behavior, 27(1), 3-14. Rasmussen, K. A., Slish, M. L., Wingate, L. R., Davidson, C. L., & Grant, D. M. (2012). Can perceived burdensomenes explain the relationship betwen suicide and perfectionism? Suicide and Life-Threatening Behavior, 42(2), 121-128. Raymond, N. C., Faris, P. L., Thuras, P. D., Eiken, B., Howard, L. A., Hofbauer, R. D., & Eckert, E. D. (1999). Elevated pain threshold in anorexia nervosa subjects. Biological Psychiatry, 45, 1389-1392. Ribeiro, J. D., Wite, T. K., Van Orden, K., Selby, E. A., Gordon, K. H.,?Joiner, T. E. (2014). Psychometric properties and construct validity of the fearlesnes about death subscale of the acquired capability for suicide scale. Psychological Asesment, 26, 115- 126. Rice, K. G., Choi, C. C., Zhang, Y., Morero, Y. I., & Anderson, D. (2012). Self-critical perfectionism, aculturative stres, and depresion among international students. The Counseling Psychologist, 40, 575-600. Rice, K. G., Richardson, C. M. E., & Clark, D. (2012). Perfectionism, procrastination, and psychological distres. Journal of Counseling Psychology, 59(2), 288-302. Rudd, M. D., Joiner, T., & Rajab, M. H. (1996). Relationships among suicide ideators, RESTRICTIVE EATING RISK FOR SUICIDE ! 47 atempters, and multiple atempters in young-adult sample. Journal of Abnormal Psychology, 105, 541-550. Sacket, P. R., & Yang, H. (2000). Correction for range restriction: An expanded typology. Journal of Applied Psychology, 85, 112. Sarason, I. G., Sarason, B. R., Shearin, E. N., & Pierce, G. R. (1987). A brief measure of social support: Practical and theoretical implications. Journal of Social and Personal Relationships, 4, 497-510. Sarason, B. R., Shearin, E. N., Pierce, G. R., & Sarason, I. G. (1987). Interrelations of social support measures: Theoretical and practical implications. Journal of Personality and Social Psychology, 52, 813-832. Selby, E. A., Smith, A. R., Bulik, C. M., Olmstead, M. P., Thornton, L., McFarlane, T. L., ?.Joiner, T. E. (2010). Habitual starvation and provocative behaviors: Two potential routes to extreme suicidal behavior in anorexia nervosa. Behaviour Research and Therapy, 48, 634-645. Sidiropoulos, M. (2007). Anorexia nervosa: the physiological consequences of starvation and the need for primary prevention eforts. McGil Journal of Medicine, 10, 20-25. Slaney, R. B., Pincus, A. L., Uliaszek, A. A., & Wang, K. T. (2006). Conceptions of perfectionism and interpersonal problems: Evaluating groups using the structural summary method for circumplex data. Assesment, 13(2), 138-153. Slaney, R. B., Rice, K. G., Mobley, M., Trippi, J., & Ashby, J. S. (2001). The revised almost perfect scale. Measurement and Evaluation in Counseling and Development, 34, 130- 145. Smith, A. R., Fink, E. L., Anestis, M. D., Ribeiro, J. D., Gordon, K. H., Davis, H., ...& Joiner, T. RESTRICTIVE EATING RISK FOR SUICIDE ! 48 E. (2012). Exercise caution: Over-exercise is asociated with suicidality among individuals with disordered eating. Psychiatry Research, 206, 246-255. Smith, P. N., Cukrowicz, K. C., Poindexter, E. K., Hobson, V., & Cohen, L. M. (2010). The acquired capability for suicide: A comparison of suicide atempters, suicide ideators, and non-suicidal controls. Depresion and Anxiety, 27(9), 871-877. Solomon, R. L., & Corbit, J. D. (1974). An opponent-proces theory of motivation: I. Temporal dynamics of afect. Psychological Review, 81? 119-145. Stein, D., Kaye, W. H., Matsunaga, H., Myers, D., Orbach, I., Har-Even, D., ... & Rao, R. (2003). Pain perception in recovered bulimia nervosa patients. International Journal of Eating Disorders, 34, 331-336. Stice, E. (2002). Risk and maintenance factors for eating pathology: A meta-analytic review. Psychological Bulletin, 12, 825-848. Stice, E., Fisher, M., Lowe, M. R. (2004). Are dietary restraint scales valid measures of acute dietary restriction? Unobtrusive observational data suggest not. Psychological Asesment, 16, 51-59. Striegel-Moore, R. H., Rosseli, F., Perrin, N., DeBar, L., Wilson, G. T.,?Kraemer, H. C. (2009). Gender diference in the prevalence of eating disorder symptoms. International Journal of Eating Disorders, 42(5), 471-474. Swanson, S. A., Crow, S. J., Le Grange, D., Swendsen, J., & Merikangas, K. R. (2011). Prevalence and correlates of eating disorders in adolescents: Results from the National Comorbidity Survey Replication Adolescent Supplement. Archives of General Psychiatry, 68, 714-723. Timko, C. A., Perone, J., & Crossfield, A. (2006) Are you currently on a diet? What respondents RESTRICTIVE EATING RISK FOR SUICIDE ! 49 mean when they say ?yes.? Eating Disorders, 14, 157-166. Tylka, T. L., & Subich, L. M. (2002). Exploring young women?s perceptions of the efectivenes and safety of maladaptive weight control techniques. Journal of Counseling & Development, 80, 101-110. Van Dam, R. M., Wilet, W. C., Rim, E. B., Stampfer, M. J., & Hu, F. B. (2002). Dietary fat and meat intake in relation to risk of type 2 diabetes in men. Diabetes Care, 25(3), 417- 424. Van Orden, K. A., Cukrowicz, K. C., Wite, T. K., & Joiner, T. E. (2012). Thwarted belongingnes and perceived burdensomenes: Construct validity and psychometric properties of the interpersonal needs questionnaire. Psychological Asesment, 24, 197- 215. Van Orden, K. A., Wite, T. K., Cukrowicz, K. C., Braithwaite, S. R., Selby, E. A., & Joiner, T. E. (2010). The interpersonal theory of suicide. Psychological Review, 117, 575-600. Van Orden, K. A., Wite, T. K., Gordon, K. H., Bender, T. W., & Joiner, T. E. (2008). Suicidal desire and the capability for suicide: Tests of the interpersonal-psychological theory of suicidal behavior among adults. Journal of Consulting and Clinical Psychology, 76, 72- 83. Von Soest, T., & Wichstrom, L. (2009). Gender diferences in the development of dieting from adolescence to early adulthood: A longitudinal study. Journal of Research on Adolescence, 19(3), 509-529. Wadden, T. A., Brownel, K. D., & Foster, G. D. (2002). Obesity: Responding to the global epidemic. Journal of Consulting and Clinical Psychology, 70(3), 510-525. Whitehouse, A. M., & Button, E. J. (1988). The prevalence of eating disorders in a UK college RESTRICTIVE EATING RISK FOR SUICIDE ! 50 population: A reclasification of an earlier study. International Journal of Eating Disorders, 7, 393-397. Wite, T. K., Didie, E. R., Menard, W., & Philips, K. A. (2012). The relationship betwen body dysmorphic disorder behaviors and the acquired capability for suicide. Suicide and Life-Threatening Behavior, 42, 318-331. Wite, T. K., Gordon, K. H., Smith, P. A., & Van Orden, K. A. (2012). Stoicism and sensation seking: Male vulnerabilities for the acquired capability for suicide. Journal of Research in Personality. ! RESTRICTIVE EATING RISK FOR SUICIDE ! 51 Appendix 1: Tables RESTRICTIVE EATING RISK FOR SUICIDE ! 52 Table 1. Demographics for Dietary Restrictor and Control Groups. Demographic Variable M (SD) n (% of sample) Dietary Restrictors (n = 99) Age 19.80 (1.74) Gender Female 85 (85.90) Male 14 (14.10) Race White/European Origin 90 (90.90) Black/African American 7 (7.10) Asian American 1 (1.00) Native American/Pacific Islander 1 (1.00) Ethnicity Non-Hispanic/Latino 96 (97.00) Hispanic/Latino 3 (3.00) Controls (n = 94) Age 19.60 (2.20) Gender Female 77 (81.90) Male 17 (18.1) Race White/European Origin 84 (89.40) Black/African American 8 (8.50) Asian American 1 (1.10) Native American/Pacific Islander 1 (1.10) Ethnicity Non-Hispanic/Latino 93 (98.90) Hispanic/Latino 1 (1.10) RESTRICTIVE EATING RISK FOR SUICIDE ! 53 Table 2. Descriptive Statistics for Use of Lifetime Restrictive and Non-Restrictive Methods Within Dietary Restrictor Group. Statistics based on participant endorsement to items in Tables 2, 3, and select items from the Eating Disorder Examination Questionaire (EDE-Q). Lifetime Restrictive Methods a M times used (SD) n (%) of sample endorsing item Fasting 5.10 (2.44) 98 (98.90) b Skiping meals 5.25 (2.57) 94 (94.94) Low calorie fod 6.85 (2.50) 96 (96.97) Special diet 5.09 (3.19) 85 (85.86) Reduce calories 6.64 (2.64) 93 (93.94) Counted calories 5.94 (2.91) 90 (90.90) Eliminate fods 6.20 (2.93) 91 (91.92) Lifetime Non-Restrictive Methods a Self-induced vomiting 1.21 (2.43) 30 (30.30) Diuretic use 0.79 (1.78) 22 (22.22) Excesive exercise 4.71 (3.10) 86 (86.87) Laxative use 0.67 (1.57) 25 (25.25) Binge eating 3.34 (2.93) 69 (69.70) Past 28 Days Restrictive Methods Fasting c 1.80 (1.18) 99 (100.00) Skiping meals d 6.73 (6.55) 87 (87.88) Low calorie fod d 13.53 (8.95) 89 (90.81) Special diet d 9.63 (9.67) 66 (67.35) Reduce calories d 14.20 (8.97) 90 (91.84) Counted calories d 10.74 (10.41) 71 (72.45) Eliminate fods d 13.28 (10.53) 77 (78.57) Past 28 Days Non-Restrictive Methods Self-induced vomiting e 0.53 (2.54) 7 (7.07) Excesive exercise e 6.00 (6.93) 69 (69.70) Laxative use e 0.19 (0.79) 8 (8.08) Binge eating f 2.64 (4.00) 60 (60.60) Note: a Means for lifetime restrictive and non-restrictive methods reported are acording to Likert scale used for items in Tables 3 and 4 (i.e., 0 = 0 times using method during lifetime; 1 = 2-3; 3 = 4-10; 4 = 11-20; 5 = 21-40; 6 = 41-60; 7 = 61-80; 8 = 81-100; 9 = 100+). b Though one participant did not endorse fasting over lifetime, this participant was stil included in analyses based on endorsing the inclusion item asesing for fasting within 28 days. c Means for past 28 days fasting based on responses to item from the Eating Disorder Examination?Questionaire (EDE- Q) using the Likert scale (0 = 0 days; 1 = 1-5; 2 = 6-12; 3 = 13-15; 4 = 16-22; 5 = 23-27; 6 = every day). d Means for past 28 days restrictive methods and binge eating based on number of days (with open-ended response) over past 28 days endorsed item. e Means for past 28 days self- induced vomiting, excesive exercise, and laxative use are based on number of times (with open-ended response) over past 28 days endorsed item from the EDE-Q. f Past 28 days binge eating means based on number of days (with open-ended response) over past 28 days endorsed EDE-Q item. RESTRICTIVE EATING RISK FOR SUICIDE ! 54 Table 3. Items Asesing Frequency of Restrictive Dietary Methods Over Lifetime and Currently. Participants were instructed to indicate number of times each method has ben used over the course of their lives. For curent endorsement of methods, item 1 was not administered, and responses were open-ended for number of days over past 28 days used method. During your life, how many times? 0 1 2-3 4-10 11-20 21-40 41-60 61-80 81-100 100+ 1. Have you gone for long periods of time (8 hours or more) while awake without eating anything in order to influence your shape or weight? 0 1 2 3 4 5 6 7 8 9 2. ?Have you skipped one than one meal (besides breakfast) in order to influence your shape or weight? 0 1 2 3 4 5 6 7 8 9 3. ?Have you eaten food low in calories fod in order to influence your shape or weight? 0 1 2 3 4 5 6 7 8 9 4. ?Have you eaten on a special diet in order to influence your shape or weight? 0 1 2 3 4 5 6 7 8 9 5. ?Have you reduced the number of calories you consume in order to influence your shape or weight? 0 1 2 3 4 5 6 7 8 9 6. ?Have you counted calories in order to influence your shape or weight? 0 1 2 3 4 5 6 7 8 9 7. ?Have you eliminated specific foods from your diet in order to influence your shape or weight? 0 1 2 3 4 5 6 7 8 9 RESTRICTIVE EATING RISK FOR SUICIDE ! 55 Table 4. Items Asesing Frequency of Non-Restrictive Methods Over Lifetime. Participants were instructed to endorse one answer to indicate number of times each method has been used over the course of their lives. During your life, how many times? 0 1 2-3 4-10 11-20 21-40 41-60 61-80 81-100 100+ 1. ..Have you made yourself sick (vomit) as a means of influencing your shape or weight? 0 1 2 3 4 5 6 7 8 9 2. ..Have you taken diuretics (water tablets) as a means of influencing your shape or weight? 0 1 2 3 4 5 6 7 8 9 3. ..Have you exercised in a "driven" or "compulsive" way as a means of controling your weight, shape or amount of fat, or to burn off calories? 0 1 2 3 4 5 6 7 8 9 4. ..Have you taken laxatives as a means of influencing your shape or weight? 0 1 2 3 4 5 6 7 8 9 5. !...Have you had an episode of overeating (i.e., you have eaten an unusualy large amount of food and have had a sense of loss of control at the time)? 0 1 2 3 4 5 6 7 8 9 RESTRICTIVE EATING RISK FOR SUICIDE ! 56 ! ! Table 5. Intercorelations Betwen al Study Variables for Dietary Restrictor Group (N = 9) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 1 Gender 1.00 2 Self-esteem -.24 * 1.00 3 Num social -.09 .21 * 1.00 4 Satis social .07 .16 .18 1.00 5 Depresive .07 -.63 * -.34 * -.21 * 1.00 6 Perfection .16 -.69 * -.18 -.19 .57 * 1.00 7 Shape .29 * -.66 * -.13 -.07 .40 * .38 * 1.00 8 Weight .28 * -.59 * -.08 -.12 .39 * .37 * .88 * 1.00 9 Eating .24 * -.48 * -.12 -.06 .38 * .33 * .70 * .74 * 1.00 10 Restraint .13 -.23 * .09 .12 .33 * .16 .50 * .55 * .61 * 1.00 11 BMI -.29 * -.02 -.09 .18 .02 -.11 .15 .16 .12 .08 1.00 12 Belong -.08 -.51 * -.32 * -.31 * .69 * .45 * .27 * .31 * .22 * .03 .10 1.00 13 Burden .01 -.34 * -.34 * -.11 .46 * .27 * .24 * .19 .09 .00 .17 .50 * 1.00 14 SI .00 -.17 -.16 -.04 .17 .15 .12 .09 .12 -.07 .15 .26 * .34 * 1.00 15 FAD -.08 .08 .01 -.05 .13 .15 -.21 -.19 -.06 -.04 .01 .17 .12 .22 * 1.00 16 Pain tolerance -.48 .12 .04 -.12 .07 .04 -.03 -.02 -.02 .14 .21 * .12 .04 .19 .26 * 1.00 17 Total restrict .18 -.18 .06 .06 .22 * .10 .32 * .37 * .29 * .45 * .10 .10 .02 .00 .04 -.05 1.00 18 Life Fasting .08 -.19 -.01 -.04 .30 * .34 * .31 * .32 * .33 * .37 * .20 * .16 .07 .06 .06 .15 .37 * 19 Life Skiping meals .10 -.14 .01 -.02 .19 .24 * .25 * .28 * .26 * .29 * .16 .08 .03 -.14 .04 -.05 .38 * 20 Life Low calorie fod .29 * -.14 .03 .06 .17 .05 .32 * .34 * .32 * .45 * .03 .11 -.15 -.03 .03 -.01 .68 * 21 Life Special diet .15 -.07 .20 .11 .10 .07 .39 * .38 * .35 * .58 * .18 -.05 -.17 -.03 -.01 .09 .57 * 22 Life Reduce calories .20 -.10 .06 .10 .19 .04 .35 * .36 * .32 * .50 * .10 .07 -.10 -.08 .04 .05 .73 * 23 Life Counted calories .21 * -.15 -.04 .02 .15 .08 .37 * .36 * .35 * .45 * .14 .05 -.07 -.03 .01 .04 .59 * 24 Life Eliminate fods .10 -.12 .04 .05 .19 .12 .42 * .47 * .49 * .63 * .17 .02 -.12 .07 -.02 .07 .67 * 25 Total non-restrict .13 -.03 -.10 .07 .04 .03 .08 .14 .20 .22 * -.12 .02 .07 -.03 .07 -.03 .32 * 26 Life Vomiting .17 -.26 * -.21 * -.13 .27 * .36 * .20 .25 * .39 * .27 * -.19 .17 .16 .09 .10 .01 .14 27 Life Diuretic use .07 -.05 -.04 .14 .07 .11 .09 .19 .22 * .19 .04 .10 .04 .01 -.03 -.05 .17 28 Life Excesive exercise -.15 .01 .14 .10 .08 .06 .12 .12 .21 * .33 * -.03 -.09 -.05 -.02 .02 .11 .28 * 29 Life Laxative use .12 .02 -.19 .07 .00 .06 .08 .08 .14 .18 -.06 -.02 -.02 -.05 .00 .01 .17 30 Life Binge eating -.06 -.28 * -.06 -.12 .29 * .19 .27 * .26 * .49 * .27 * .18 .20 .09 .20 * .11 .14 .20 31 Curent Fasting -.17 -.08 -.06 -.05 .26 * .05 .16 .16 .24 * .40 * .27 * .07 .13 -.10 .04 .15 .12 32 Curent Skip meals .02 -.17 .04 -.14 .24 * .24 * .14 .16 .16 .29 * .08 .11 .02 -.12 .01 -.03 .25 * 33 Curent Low calorie fod .14 -.09 .13 .03 .16 .10 .27 * .36 * .34 * .53 * .09 .12 -.15 -.13 -.03 .03 .50 * 34 Curent Special diet .01 -.11 .22 * .04 .11 .02 .27 * .31 * .24 * .53 * .06 .02 -.08 -.05 -.02 .16 .33 * 35 Curent Reduce calories .00 -.05 .08 .09 .15 .04 .21 * .25 * .34 * .50 * .19 .05 -.04 -.14 .05 .14 .43 * 36 Curent Counted calories .16 -.06 .03 .05 .11 .02 .29 * .32 * .40 * .53 * .11 .03 -.06 -.05 -.08 .07 .36 * 37 Curent Eliminate fods .13 -.05 .10 .12 .15 .03 .34 * .41 * .35 * .61 * .07 .00 -.06 .06 -.03 .16 .39 * 38 Curent Vomiting .07 -.20 * -.04 -.15 .19 .11 .25 * .29 * .33 * .32 * -.10 .06 .04 -.07 -.09 -.11 .07 39 Curent Exercise .03 -.02 .10 -.04 .15 .07 .18 .32 * .21 * .42 * -.10 .08 .04 -.05 -.16 .21 * .18 40 Curent Laxatives .10 .07 .08 .07 -.13 .00 .07 .04 .14 .19 -.06 -.01 .06 -.08 -.10 -.09 .10 41 Curent Binge eating -.22 * -.16 -.11 -.19 .17 .08 .20 .20 * .36 * .07 .27 * .17 .17 .12 .03 .24 * -.06 RESTRICTIVE EATING RISK FOR SUICIDE ! 57 Note: Corelations are for total scores for variables 2-16. Gender: male = 0; female = 1. Num social = number of social suports; Satis social = satisfaction with social suport; Depresive = depresive symptoms; Perfection = maladaptive perfectionism; Shape = shape concern subscale; Weight = weight concern subscale; Eating = eating concern subscale; Restraint = restraint subscale; BMI = body mas index; Belong = belongingnes; Burden = burdensomenes; SI = suicidal ideation; FAD = fearlesnes about death; Pain tolerance = average across five trials. Total restriction = total score for number of restrictive methods used over lifetime; Total non-restrict = total score for number of non-restrictive methods used over lifetime; Variables 18-24 and 32-37 refer to endorsement of items from Table 3, which indicate frequency of restrictive methods over the lifetime and curently; Variables 26-30 refer to endorsement of items from Table 4, which indicate frequency of non-restrictive methods over lifetime; Variables 31, 38-41 refer to endorsement of current (past 28 days) restrictive and non-restrictive methods from Eating Disorder Examination Questionaire (EDE-Q). * p ? .05 ! Table 5 (continued). 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 18 Life Fasting 1.00 19 Life Skiping meals .84 * 1.00 20 Life Low calorie fod .45 * .41 * 1.00 21 Life Special diet .45 * .38 * .59 * 1.00 22 Life Reduce calories .50 * .46 * .87 * .66 * 1.00 23 Life Counted calories .42 * .37 * .74 * .57 * .80 * 1.00 24 Life Eliminate fods .56 * .51 * .70 * .74 * .73 * .66 * 1.00 25 Total non-restrict .17 .16 .21 * .21 * .18 .24 * .23 * 1.00 26 Life Vomiting .33 * .25 * .20 * .05 .15 .25 * .23 * .48 * 1.00 27 Life Diuretic use .17 .16 .17 .17 .18 .20 .17 .62 * .34 * 1.00 28 Life Excesive exercise .26 * .25 * .23 * .34 * .24 * .17 .34 * .29 * .29 * .15 1.00 29 Life Laxative use .08 .00 .15 .15 .12 .21 * .15 .57 * .37 * .40 * .20 1.00 30 Life Binge eating .23 * .17 .18 .25 * .20 .22 * .28 * .20 * .27 * .15 .42 * .19 1.00 31 Curent Fasting .53 * .48 * .06 .25 * .08 .11 .29 * .02 .11 -.01 .26 * .08 .25 * 1.00 32 Curent Skip meals .67 * .75 * .26 * .33 * .29 * .19 .38 * -.01 .14 .05 .12 -.07 .15 .62 * 1.00 33 Curent Low calorie fod .36 * .26 * .58 * .48 * .58 * .51 * .54 * .14 .09 .23 * .09 .19 .14 .12 .28 * 1.00 34 Curent Special diet .24 * .18 .34 * .59 * .33 * .22 * .46 * .19 .07 .13 .22 * .06 .12 .23 * .18 .59 * 1.00 35 Curent Reduce calories .35 * .30 * .49 * .37 * .57 * .44 * .40 * .10 .10 .19 .12 .13 .22 * .17 .32 * .74 * .43 * 36 Curent Counted calories .30 * .21 * .44 * .39 * .44 * .63 * .43 * .21 * .22 * .19 .13 .20 .16 .20 .16 .64 * .41 * 37 Curent Eliminate fods .22 * .11 .34 * .42 * .36 * .23 * .51 * .14 .09 .14 .23 * .06 .12 .19 .14 .57 * .64 * 38 Curent Vomiting .11 .08 -.02 .02 -.03 .10 .09 .13 .58 * .07 .16 .08 .17 .11 .08 .09 .15 39 Curent Exercise .18 .11 .18 .36 * .17 .13 .29 * .21 * .18 .05 .53 * .18 .14 .18 .15 .23 * .42 * 40 Curent Laxatives .08 .14 .04 .11 .05 .11 .11 .29 * .10 .34 * .12 .34 * .17 .02 .11 .06 -.01 41 Curent Binge eating -.02 -.01 -.11 .01 -.07 -.05 .01 .06 .13 .05 .00 .03 .62 * .08 .00 .00 .05 35 36 37 38 39 40 41 35 Curent Reduce calories 1.00 36 Curent Counted calories .68 * 1.00 37 Curent Eliminate fods .47 * .50 * 1.00 38 Curent Vomiting .08 .24 * .18 1.00 39 Curent Exercise .21 * .26 * .37 * .18 1.00 40 Curent Laxatives .08 .13 .00 .01 .13 1.00 41 Curent Binge eating .14 -.01 .03 .11 -.01 .05 1.00 RESTRICTIVE EATING RISK FOR SUICIDE ! 58 Table 6. Intercorrelations Between all Study Variables for Control Group (N = 94). 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 1 Gender 1.00 2 Self-esteem -.05 1.00 3 Num social .08 .27 * 1.00 4 Satis social -.05 .38 * .183 1.00 5 Depresive .03 -.70 * -.46 * -.43 * 1.00 6 Perfection -.20 -.63 * -.32 * -.22 * .61 * 1.0 7 Shape .30 * -.40 * -.03 -.33 * .29 * .02 1.0 8 Weight .29 * -.30 * -.02 -.15 .24 * -.04 .78 * 1.0 9 Eating .16 -.24 * -.01 -.18 .21 * .05 .52 * .55 * 1.00 10 Restraint .13 .02 -.07 -.26 * -.08 -.13 .36 * .23 * .21 * 1.0 11 BMI -.17 .01 .03 -.04 -.03 -.12 .14 .39 * .23 * -.06 1.0 12 Belong -.11 -.57 * -.54 * -.45 * .76 * .47 * .28 * .20 .20 -.03 .06 1.0 13 Burden .09 -.63 * -.35 * -.53 * .74 * .41 * .35 * .32 * .21 * -.05 .02 .69 * 1.0 14 SI .09 -.38 * -.29 * -.33 * .45 * .26 * .20 .15 .06 -.02 -.11 .40 * .61 * 1.00 15 FAD -.29 * .19 .01 .07 -.13 -.12 -.14 -.17 -.23 * -.05 .13 -.00 -.00 .01 1.0 16 Pain tolerance -.50 * -.00 -.07 .03 -.03 .18 -.20 -.11 -.14 -.06 .26 * .01 .11 .13 .25 * 1.00 Note: Corelations are for total scores for variables 2-16. Gender: male = 0; female = 1. Num social = number of social suports; Satis social = satisfaction with social suport; Depressive = depressive symptoms; Perfection = maladaptive perfectionism; Shape = shape concern subscale; Weight = weight concern subscale; Eating = eating concern subscale; Restraint = restraint subscale; BMI = body mass index; Belong = belongingness; Burden = burdensomenes; SI = suicidal ideation; FAD = fearlessness about death; Pain tolerance = average across five trials. * p ? .05 RESTRICTIVE EATING RISK FOR SUICIDE ! 59 Table 7. Results of Hypotheses 1-4. Betwen-subjects efects and group mean diferences for al study hypotheses. Psychosocial Variable Total Score Dietary Restrictors M (SD) Controls M (SD) F(df) p Partial ? 2 Hypothesis 1 Self-esteem 64.27 (13.73) 77.70 (11.30) 22.93 (1, 186) .00 .11 Number of social supports 5.09 (2.11) 5.59 (2.23) 0.10 (1, 186) .82 .00 Satisfaction with social support 5.29 (1.15) 5.58 (.81) 2.22 (1, 186) .15 .01 Depresive symptoms 16.29 (9.96) 10.71 (8.68) 5.97 (1, 186) .02 .03 Maladaptive perfectionism 50.70 (15.83) 38.55 (15.34) 11.62 (1, 186) .00 .06 Shape concern 3.56 (1.42) 0.89 (0.84) 98.70 (1, 186) .00 .35 Weight concern 3.01 (1.45) 0.62 (0.69) 72.01 (1, 186) .00 .28 Eating concern 1.40 (1.12) 0.11 (0.19) 29.27 (1, 186) .00 .14 Restraint 2.82 (1.60) 0.74 (1.05) 30.66 (1, 186) .00 .14 Body mas index 24.74 (3.91) 22.63 (3.30) 13.28 (1, 186) .00 .07 Hypothesis 2 Belongingnes 10.57 (9.84) 8.38 (9.95) 0.84 (1, 189) .36 .00 Burdensomenes 1.48 (2.86) 1.30 (3.17) 0.04 (1, 189) .85 .00 Hypothesis 3 Suicidal ideation 0.79 (2.48) 0.26 (1.17) 2.46 (1, 190) .12 .01 Hypothesis 4 FAD 13.72 (6.70) 14.64 (5.44) 1.11 (1, 176) .29 .00 Pain tolerance 236.74 (132.92) 254.57 (130.90) 0.15 (1, 176) .70 .00 Note: All results control for total score lifetime non-restrictive methods. The Restraint subscale reflects total scores not including the item asesing for fasting. Group means difer at p < .05 using Benjamini and Hochberg?s (195) procedure to control for the false discovery rate in post hoc pairwise comparisons. !