Differentiating between Children with and without ADHD Using the Delis-Kaplan 
Executive Function System and Conners? Continuous Performance Test-II
by
Sarah Nicolette Peden
A dissertation submitted to the Graduate Faculty of
Auburn University
in partial fulfillment of the
requirements for the Degree of 
Doctor of Philosophy
Auburn, Alabama
December 13, 2010
Keywords: D-KEFS, Executive Functioning, ADHD, CCPT-II
Copyright 2010 by Sarah Nicolette Peden
Approved by
Steve Shapiro, Chair, Associate Professor of Psychology
Christopher Correia, Assistant Professor of Psychology
Martha Escobar, PhD, Assistant Professor of Psychology
Jennifer Gillis, Assistant Professor of Psychology
Abstract
The Delis-Kaplan Executive Function System (D-KEFS) is a relatively new measure of 
executive functioning (EF) that has not been fully evaluated for its potential of differentiating 
between children with and without ADHD.  The Conners? Continuous Peformance Test (CCPT) 
is a computerized task that has been studied extensively in ADHD populations and been found to 
have moderate success at predicting diagnostic status.  The present study examined diagnostic 
group differences on the D-KEFS and CCPT between children with ADHD (n = 32) and a 
control group (n = 37).  Results replicate previous findings in both the D-KEFS and general EF 
literature (Corbett et al., 2009; Frazier et al., 2004; Wodka et al., 2008a; Wilcutt et al., 2005). 
Specifically, ADHD group performance on several measures of planning, reconstitution, and 
inhibition was found to be below average or significantly lower than that of the control group. 
Additionally, performance on most D-KEFS variables, including some less executive tasks, was 
associated with diagnostic group status.  Inattentive and Hyperactive/Impulsive symptoms, as 
rated by parents on the CPRS, were significantly related to almost all D-KEFS variables. 
However, higher magnitude correlations between Inattention and the Trailmaking and Color-
Word Interference tests were noted.  On the CCPT, between-group differences were found for 
Hit Reaction Time Standard Error, Detectability, and Omission errors.  Intercorrelations between 
the CCPT and CPRS symptom scales revealed Reaction Time variables related to both symptom 
domains, Omissions related to Inattention, and Detectability and Response Style related to 
Hyperactivity/Impulsivity.  Significant correlations between several CCPT and D-KEFS 
ii
variables were found, with some planning/inhibition/switching and more basic skill D-KEFS 
measures being associated with CCPT scores.  Moderate predictive utility of diagnostic group 
was found for both the D-KEFS and CCPT individually.  Combined predictive utility of both 
measures resulted in non-significant models.  Recommendations are provided regarding the use 
of the D-KEFS and CCPT in evaluating ADHD, as well as important future research directions.
iii
Acknowledgments
I wish to thank Steve Shapiro for his unfailing support and guidance throughout the 
dissertation process.  Without the countless meetings, consideration of new ideas, hours of 
statistical analyses, and lengthy phone conferences, this project would never have seen 
completion.  I also thank Steve for his mentorship and continued interest in my future career as a 
clinician and researcher.  My doctoral committee members Chris Correia, Martha Escobar, Jen 
Gillis, and Joseph Buckhalt also deserve my appreciation.  I appreciate their flexibility and 
helpful feedback.  I also wish to thank Andy Cohen, Bill Ferguson, and Clarissa Mooney for 
their support in the Shapiro Lab meetings, as well as their assistance with project ideas and data 
collection.  We all survived the Thach basement sauna!  The data entry and checking efforts of 
Laura Coursen and Kristen Hiott are not to go unnoticed, as they diligently helped with the 
tedious aspects of the project.  While on internship, many individuals provided insight and 
guidance for my project, including David Tupper, Pat McMahon, Andrea Szporn, and my fellow 
interns.  Finally, this project would not have been completed without the financial assistance of 
the Brown Finch Foundation and the support of Emily Lambeth. 
To Mom, Dad, and Luke, your support throughout the dissertation process, as well as all 
my years of schooling, has been incredible.  Through the good times and trying times, you are 
always there to encourage me, cheer me up, and tell me how proud you are of me.  I could not 
ask for a better family, as I know we can always count on each other.  Mom and Dad, thank you 
for encouraging your children to follow their dreams and for providing guidance along the way.
iv
Table of Contents
Abstract.........................................................................................................................................ii
Acknowledgments.......................................................................................................................iv 
List of Tables.............................................................................................................................viii 
Introduction  .................................................................................................................................1
Attention-Deficit/Hyperactivity Disorder  .......................................................................2
What is Executive Functioning?  ......................................................................................2
Theories of ADHD and Executive Functioning................................................................4
          Barkley?s theory of ADHD and EF.........................................................................4
          Nigg?s Theory of ADHD and EF ...........................................................................8
                      Other theories of ADHD.........................................................................................9
Developmental Considerations of Executive Functioning  ........................................................12
Associations between EF measures of Behavioral Inhibition and Planning/Reconstitution and 
           ADHD in Children   ........................................................................................................14
                        
            Behavioral Inhibition (Control of Motor Response and Behavior)  ...............................15
 Continuous Performance Tasks.............................................................................15
 Stroop Color-Word Test........................................................................................18
             Reconstitution/Planning.................................................................................................20
 Tower Tests...........................................................................................................20
 Trail Making Tests................................................................................................21
 Controlled Oral Word Association Test................................................................21
v
Delis-Kaplan Executive Function System (D-KEFS).................................................................23
Rationale for the Current Study...................................................................................................28
Hypotheses for the Current Study...............................................................................................31
Method.........................................................................................................................................33
               Participants...................................................................................................................33
Measures.......................................................................................................................36
    Conners? Parent Rating Scales-Revised, Long Form. (CPRS-R:L)................36
    Conners? Continuous Performance Test-II (CCPT-II)....................................37
    Delis-Kaplan Executive Function System (D-KEFS).....................................38
   Wechsler Abbreviated Scale of Intelligence (WASI)......................................39
     Independent Measures...................................................................................................39
Between-group.................................................................................................39
              Dependent Measures.....................................................................................................39
 CCPT-II...........................................................................................................39
 D-KEFS...........................................................................................................39
              Procedure.......................................................................................................................40
Results.........................................................................................................................................43
Statistical Considerations.............................................................................................43
Diagnostic Group Performance....................................................................................45
             D-KEFS...........................................................................................................45
             CCPT...............................................................................................................50
Correlations between D-KEFS, CCPT, and CPRS Scores...........................................51
Predictive Utility..........................................................................................................55
vi
             D-KEFS...........................................................................................................55
             CCPT...............................................................................................................57
             Combined Tasks..............................................................................................58
D-KEFS Prediction of ADHD Symptoms by Age Group............................................60
Discussion....................................................................................................................................64
D-KEFS Between-Group Performance........................................................................64
CCPT Between-Group Performance............................................................................67
Predictive Utility of Diagnostic Group.........................................................................70
              D-KEFS..........................................................................................................70
              CCPT..............................................................................................................72
              D-KEFS and CCPT........................................................................................73
Concurrent Validity Between the D-KEFS and CCPT................................................74
Predictive Utility of the D-KEFS by Age Group.........................................................76
Limitations....................................................................................................................79
Conclusions and Future Directions..............................................................................81
References...................................................................................................................................85
Appendix A  ...............................................................................................................................97
Appendix B  ..............................................................................................................................101
vii
List of Tables
Table 1. Diagnostic Group Descriptives ....................................................................................36
Table 2. CPRS Group Comparison ............................................................................................36
Table 3. CCPT-II Variables ........................................................................................................38
Table 4. D-KEFS Variables ........................................................................................................40
Table 5. Group Performance on D-KEFS ..................................................................................47
Table 6. Point Biserial Correlations between Diagnostic Group and D-KEFS Variables .........47
Table 7. Intercorrelations among D-KEFS Variables and IQ ....................................................49
Table 8. Group Performance on CCPT Variables.......................................................................50
Table 9. CCPT and IQ Intercorrelations......................................................................................50
Table 10. Correlations between CCPT and D-KEFS Variables..................................................53
Table 11. Correlations between D-KEFS, CCPT, and CPRS Scores .........................................55
Table 12. Logistic Regression for D-KEFS Planning/Inhibition/Switching Variables...............56
Table 13. Logistic Regression for all D-KEFS Variables...........................................................57
Table 14. Logistic Regression for CCPT Variables....................................................................58
Table 15. Logistic Regression for Combined D-KEFS Planning/Inhibition/Switching and 
                 CCPT Variables..........................................................................................................59
Table 16. Logistic Regression for Combined CCPT and Exploratory D-KEFS Variables.........60
Table 17. D-KEFS Logistic Regression for Younger Age Group...............................................62
Table A1. Meta-analysis Results for Selected EF Measures.......................................................98
viii
Table A2. Predictive Validity of ADHD by Selected EF Tests..................................................98
Table A3. Internal Consistency and Test-Retest Reliability Values Reported 
                  in D-KEFS manual....................................................................................................99
Table A4. Intercorrelations of D-KEFS Variables Reported in the Technical Manual............100
ix
Introduction
Reflecting the substantial prevalence of and documented associated functional impairment, 
Attention-Deficit/Hyperactivity Disorder (ADHD) has been the focus of considerable 
psychological and medical research.  Recent theories proposed by Barkley (1997; 2006) and 
others (e.g. Nigg, 2006) point to Executive Functioning (EF) as a particular area of interest in 
ADHD.  Specifically, it is postulated that deficits in EF are key features of individuals who meet 
criteria for the disorder. 
 However, finding assessment tools to qualitatively and quantitatively measure a construct 
such as EF can be challenging.   The incremental development of EF throughout childhood and 
adolescence further complicates the measurement of EF, as their skills may be more in flux than 
those of adults.  Executive functioning is biologically based in the sense that the processes have 
been found to be regulated by the prefrontal cortex, inferior frontal junction, premotor cortex, 
anterior cingulate, and cerebellum (Davidson, Amso, Anderson, & Diamond, 2006).  Thus, some 
neuroimaging techniques are available to measure brain function.  However, most psychologists 
and neuropsychologists measure EF through a variety of cognitive tests, varying from hands-on 
manipulative tasks to computerized procedures.  The purpose of this study is to explore the 
utility of a fairly new measure of EF, the Delis-Kaplan Executive Function System (D-KEFS; 
Delis, Kaplan & Kramer, 2001) in differentiating between children with and without ADHD.  In 
addition, concurrent and incremental validity of a measure commonly used in ADHD 
evaluations, the Conners? Continuous Performance Test-II (CCPT-II; Conners, 1995), will be 
assessed.
1
Attention-Deficit/ Hyperactivity Disorder (ADHD)
ADHD is the most common neurodevelopmental disorder seen in psychology practice 
around the United States.  Its prevalence varies by gender, age, social class, and population 
density, with 2-18% of children exhibiting clinically significant symptoms (Rowland, Lesene, & 
Abramowitz, 2002).  Furthermore, males are generally more likely than females to meet 
diagnostic criteria, with a ratio of 3:1 in community samples and a range of 5:1 to 9:1 in clinical 
samples (American Psychiatric Association, 2000).  
The Diagnostic and Statistical Manual of Mental Disorders (DSM-IV-TR; American 
Psychiatric Association [APA], 2000) characterizes ADHD along two symptom domains, 
inattention-disorganization and hyperactivity-impulsivity, which yields three clinical subtypes: 
Predominantly Inattentive (PI), Predominantly Hyperactive-Impulsive (HI), and Combined (C). 
These two symptom domains were derived from empirical studies and factor analysis of parent 
and teacher ratings in preparation for DSM-IV (Lahey, Applegate, McBurnett, Biederman, 
Greenhill, et al., 1994).  Inattentive symptoms include daydreaming, ?spacing out? or being ?lost 
in a fog,? staring frequently, and being easily confused, lethargic, and passive. Hyperactive-
impulsive symptoms include being ?always on the go? (e.g., always out of seat), acting as if 
driven by a motor (e.g., constantly moving arms and legs while at task), and acting impulsively 
(e.g., blurting out answers before questions are completed). The Combined subtype includes 
symptoms of both the Inattentive and Hyperactive-Impulsive variants.
What is Executive Functioning?
Executive functioning (EF) is the ability to maintain an appropriate and efficient 
problem-solving set for attainment of a future goal (Bianchi, 1922; Lezak, 1985; Luria, 1966). 
Brown (2006), in describing EF, uses a metaphor of a conductor of a symphony.  Just as there are 
2
numerous parts of the brain that all function differently, so are there many musicians in a 
symphony.  The role of the conductor is to integrate the functions of these musicians in the 
execution of a musical composition, from indicating when to start the piece, to keeping time, to 
modulating dynamics.  The frontal lobe of the brain performs a similar function in that it helps to 
manage cognitive functions that assist in tasks such as programming and planning of goal-
oriented motor skill behavior, modulation of behavior in light of expected future consequences, 
anticipation of events in the regulation of behavior, learning of contingency rules and the ability 
to use feedback cues, inhibition of response set and flexibility (versus perseveration), abstract 
reasoning, problem-solving, sustained attention, and concentration (Seguin, Phil, Harden, 
Tremblay, & Boulerice, 1995).  
EF has been found to be a distinct set of cognitive skills when compared to intelligence. 
Although some skills may overlap, individuals who have sustained damage to areas of the 
prefrontal cortex show deficits in reasoning tasks related to EF while sustaining normal and 
intact levels of intelligence (Waltz, Knowlton, & Holyoak, 1999).  Some investigators have 
found specific IQ test profiles that show lower performance by children with ADHD on the 
Arithmetic, Coding, Information, Digit Span, and Symbol Search subtests of the Wechsler 
intelligence tests (Dykman, Ackerman, & Oglesby, 1980).  The overall IQ scores of these 
individuals are then lowered because of performance on these specific subtests.  Besides the 
Information subtest, all others are part of the Working Memory or Processing Speed indexes of 
the WISC-IV (Wechsler Intelligence Scale for Children, Fourth Edition; Wechsler, 2003).  
According to the Cattell-Horn-Carroll (CHC) theory of intelligence, general intelligence 
is a latent structure that includes performance in numerous areas of functioning (fluid reasoning, 
language knowledge and use, memory and learning, visual and auditory perceptive, retrieval 
3
ability, cognitive speed, and reaction time speed) (Carroll, 1993).  Therefore, it is possible that 
children with ADHD have difficulty with some of these skills because they are very similar to 
those measured by EF tests.  However, performance on these tasks is combined with that on 
verbal and perceptual reasoning tasks for an overall IQ score, which would then dilute the 
correlation between EF and intelligence.  Schuck and Crinella (2005) found minimal correlations 
between EF measures and all subtests of the Wechsler Intelligence Scale for Children ? Third 
Edition (WISC-III; Wechsler, 1992) in 123 boys with ADHD.  Therefore, because intelligence is 
typically measured combining many factors of the CHC theory, the potential for high 
correlations with EF is lowered.
EF deficits have also been related to several psychological disorders and developmental 
delays, such as Attention-Deficit/Hyperactivity Disorder (ADHD) and autism (Pennington & 
Ozonoff, 1996).  Individuals with ADHD, phenylketonuria (PKU), and specific learning 
disabilities have been found to have impaired levels of EF performance while exhibiting general 
intelligence within the normal range (Blair, Zelazo & Greenberg, 2005). Thus, measured 
intelligence can appear unaffected while measured EF can be much lower than expected given 
intelligence scores, suggesting that multiple cognitive processes are at work.
Theories of ADHD and Executive Functioning
Barkley?s Theory of ADHD and EF
Barkley (1997; 2006) articulated an integrative theory of ADHD because he believed that 
the existing models of ADHD merely described two behavioral deficits (inattention and 
hyperactivity-impulsivity) and failed to account for many other cognitive and behavioral deficits 
associated with the disorder. Therefore, he set out to create a more unifying theory of the 
disorder, proposing that the primary deficit in ADHD is impairment in response inhibition, which 
4
leads to disruption in performance of EFs. Barkley (1997; 2006) posited that successful 
behavioral inhibition sets the stage for four executive functions to occur: nonverbal working 
memory, verbal working memory, self-regulation of affect/motivation/arousal, and 
reconstitution. Therefore, once a person successfully inhibits an automatic or ongoing response 
pattern, that individual may then proceed to the executive function processes. However, a person 
with ADHD is likely to have poor behavioral inhibition, leading to deficits in these areas of 
executive functioning which may lead to less overall success in the execution of goal-directed 
behavior.
Barkley (1997; 2006) defined behavioral inhibition as three interrelated processes: ?(a) 
inhibition of the initial prepotent response to an event; (b) stopping of an ongoing response, 
which thereby permits a delay in the decision to respond; and (c) the protection of this period of 
delay and the self-directed responses that occur within it from disruption by competing events 
and responses? (p. 67).  In this light, Barkley characterized the primary underlying mechanism of 
ADHD as a deficit in behavioral inhibition and executive control of behavior, through which 
inattention becomes a secondary symptom.
It is believed that behavioral inhibition is the first self-regulatory act in responding, as it 
allows more time to generate alternate responses, anticipate consequences of various responses, 
and make an appropriate choice of future behavior. Barkley (1997) asserted that the four 
executive functions affect two types of sustained attention: contingency-shaped and motor 
control. Specifically, he argued that immediate contextual factors (e.g., novelty of the task, 
reinforcement, and delay of reinforcement) govern the attention level demonstrated, which could 
then affect task performance.  It has been found that performance of individuals with ADHD 
tends to be more easily influenced by these contextual factors (response contingencies) compared 
5
to normal control counterparts (Douglas, 1985; Haenlein & Caul, 1987).
Research findings indicate that the cognitive deficit seen in ADHD is at the motor control 
rather than the attentional or information-processing stage (Sergeant, 2005), which supports 
Barkley?s assertion that inattention in ADHD is secondary to, and results from, deficits in 
behavioral inhibition.  Specifically, in a review of the relationship between behavioral inhibition 
and ADHD, Nigg (2001) reported that there is more evidence to support the inhibitory deficit 
when it involves suppression of a pre-potent motor response, such as the Stop or Go/No-go tests, 
as opposed to secondary response inhibition, such as that measured by the Stroop test.  Barkley 
also included a motor control aspect in which self-regulation, internal representation of 
motivation and goals, and novel chains of responses all influence the inhibition of impulsive 
motor behavior. 
The first EF implicated in Barkley?s theory is nonverbal working memory.  Individuals 
with ADHD tend to have significant difficulties in nonverbal working memory when complex 
information must be held in mind over lengthy delays (Seidman, Biederman, & Faraone, 1995). 
Barkley (1997; 2006) suggested that this inability to hold information in mind can lead to 
impairments in imitating complex, novel, and lengthy behavioral sequences, temporal 
organization and regulation, and consequently, the disorganization of the syntax of motor 
planning and execution. Therefore, a deficiency in nonverbal working memory makes it difficult 
for those with ADHD to determine the times and places for appropriate and adaptive behavior, as 
well as the steps and sequences required to complete nonverbal tasks and activities.  
The second EF proposed by Barkley is verbal working memory, which includes self-
directed (internalized) speech, following rules, and moral reasoning (Kochanska, DeVet, & 
Goldman, 1994). Deficits in the internalization of speech have been seen among children with 
6
ADHD who had difficulty complying with directions and commands, following rules when the 
rules competed with rewards available for rule violation, and transferring initially learned rules 
to novel tasks (Conte & Regehr, 1991; Danforth, Barkley, & Stokes, 1991; Hinshaw & Melnick, 
1992).  Therefore, Barkley (1997; 2006) suggested that people with ADHD show deficits in 
creating an internalized dialogue of what they should be doing and with keeping long sequences 
of verbal information in mind, especially when required to manipulate the information and use it 
for a task.
Next, Barkley implicated self-regulation of affect/motivation/arousal as an EF associated 
with ADHD, with deficits resulting in greater emotional reactivity, fewer anticipatory emotional 
reactions, less capacity to regulate emotional drive and motivation for future goals, and greater 
dependence on external sources for drive, motivation, and arousal. Therefore, Barkley argued 
that the behavior of people with ADHD is primarily under the control of immediate and external 
sources of reinforcement, rather than internal motivation. 
Finally, Barkley included reconstitution as the fourth EF, which he described as the 
ability to create multiple, novel, complex, alternative response sequences. Researchers have 
found that people with ADHD perform poorly on verbal and organizational fluency tests 
(Grodzinsky & Diamond, 1992). Also included in this domain are analysis and synthesis of 
behavior, in which components or steps toward a response are broken down and reorganized 
based on the demands of a new task.  Furthermore, planning and goal-directed behavior are 
important in reconstitution.  Barkley reasoned that if a person has difficulty with behavioral 
inhibition, that person consequently may not be able to envision multiple behavioral responses 
and would have a deficit in the EF of reconstitution. 
7
Nigg?s Theory of ADHD and EF
Nigg (2006) set forth a similar theory of ADHD to Barkley?s, implicating cognitive 
control, or EF, as primary deficits in the disorder.  However, Nigg eliminates the hierarchy of 
behavioral inhibition with the other EFs and simply suggests that all aspects of EF are equally 
important.  Nigg breaks down the cognitive control EFs into four areas: control of attention, 
control of motor response and behavior, working memory, and state regulation.  
Nigg (2006) describes ?control of attention? as important in selection and working 
memory, conflict detection, and control of interfering information/responses.  Thus, this area of 
EF serves to filter competing information, suppress this information from working memory, and 
allow for more appropriate current responses.  Nigg suggests that interference control and 
cognitive inhibition also fall into this category, in which one would put out of mind thoughts that 
are not relevant to the task and select appropriate cognitions to facilitate task completion.  Nigg?s 
control of attention category seems to fit most closely with Barkley?s (2006) verbal working 
memory EF component in that both require the filtration of interfering cognitions for more task-
appropriate thoughts.
Next, Nigg (2006) suggests that control of motor response and behavior is critical for the 
suppression or interruption of a prepared response, as well as the delay of any or all responding. 
This second area of ?control? in Nigg?s theory is more motor response and behaviorally oriented, 
rather than being cognitive and thought-process oriented.  One must be able to stop a prepared or 
previously performed motor response in order to adapt to new task demands.   Therefore, Nigg?s 
description of this area maps on most closely to Barkley?s (1997) description of behavioral 
inhibition.  
Nigg?s (2006) third EF in his theory of ADHD is working memory, which includes 
8
auditory, spatial, and location working memory.  Specifically, Nigg asserts that working memory 
is distinct from interference control, as it involves additional demands such as the ability to 
manipulate information over a short period of time and protect it from interfering thoughts and 
stimuli.  In addition, planning is included within the working memory domain and is described 
as the ability to mentally organize the steps for solving a problem and determining the 
appropriate sequence for these steps.  Nigg?s description of working memory seems to overlap 
most closely with Barkley?s (2006) category of nonverbal working memory.  However, with the 
addition of planning in Nigg?s category, it may also include some of Barkley?s reconstitution 
functions.
Finally, Nigg (2006) suggests state regulation to be the fourth EF in his theory, and it is 
described as activation, readiness to respond, or motor preparation.  Furthermore, Nigg 
distinguishes activation from attention or vigilance in that it is the preparedness to respond and is 
not simply the overall alertness to surroundings.  Nigg?s proposed state regulation category of 
EF is most similar to Barkley?s (2006) self-regulation of affect/motivation/arousal.  Both 
describe the readiness to respond, where the brain and body have to be active and attentive to the 
task. 
Other theories of ADHD 
Some other theories are prominent in the ADHD literature, one of which is the dual 
pathway model (Sonuga-Barke, 2002, 2005).  This model proposes two possible mechanisms 
toward ADHD: an inhibitory deficit and an altered reward/reinforcement deficit.  The pathways 
are associated with two distinct subtypes of ADHD, one that results from dysregulation of action 
and poor inhibitory motor control and one that stems from delay aversion (preferring immediate 
small rewards over larger delayed rewards).  The dysregulation of thought and action pathway 
9
(DTAP) is characterized by disinhibition, which results in consequences for both behavioral 
(impulsiveness, inattention, hyperactivity) and cognitive (quality and quantity of task 
engagement) processes.  This pathway is thought to be more biological and less context-
dependent, and individuals with the DTAP form of ADHD are expected to be more generally 
cognitively impaired.  The motivational style pathway (MSP) suggests that delay aversion is an 
acquired characteristic stemming from shortened dopaminergic reward circuits in the brain, 
combined with environmental factors.  Tendency to be delay aversive is strengthened over time 
by an individual?s experiences in situations where reward is delayed and reinforcement is given 
to shortening the delay (thus, a form of conditioning). It is suggested that delay aversion leads to 
both impulsivity and inattention, as an individual is likely to act quickly and impulsively if 
wanting rewards sooner and may also try to find stimulation in the environment to increase how 
?rewarding? a task is.  Both pathways are proposed to be separate manifestations of ADHD 
symptoms, and Sonuga-Barke uses this multiple pathway model to possibly explain some of the 
inconsistencies in research on neuropsychological profiles of ADHD.
The other theory mentioned in ADHD literature is the cognitive-energetic model 
(Sergeant, 2005).  This model includes deficits associated with ADHD on three levels: cognitive 
response output (encoding, search, decision, and motor organization), energetic activation and 
effort (effort, arousal, and activation), and EF control (planning, monitoring, detection of errors, 
and error correction).  It is posited that ADHD results from the inability of an individual to 
modulate physiological state to meet task demands, with problems occurring at one or more of 
the three levels.  Specifically, Sergeant posits that information processing is likely intact in 
individuals with ADHD, but that the failure of inhibition (or other EFs) might be a failure of the 
activation of the inhibitory mechanism (activation and effort factors) rather than a deficit in the 
10
inhibitory skill itself.  Therefore, his theory combines factors from the more pure EF theories and 
the dual pathway model.
11
Developmental Considerations of Executive Functioning
EF skills, as noted earlier, have been found to be associated with brain activity in the 
prefrontal cortex, inferior frontal junction, premotor cortex, anterior cingulate, and cerebellum. 
With normal brain development, EF skills develop and become more fine-tuned throughout 
childhood and adolescence.  Some researchers consider EF as a multi-faceted cognitive 
component in which different developmental trajectories occur and maturation is reached in 
stages (Anderson, Anderson & Lajoie, 1996).  It has been found that the brain develops and fine-
tunes itself in a back-to-front direction, beginning in the primary motor cortex, moving forward 
with the prefrontal cortex developing last (Gogtay, Giedd, Lusk, Hayashi, Greenstein, et al., 
2004).   Therefore, cognitive development follows the sequence of functional developmental 
milestones, such as primary motor and sensory development, then spatial orientation, speech, and 
language development.  Finally, the executive function and attention areas seem to be the last to 
develop.  Evolutionarily, this order makes sense because the least important skills for survival 
(EF) develop last, preceded by more basic human functions.
Barkley?s and Nigg?s theories provide a framework to examine the literature to-date that 
explores links between ADHD and EF, specifically in children.  Developmental considerations 
are particularly important in examining the reliability and validity of using EF tasks because of 
the wide variety of measures used.  Additionally, some EF tasks seem to measure abilities that 
are not yet formed in younger children.  Therefore, it is critical to understand the applicability of 
such tests within a developmental framework before using them for research or diagnostic 
purposes.
12
ADHD is described as a disorder of developmental deficits of EF, such that children may 
not perform comparably to their same-age counterparts on various measures of EF.  This is not to 
say that individuals with ADHD will never have the EF skills, but rather, that they will likely fall 
behind their peers in EF development.  Furthermore, deficits of EF implicated in children with 
ADHD may vary across the age groups, and distinct deficits may manifest themselves at 
different ages or developmental stages.  For example, Brocki and Bohlin (2006) found that 
younger children?s (ages 6.0 to 9.7) ADHD symptoms were related to their performance on tasks 
of inhibition; whereas, these tasks did not predict ADHD symptoms in the older age group (9.8 to 
13.0 years).  The older group?s ADHD symptoms were significantly related to verbal working 
memory, fluency, and speed/arousal, while these EFs did not predict ADHD symptoms in the 
younger group.
13
Associations between EF Measures of Behavioral Inhibition and Planning/Reconstitution and 
ADHD in Children
An underlying EF deficit in ADHD that spans across both Barkley?s (2006) and Nigg?s 
(2006) models is that of inhibiting prepotent, automatic responses for more appropriate, planned 
responses.  Barkley describes this construct as behavioral inhibition whereas Nigg terms it 
?control of motor response and behavior.?  When approaching goal-directed behavior, a person is 
required to inhibit inappropriate responses while at the same time keeping in mind the rules of 
the task and planning more accurate, situation-specific responses.  Therefore, reconstitution and 
planning are also an important part of goal-directed behavior.  A number of research studies 
suggest EF deficits in both behavioral inhibition and reconstitution/planning in individuals with 
ADHD, and these will be described below.  Because of the natural link between these two 
aspects of EF, they have been chosen to be the main focus of this current study.  
One difficulty with studying EF is that many tasks overlap by measuring several EFs. 
Due to the focus of the current study, EF measures have been chosen based on their historical 
and theoretical association with behavioral inhibition or reconstitution/planning.  Specifically, 
the D-KEFS (Delis, et al., 2001) incorporates a standardized scoring procedure across various 
well-established tasks of EF that will be reviewed below.  Furthermore, the D-KEFS produces 
scores for separate tasks within subtests, which allows for description of both EFs, as well as the 
more basic non-EF cognitive processes.  Therefore, separate processes within subtests and their 
relationship to ADHD can be explored.  Given significant results from research with older 
measures of EF and ADHD, the next logical step is to explore the relationship between the D-
14
KEFS and ADHD, since the subtests on the D-KEFS are derived from established EF tasks. 
Two extensive meta-analyses were recently conducted with a variety of EF measures and 
their relationship to ADHD (Frazier, Demaree, & Youngstrom, 2004; Willcutt, Doyle, Nigg, 
Faraone, & Pennington, 2005).  Frazier et al. (2004) analyzed 123 studies that each included a 
control group and an ADHD group (diagnosed by DSM-III, III-R, or IV criteria).  Effects were 
combined for all ADHD subtypes, and weighted mean effect sizes for the 11 reported measures 
of EF ranged from small to medium, with only one from the CPT falling in the large range. 
Willcutt et al. (2005) analyzed a total of 83 studies that included 13 measures of EF. 
Standardized effect sizes were calculated for a group of all ADHD subtypes (diagnosed by DSM-
III, III-R, or IV criteria, ICD-10 criteria, or rating scales) versus a control group, and all effect 
sizes fell within the medium range.  Weighted mean effect sizes for all EF variables from both 
meta-analyses are included in Table A1, and individual results will be discussed within each 
section below as the evaluation measure is presented.  The predictive utility of several EF 
variables are included in Table A2.
Behavioral Inhibition (Control of Motor Response and Behavior)
Continuous Performance Tasks.  Several studies have examined behavioral inhibition in 
children with ADHD using continuous performance, go/no-go, and stop tasks that require motor 
inhibition and the ability to control response perseveration.  Continuous performance tasks 
(CPT), typically administered on a computer, examine motor inhibition, perseveration, and 
attention.  There are several variations of the task, including the Stop signal and Go/No-go 
measures.  The Stop Signal Task (Logan, Schachar, & Tannock, 1997) requires participants to 
press a key when the letter X or O appears on the screen (go trial) and inhibit responding when a 
tone is presented immediately prior to the letter (stop trial).  Another example of one of these 
15
measures is the Conners? Continuous Performance Test-II (CCPT-II; Conners, 1995) in which 
examinees are required to respond to all letters that appear on the screen except for the letter 
?X,? for which a response should be inhibited.  The Go/No-go task requires motor responses to 
red and blue colored squares appearing on the screen, such that the participant must determine a 
rule for responding (i.e. response to red squares, no response to blue squares) (Drewe, 1975).  
In an epidemiological study, 816 children completed the CCPT (Conners et al., 2003). 
Twenty-one children met criteria for diagnosis of ADHD, and their performance was compared 
to that of nondiagnosed children.  Results revealed significantly worse performance by the 
ADHD group on Hit Reaction Time Standard Error (RT SE), errors of commission, errors of 
omission, and detectability (d?).  Reaction Time (RT) and Response Style (B) were also reported, 
as they were considered by the authors as primary response measures; however, no differences in 
group performance were noted on these two variables.  Explanation of the CCPT variables can 
be found in Table 4, where they are outlined for use in the current study.  Relationship between 
CCPT variables and specific ADHD symptoms was explored for the same sample by Epstein et 
al. (2003), with Hit RT SE, commission, omission, detectability, and response style 
demonstrating relationships across the ADHD symptom domains (inattention, hyperactivity, and 
impulsivity).  Detectability and response style showed the strongest relationship with ADHD 
symptoms. 
In a rare study that included all female participants (Hinshaw, Carte, Sami, Treuting, & 
Zupan, 2002), the CCPT was administered to girls ages 6 to 12 years with ADHD-C (n=93), 
ADHD-PI (n=47), and an age- and ethnicity-matched control group (n=88). Results indicated 
that the ADHD-C and ADHD-PI groups performed significantly worse than controls on CPT 
commission errors.  Furthermore, the ADHD-C group had significantly more commission errors 
16
than the ADHD-PI group.
Wodka and colleagues (2007) administered three variations of go/no-go tests to children 
with ADHD (n=58; CT=35, HI=3, PI=20; 52.4% male) and controls (n=84; 62.1% male).  One 
task was the typical ?motor? go/no-go task, whereas the other two required either more cognitive 
effort (refrain from pressing the button when the red spaceship was preceded by an odd number 
of green spaceships) or was motivation-linked (a screen indicated whether points were gained or 
lost based on accuracy and speed).  Covarying age, children with ADHD had more commission 
errors and had greater reaction time variability than controls on the basic go/no-go task. 
However, no differences were found for omissions or response time, and no gender differences 
were noted.  The cognitive go/no-go task produced a significant difference in commission errors 
for the ADHD group, and no effects for omissions, response time, and reaction time variability 
were noted.  Furthermore, gender differences were noted such that boys made significantly more 
commission errors and had greater reaction time variability than girls.  On the motivational 
go/no-go task, ADHD children made more omissions and commissions and showed greater 
reaction time than controls.  No significant differences were noted for any of the measures 
between ADHD subtypes.
Meta-analyses have been conducted on a variety of inhibition measures.  CPT 
commission errors (responding to the stimulus when inappropriate) and omission errors (not 
responding to the stimulus when response is required) indicated moderate effect sizes in 
differentiating between ADHD and non-ADHD groups (Frazier et al., 2004; Willcutt et al., 2005; 
see Table 1).  CPT hits (number of hits recorded) received the largest weighted mean effect size 
of all EF measures evaluated (Frazier et al., 2004).  However, all CPT variables analyzed by 
Frazier et al. (2004) had heterogeneous effect sizes, indicating that the results were inconsistent 
17
across studies.  Stop-signal task reaction times produced medium effect sizes across studies, with 
homogeneous distribution (see Table 1).  Therefore, according to these meta-analyses, stop-
signal tasks may be more efficient and accurate in differentiating between ADHD and control 
groups than the CPT, as the effects were more homogeneous across studies.
A different, non-computerized, continuous performance task was used in a study 
comparing 6 girls and 19 boys with ADHD CT to age and sex matched controls (Klimkeit, 
Mattingley, Sheppard, Lee, & Bradshaw, 2005).  This selective reaching task required responding 
as rapidly as possible to a target, which could appear to the left or right of fixation, while at times 
having to ignore a simultaneous distractor.  Scoring procedures allowed for parsing out of 
reaction and movement times as control measures.  Results indicated that unmedicated children 
with ADHD showed significantly slower reaction times than controls but were not impaired in 
movement execution time. They also made more inattentive (misses) and impulsive (premature 
responses) errors.  In all of these measures, medicated ADHD participants performed similarly to 
control participants, indicating that stimulant medication proved beneficial to task performance. 
However, the medicated group was composed of only 7 children, possibly limiting the 
significance of this finding.
Stroop Color-Word Test.  The Stroop Color-Word Test has been described as a measure of 
selective attention and response inhibition (Stroop, 1935).  Scoring is typically calculated based 
on the time required to complete the task.  Often, three tasks are administered in the Stroop Test: 
naming colors of printed squares, reading names of colors written in black ink, and the 
interference task (which assesses color-word interference skills by presenting a subject with a 
series of names of colors written in ink of non-congruent colors).  In the latter task, the 
participant is required to name the color of the printed ink and inhibit the more automatic 
18
response of reading the word. 
A meta-analysis of seventeen independent studies that implemented the Stroop Color-
Word Test was conducted to determine the overall evidence for deficits in interference control in 
individuals with ADHD aged 6-27 years (van Mourik, Oosterlaan, & Sergeant, 2005).  The 
analysis took into account only studies that used an interference score calculated in one of two 
ways: subtracting the color-word card score from the score on the color card (C ? CW) and the 
Golden method in which correction for color naming and word reading is taken into account 
(CW score ? [(W score * C score) ? (W score + C score)] (Golden, 1978).  Participants in each 
study included at least one ADHD group and a comparison group of normal controls.  Results of 
the meta-analysis indicate a small (.35) but significant effect size for interference scores and a 
heterogeneous distribution of effect sizes across studies, with eight studies having an effect size 
of zero.  Therefore, the authors conclude that the Stroop Test may not be a valid measure of 
interference or inhibitory control deficits in ADHD.
Van Mourik et al. (2005) attribute the lack of consistency in findings comparing Stroop 
performance in individuals with and without ADHD to the method used to calculate interference 
scores.  They found the effect size for C ? CW non-significant and heterogeneous (.26; p = .04) 
and the effect size for the Golden method as significant but also heterogeneous (.40; p = .01).  In 
terms of ADHD subtypes, three studies compared children with ADHD PI and ADHD CT 
(Houghton, Douglas, West, Whiting, Wall, et al., 1999; Nigg, Blaskey, Huang-Pollock, & 
Rappley, 2002; Scheres, Oosterlaan, Geurts, Morein-Zamir, Meiran, et al., 2004), all finding no 
differences.  However, results of the meta-analysis indicate a small but significant and 
homogeneous effect size, suggesting that children with ADHD PI perform worse on measured 
interference than children with the CT.  Thus, the Stroop test may be beneficial in distinguishing 
19
between ADHD subtypes even though it does not seem to be able to accurately differentiate 
between ADHD children and controls.  It is possible that individuals with PI have a more 
pronounced deficit associated with this task, and, since many of the studies combine ADHD 
subtypes into one group, this effect may have been lost.  Further research comparing ADHD 
subtypes on the Stroop task could help illuminate the conflicting results.
Reconstitution/Planning
Tower Tests.  Several tests are typically used to examine reconstitution or planning, such 
as tower tests, mazes, trailmaking, and card sorting tasks.  The Tower of Hanoi or Tower of 
London tasks (TOH; Borys, Spitz, & Dorans, 1982; TOL; Culbertson & Zillmer, 1999; Shallice, 
1982) measure planning, problem-solving, inhibition, and working memory using three colored 
beads that can be placed on pegs of three different heights.  Participants are required to match a 
prescribed pattern of beads on the pegs by following set rules, and they are graded on number of 
moves, difficulty of problem, and time taken to make the first move (initial planning time). 
Meta-analyses of the tower tests indicated medium effect sizes (.51 and .69) for distinguishing 
between ADHD and control groups (Willcutt et al., 2005).
A computerized version of the Tower of London (TOL) task, which measures planning, 
problem-solving, inhibition, and working memory, was used to assess 103 children aged 7 to 15 
years with only ADHD and ADHD with comorbid disorders.  Sixty-nine boys and 26 girls were 
included in the study, and results indicated a gradual decrease in initiation time as children got 
older.  Furthermore, boys were found to have a shorter total initiation time than girls, indicating 
that boys? EFs of planning and inhibition might mature at a slower rate than girls.  No significant 
effects in performance on the TOL were noted for comorbid disruptive behavior, anxiety, or 
mood disorders (Sarkis, Sarkis, Marshall, & Archer, 2005).  
20
Trail Making Tests.  Trail Making tests are part of the Halstead-Reitan 
Neuropsychological Test Battery (Reitan & Wolfson, 1985), and they are often used to examine 
set-shifting, sequencing, and mental flexibility.  Two tasks are typically presented, one in which 
the individual is required to use a pencil to connect numbers (part A) and one in which the 
requirement is to connect a series of numbers and letters in ascending order, alternating between 
numbers and letters (part B).  The task demands of the Trail Making Test necessitate visual 
search and motor skills as well.  Since the typical scoring procedure is based on time taken for 
task completion, true measurement of the EFs versus fine motor skills and speed of responding is 
questionable.  
Meta-analyses of the Trail Making Test indicated a small effect size for Trails A (.40) and 
medium effect sizes for Trails B (.59 and .55) when comparing ADHD children to normal 
controls (Frazier et al., 2004; Willcutt et al., 2005).  In an independent study looking at ADHD 
and control twin pairs (ages 8-18 years), results indicated a significant difference between 
controls and ADHD children on Trails B (d=.69) (Bidwell et al., 2007).  Additionally, a small 
effect size (.39) was noted between non-ADHD co-twins and their ADHD siblings.  
Controlled Oral Word Association Test.  The Controlled Oral Word Association Test 
(COWAT; Benton & Hamsher, 1976) examines verbal working memory and fluency skills by 
requiring the individual to generate as many words as possible beginning with certain letters (F, 
A, S) in one minute, a measure of effortful phonemic word fluency.  Since the original task, 
versions measuring semantic fluency have also been developed, in which the individual is 
required to generate items from overlearned concepts, such as animals or things to eat (category) 
(Newcombe, 1969). In addition, a condition requiring simultaneous switching between two 
semantic (overlearned) categories was included on some versions (Newcombe, 1969).  The EF 
21
assessment measure being used in the current study (D-KEFS) includes the category switching 
component of the Verbal Fluency subtest, in which an individual is required to switch back and 
forth between naming fruits and pieces of furniture.  This additional trial requires the EF of 
reconstitution and planning, as the new rules must be kept in mind and responses are planned 
accordingly.  Thus, a review of the COWAT tests as a measure of EF deficits in individuals with 
ADHD is being included, despite the lack of the switching aspect on the most commonly used 
versions of the test.
Brocki and Bohlin (2006) combined Digit Span and COWAT z-scores from their sample 
to form a verbal working memory/fluency composite score, and this composite significantly 
predicted ADHD inattentive (but not hyperactive/impulsive) symptoms.  Separate regression 
analyses were performed for each age group, which were formed by median split of age 
distribution.  Results revealed that verbal working memory/fluency was only predictive of these 
symptoms in the older age group (9.8 to 13 year olds) and not in the younger age group (6 to 9.7 
year olds).  The authors give two hypotheses that could account for these results: 1) the abilities 
are not yet fully formed in the younger children, and there may be little or no variation in scores 
(i.e. tasks are too difficult) or 2) verbal working memory deficits are only apparent in older 
children with ADHD compared to normal controls because their development in this area of EF 
is significantly delayed. 
22
Delis-Kaplan Executive Function System (D-KEFS)
A recently developed measure of EF is the Delis-Kaplan Executive Function System (D-
KEFS). The D-KEFS (Delis et al., 2001) is a nine-test battery that assesses key components of 
executive functions, including flexibility of thinking, inhibition, problem solving, planning, 
impulse control, concept formation, abstract thinking, and creativity in both verbal and spatial 
modalities.  Several of the tests are similar to existing and well-established measures of 
executive functioning described above.  The standardization on a contemporary stratified sample 
of children, adolescents, and adults (n = 1,750) is one of several practical and empirical 
advantages of this instrument.  Standard scores are available for individuals between the ages of 
8 and 89 years, thus reflecting developmental trends and the applicability of tasks for children 
under 8 years of age.  Norms were developed specifically for each age year from 8 years to 15 
years, which allows for more age-specific information on EF and reflects rapid developmental 
changes in EF for this age group.  
The D-KEFS manual provides some evidence of reliability and validity (Delis et al., 
2001).  Internal consistency values included split-half reliability estimates within each subtest 
that ranged from moderate to high for the tasks being used in the current study.  Test-retest 
reliability estimates range from low to high (.20 to .90), suggesting the relative stability of most 
of the constructs being measured over time.  See Table A3 for a summary of scores presented 
throughout the D-KEFS manual.  The authors also noted that practice effects were observed 
across measures, as performance improved from the first to the second testing session for 
virtually all variables.  Alternate forms are available in order to minimize the effects of 
23
readministration.  
Minimal evidence for convergent and discriminant validity was presented in the D-KEFS 
manual.  The authors presented correlations between an early version of the D-KEFS sorting test 
(California Card Sorting Test; CCST) and the Wisconsin Card Sorting Test (WCST) with a small 
sample (n = 23), and they found moderate correlations.  The CCST was also used in several other 
validity studies, described in the D-KEFS manual, in which patients with frontal lobe lesions, 
Korsakoff syndrome, Parkinson?s disease, chronic alcoholism, and schizophrenia all exhibited 
lower levels of performance on the task than healthy adults.  In addition, D-KEFS performance 
was compared to that on the California Verbal Learning Test ? 2nd Edition (CVLT-II) in a sample 
of 292 adults, with generally non-significant correlations suggestive of discriminant validity.   
Since the D-KEFS is a relatively new measure, limited independent research is available 
on its reliability and validity in measuring EF in children and adolescents.  A study comparing 
the Woodcock Johnson III Tests of Cognitive Abilities (WJ COG-III; Woodcock, McGrew, & 
Mather, 2001) and the D-KEFS suggested moderate correlations between measures (Floyd, 
McCormack, Ingram, Davis, Bergeron, et al., 2006).  The Color-Word Interference, Design 
Fluency, Verbal Fluency, and Trail Making tests were administered to 12 high-functioning adults 
and adolescents with autistic disorder or Asperger?s disorder (Kleinhans, Akshoomoff, & Delis, 
2005).  Results indicated significantly lower performance on an aggregated EF score than the 
normative sample.  Children with prenatal alcohol exposure have also been found to perform 
lower than normative controls on the Trail Making, Color-Word, Tower, Word Context, Verbal 
Fluency, and Design Fluency subtests of the D-KEFS (Mattson, Goodman, Caine, Delis, & 
Riley, 1999; Schonfield, Mattson, Lang, Delis, & Riley, 2001).  In addition, children ages 8 to 16 
years who met criteria for a fetal alcohol spectrum disorder, performed significantly lower than 
24
the normative mean on the Trail Making letter sequencing and number-letter switching tests, the 
Verbal Fluency letter fluency and category switching tasks, and all conditions of the Color-Word 
Interference test (Rasmussen & Bisanz, 2009).
A recent study used four subtests of the D-KEFS (Color-Word Interference, Trail Making, 
Tower, and Verbal Fluency) to predict ADHD in boys and girls aged 8-16 years (Wodka, Loftis, 
Mostofsky, Prahme, Larson, et al., 2008a).  Comparison of ADHD (n = 54; 59% boys) versus 
control (n = 69; 51% boys) groups revealed poorer performance by the ADHD group on only the 
Color-Word Interference (Color Naming, Word Reading, Inhibition, and Inhibition Switching) 
and Tower (total achievement) tests.  However, even though ADHD participants performed 
worse on these subtests, all scores were still within average range for both control and ADHD 
participants.  Furthermore, prediction of ADHD in girls resulted in 67.3% being accurately 
classified based on contribution from scores on the Tower test and Trail Making test.  Boys with 
ADHD were accurately classified at 58.5%, solely based on the Combined Color Naming + 
Reading Composite score on Color-Word Interference.  
Another study analyzed the ?process? scores from the same participants? performance on 
the D-KEFS (Wodka, Mostofsky, Prahme, Larson, Loftis, et al, 2008b).  They found no 
significant group, sex, or ADHD subtype effects; however, a significant interaction for Verbal 
Fluency Total Repetition Errors emerged.  Specifically, boys with ADHD-HI performed better 
than girls with ADHD-C.  On the other hand, girls with ADHD-PI performed better than boys 
with ADHD-PI.  Overall, the authors concluded that these subtests of the D-KEFS did not 
provide sufficient predictive power to be used alone in identifying ADHD.  Furthermore, the 
process scores did not prove useful in distinguishing between ADHD and control groups. 
Several limitations to their sample included the use of a ?pure ADHD? sample, thus ruling out 
25
participants with all other often comorbid disorders except Specific Phobias and Oppositional 
Defiant Disorder.  Additionally, the majority of participants were Caucasian, limiting the 
generalizability of findings across ethnic groups.  Finally, both groups had above-average IQ 
scores, and the authors cited literature suggesting lower sensitivity of EF measures in individuals 
with higher IQs.  
Another study examined EF performance (which included some D-KEFS subtests) 
between typically developing children, those with ADHD, and those with Autism Spectrum 
Disorders (ASD) (Corbett, Constantine, Hendren, Rocke, & Ozonoff, 2009).  MANCOVAs were 
conducted using IQ as a covariate.  Children with ASD performed significantly worse than 
controls on the Color-Word Interference Inhibition and Letter Fluency tasks.  The ASD group 
also performed significantly worse than ADHD participants on the Verbal Fluency Category 
Switching condition.  No differences between ADHD and control groups were noted on the 
above measures, nor on the Category Fluency condition. 
 Based on the standardization across EF tasks that are similar to older, more widely used 
EF measures, the use of the D-KEFS to study EF deficits in ADHD seems warranted.  The D-
KEFS allows for intra- and inter-individual comparisons across tasks that tap into distinct EFs, 
hopefully providing a clearer picture of the relationship between EF and ADHD.  ?Contrast? and 
process scores are generated by the D-KEFS scoring software, which describe various aspects of 
performance on each D-KEFS test.  Therefore, the utility of the D-KEFS should be higher than 
that of previous measures of EF that did not separate the tasks into specific components. 
However, further research is warranted to establish the utility of the D-KEFS in examining EF, 
particularly in a sample of children with ADHD.  Independent research, as well as evidence from 
the D-KEFS manual, suggests adequate psychometric properties of the variables to be used from 
26
this measure.  Therefore, the current study built upon these findings to further explore the utility 
of the D-KEFS.  
27
Rationale for the Current Study
Several limitations in the current literature linking ADHD and EF deficits should be 
noted in order to focus future research in a direction that will be most beneficial to the field.  
First, measures of EF seem to have adequate positive predictive power (i.e. probability of having 
the disorder if exhibiting the symptoms).  However, they also have poor negative predictive 
power, suggesting that individuals with ADHD may not exhibit all of the associated symptoms. 
In this case, EF performance is expected to be lower in people who are diagnosed with ADHD. 
In addition, many of the EF tests exhibit low sensitivity, indicating that ADHD diagnosis does 
not always predict abnormal EF scores.  Finally, specificity seems to be good in that individuals 
with normal test scores typically do not meet criteria for diagnosis.  
Table A2 highlights these predictive statistics for several EF tasks in which cutoff criteria 
were based on ADHD scores that were 1.5 SD above or below the mean of controls (Doyle, 
Biederman, Seidman, Weber, & Faraone, 2000; Grodzinsky & Barkley, 1999).  Grodzinsky and 
Barkley?s study (1999) had a sample of 66 boys with ADHD and 64 control boys, all aged 6-11 
years.  The Doyle et al. (2000) study included boys ages 6 to 17 years with ADHD (n=113) and 
without ADHD (n=103).  Neither study specified ADHD groups by subtype.  Judging by these 
results, EF deficits do not seem to be necessary nor sufficient in order to render a diagnosis of 
ADHD, given the current EF testing measures? poor discriminative validity (Seidman, 2006). 
Whereas 5-10% of controls typically exhibit abnormal EF scores across measures, 30-50% of 
individuals with ADHD have been found to display some type of EF deficit (Doyle et al., 2000). 
Further research is necessary to determine if there is a subtype of ADHD that includes EF 
28
deficits or if the current tasks designed to measure EFs are not tapping into the areas that are 
most specifically impaired in individuals with ADHD.  The D-KEFS will hopefully shed some 
light onto this matter, as it generates scores for individual components of each subtest that might 
allow for better discrimination between ADHD and control groups, suggesting that components 
rather than broad processes are contributing factors. 
Many of the instruments used to measure EF in children pose potential problems for 
research in this area.  A significant conceptual issue with EF tasks is that some of them require 
the use of several skills, such as the EFs of planning, organization, or inhibition but also other 
skills like reading, visual and auditory processing, and motor responding.  Furthermore, 
numerous measures and tasks are available to test EFs (as evident by the review above), which 
poses a problem for consistency and consensus across researchers.  Some newer measures, such 
as the D-KEFS (Delis et al., 2001), include scores on separate tasks within subtests, such as 
reading, speed of processing, and similar concepts that can be taken into account for analyses to 
better differentiate between more basic processes or higher order EF processes.  
Another limitation with research in this area is that, across studies, identical EF measures 
(e.g. the Stroop test) may be scored using various techniques, or scores may be aggregated or 
manipulated in some other way prior to analyses that makes comparison of results difficult.  Use 
of measures with standardized scoring procedures and multi-domain scores that can be compared 
intra-individual would alleviate this issue.  Thus, the D-KEFS is a logical choice to address 
previous scoring irregularities in the literature.
Furthermore, most EF assessment tools used with children were originally designed for 
adults.  Developmental considerations are critical in this area because of the research cited earlier 
that clearly shows that EF capabilities become more refined and efficient throughout childhood 
29
and adolescence.  Therefore, measures originally created to assess adult levels of EF may not be 
sensitive to the more subtle differences that develop between age groups in children.  In addition, 
some tasks, such as the Stroop test, require non-EF abilities like reading, which eliminates any 
utility this test would provide in examining EFs in children who are too young to read.  The D-
KEFS standardization sample included children of each age level from 8 years to 15 years, thus 
establishing individual age-based norms.  Age-based norms are critical in understanding 
developmentally appropriate performance on a measure, and the D-KEFS aptly takes this into 
account.  Previous literature suggests that skills for some tests of EF, such as the Stroop test, do 
not develop until age 7 or 8 years. Thus, the developers of the D-KEFS did not norm their 
measure on children younger than 8 years.
The current study takes into account the limitations of EF and ADHD research to-date, 
exploring EF performance in children with and without ADHD.  Specifically, the D-KEFS, a 
relatively new assessment battery, was used.  The D-KEFS incorporates well-established 
measures of EF, has standardized scoring, and has been normed on an age-stratified sample. 
Furthermore, performance on a basic test of behavioral inhibition (CCPT-II) was compared with 
that on a measure of more complex EFs (D-KEFS) to determine concurrent as well as 
incremental validity in predicting ADHD.  The measures used in this study have been normed on 
children as young as 8 years old, which suggests that children of that age are able to comprehend 
and complete the tasks.    Finally, the D-KEFS provides scores for the more basic tasks (e.g. 
motor speed, reading) within subtests that can help parse out the influence of basic skills on the 
more complex EF tasks.
30
Hypotheses for the Current Study
Several hypotheses were tested in this study.  First, the relationship between diagnostic 
group and performance on EF measures was explored.  Specifically, it was expected that 
individuals with ADHD would perform significantly worse than controls on measured variables 
of complex EF (inhibition/switching), but not on measures of the less executive aspects of a task 
(reading, number sequencing, etc.).  Furthermore, more severe ADHD symptoms, as measured 
by the CPRS, were expected to be associated with more deficient EF performance. CCPT-II 
performance on all variables measured was expected to be worse for the ADHD group, as 
various research groups have found CPT performance deficits across variables (Conners et al., 
2003; Frazier et al., 2004).
Next, concurrent validity between D-KEFS ?inhibition? measures and the CCPT-II was 
evaluated.  It was hypothesized that scores measuring inhibition on the D-KEFS and scores on 
the CCPT-II would be moderately correlated.  
In addition, utility of all study measures in predicting ADHD versus control group 
membership was explored.  It was expected that the D-KEFS variables that include inhibition 
and planning components would significantly contribute to group prediction.  Specifically, 
results from Wodka et al. (2008a) suggest that scores from the Tower Test and Color-Word 
Interference subtests would significantly predict groups.  It was also expected that CCPT-II 
scores would accurately predict group membership based on previous research that found 
omission errors, commission errors, and hits to significantly differentiate between ADHD and 
controls (Frazier et al., 2004; Willcutt et al., 2005).  Additionally, it was hypothesized that 
31
combining D-KEFS and CCPT-II scores would increase the utility of predicting ADHD versus 
control group membership.
Finally, analyses were conducted to explore the prediction of ADHD symptoms based on 
age.  Given Brocki and Bohlin?s (2006) research, it was expected that younger children?s ADHD 
symptoms would be more related to performance on inhibition measures; whereas older 
children?s ADHD symptoms would be predicted by verbal working memory, fluency, and 
speed/arousal scores.   
32
Method
Participants
Participants included children with and without ADHD ranging in age from 8 years to 14 
years.  The use of this age range was to facilitate a sample that might allow for age group 
comparisons and sufficient distribution characteristics.  Eight years was selected as the lowest 
end of the spectrum due to D-KEFS norms beginning at that age.  Fourteen years was selected at 
the highest end of the spectrum as the sample would then extend through elementary and middle-
school age groups, and would exclude individuals in junior high and high school.  Typically, 
adolescents begin high school around the age of 15 years, and recruitment of the high school 
population is difficult due to increased independence and higher levels of involvement in 
extracurricular activities.  Therefore, in expectation of possible recruitment difficulties with older 
adolescents, it was decided to stop recruitment at age 14 years.  
Recruitment occurred via flyers distributed to local agencies, including physicians? 
offices, gyms, recreation centers, daycare centers, elementary schools, and mental health 
professionals.  Furthermore, participants in the study were given flyers to pass on to friends or 
family members who might be interested.  Approval was obtained from the Auburn University 
Institutional Review Board.  
Parents or guardians who were interested in their children participating in the study 
completed a screening packet, which included a demographic questionnaire (see Appendix B) 
and the CPRS-R:L (Conners, 2000).  Inclusion criteria for the ADHD group required 
endorsement of a prior ADHD diagnosis and scores 1.5 or more standard deviations above the 
33
mean on the CPRS-R:L DSM-IV Inattentive Symptoms Scale and/or DSM-IV 
Hyperactive/Impulsive Symptoms Scale.  Children who were currently taking short-acting 
psychostimulant medication for ADHD were eligible for the study, and their parents/guardians 
were instructed to complete the CPRS-R:L based on the child?s behavior while unmedicated. 
Membership in the control group required a child to have no prior diagnosis of ADHD and scores 
no more than 1 standard deviation above the mean on the CPRS-R subscales noted above. 
Exclusion criteria for both groups included current use of any long-acting psychotropic 
medication, uncorrected vision, history of seizures, brain damage, or psychosis.  
Of note, ADHD assessment practices typically include a comprehensive evaluation for 
clinical purposes but have focused on a DSM-based norm-referenced scale for research purposes 
(Pelham, Fabiano, & Massetti, 2005).  In addition, previous research using ADHD rating scales 
for diagnostic purposes has primarily been validated with maternal parent report, and no 
incremental utility has been found when combining parent ratings with structured interviews 
(Pelham, et al., 2005).  Therefore, the current study used previously validated rating scale 
procedures in verifying ADHD symptom criteria and attempted to garner maternal report unless 
the child was not in the mother?s custody.
Thirty-two children with ADHD and thirty-seven children without ADHD met study 
criteria and completed the laboratory portion of the study.  One Caucasian female ADHD 
participant was excluded from analyses because her full-scale IQ (standard score = 134)resulted 
in violating the assumption of homogeneity of variance of IQ scores across diagnosis and sex 
defined groups.  All statistical analyses were run with and without her in them, and no significant 
changes were noted in the results.
Table 1 presents diagnostic and sex group descriptive statistics.  Of the 32 children with 
34
ADHD, scores on the CPRS indicated 5 displaying significant level of symptoms consistent with 
meeting criteria for ADHD-H, 4 for ADHD-I, and 23 for ADHD-C.  MANOVA produced 
significant diagnostic group differences on mean CPRS scores for the Inattentive symptoms and 
the Hyperactive/Impulsive symptoms subscales, with parents of children with ADHD reporting 
higher symptom levels, F(2, 66) = 201.7, p < .001 (Table 2).
Control participants were matched as best as possible for age and sex.  The ADHD group 
included 27 males and 5 females, with a mean age of 10.8 years (SD = 1.97). The control group 
included 28 males and 9 females, with a mean age of 11.1 years (SD = 1.98).  No diagnostic 
group differences were found for age, t(67) = -0.60, p > .05, and proportion of sex, ?2(1, N = 69) 
= 0.80, p > .05.  The ADHD group included 18 Caucasian and 14 African-American participants; 
whereas, the control group included 31 Caucasian and 6 African-American participants.  This 
resulted in a significant between-group difference in the proportion of race, ?2(1, N = 69) = 6.32, 
p < .05.
Full-scale IQ, as measured by the WASI, produced significant diagnostic group 
differences, t(67) = -3.53, p = .001, with the ADHD group mean of 99.1 (SD = 15.8) and the 
control group mean of 112.9 (SD = 16.6).  In addition, significant IQ differences by race were 
noted, t(67) = 4.96, p <.001, as the mean IQ for Caucasian participants was 112.2 (SD = 16.0) 
and African American participants was 92.4 (SD = 12.6).  However, the interaction between 
diagnostic group and race for IQ was not significant, F(3,65) = .388, p = .762.  
35
Table 1.  Diagnostic Group Descriptives
Table 2.  CPRS Group Comparison
CPRS Scale ADHD Group 
Mean (SD)
Control Group
Mean (SD)
Univariate 
F
Inattention T-score 72.9 (8.4) 46.3 (5.3) 254.7*** 
Hyperactivity-
Impulsivity T-score 78.0 (11.3) 47.2 (4.2) 237.4***
*** p < .001
Measures
Conners? Parent Rating Scales-Revised, Long Form. (CPRS-R:L).  The CPRS-R:L 
(Conners, 2000) is a norm-referenced questionnaire that assesses for core DSM-IV symptoms of 
ADHD and related problem areas.  Reviews of evidence-based ADHD assessment indicate that 
the CPRS-R:L is one of the most widely used rating scales for this purpose and has the benefits 
of a large standardization sample (2,482 children and adolescents; 3 to 17 years), ease of 
administration and interpretation, and sound psychometric properties (Collett, Ohan, & Myers, 
2003; Conners, 2000).  Specifically, test-retest reliability coefficients after 6 to 8 weeks ranged 
from .73 to .94, and internal consistency Cronbach?s alpha coefficients ranged from .47 to .85 
(Collett et al., 2003).  In addition, the CPRS-R technical manual reports on acceptable levels of 
internal, convergent, and divergent validity, as well as discriminative validity between children 
with and without ADHD (Conners, 2000).  Subscales used in this study included DSM-IV 
36
ADHD Group (n =32)
Mean (SD)
Control  Group (n = 37)
Mean (SD)
Sex Male (n=27) Female (n=5) Male (n=28) Female (n=9)
Age in years 10.8 (0.40) 9.9 (0.90) 11.2 (0.38) 10.5 (0.67)
Verbal IQ 102.2 (16.2) 85.0 (8.1) 111.9 (18.5) 115.7 (13.2)
Performance IQ 100.8(16.0) 88.2 (13.3) 109.7 (14.4) 112.3 (15.4)
Full Scale IQ 101.7 (15.4) 85.0 (10.2) 111.9 (17.2) 115.9 (15.1)
CPRS ? Inattention 
T-score
72.0 (6.8) 77.6 (14.7) 46.3 (5.6) 46.3 (4.2)
CPRS ? 
Hyperactivity-
Impulsivity T-score
76.6 (11.3) 85.2 (9.1) 47.0 (4.3) 47.7 (4.3)
Inattentive Symptoms Scale, the DSM-IV Hyperactive/Impulsive Symptoms Scale, and the DSM-
IV Total ADHD Symptoms Scale.
Conners? Continuous Performance Test-II (CCPT-II).  The CCPT-II (Conners, 1995) is a 
computerized task that was administered on a Windows desktop computer (17 in (43.18 cm) 
monitor, 1,024 x 768 resolution, 75 Hz refresh rate).  Three-hundred and sixty letters 
(approximately 1 inch high) appear on the screen, one at a time, for approximately 250 ms.  The 
participant is required to press the spacebar for each letter as it appears on the screen and to 
refrain from pressing the spacebar when an ?X? appears.  The 360 letter trials are divided into 18 
blocks, each consisting of 20 trials.  Each block is randomly assigned to one of three 
interstimulus interval (ISI) conditions (1, 2, or 4 s), and all three ISI conditions occur every three 
blocks.  Across all blocks, the percentage of trials in which letters other than X appear is 90%. 
The test administration duration was 14 minutes.  Variables used are those outlined by the task 
authors as the ?standard set of performance measures? calculated in most continuous 
performance tasks (Conners, Epstein, Angold, & Klaric, 2003).  These include variables of 
accuracy: commission errors (responding to a non-target stimulus), omission errors (failing to 
respond to target stimulus), detectability (d?, signal detection or perceptual sensitivity to targets), 
and response style (B, tendency to respond too little or too much relative to the actual 
distribution of the signal).  Descriptions of variables are presented in Table 3.  T-scores ? 65 
indicate markedly impaired functioning (Conners, 1995).
37
Table 3. CCPT-II Variables
Score Measures Feature
Omissions Not responding to target (non-X) stimuli Accuracy/Inattention
Commissions Responding to non-target (X) stimuli Accuracy/Inattention & 
Impulsivity
Detectability (d?) Derived from distance between target distribution 
and non-target distribution (greater distance = 
better detection of targets).  Is dependent upon 
frequency of target and sensitivity of respondent, 
and reflects the ability to discriminate among 
stimuli.  
Accuracy/Inattention
Response Style (B) Derived from response frequency compared to 
actual target vs. non-target distribution. 
Calculated by comparing if the participant 
responds too little or too much versus the actual 
distribution of target stimuli. Reflects the degree 
to which respondent is being conservative or 
impulsive in responding, which affects likelihood 
that the correct response will be made.
Accuracy/Impulsivity
Hit Reaction Time (RT) Mean time taken to respond to all targets Reaction 
Time/Inattention & 
Impulsivity
Hit Reaction Time   
 Standard Error (RT SE)
Variability/consistency of response time for all 
targets
Variability/Inattention 
& Impulsivity
Delis-Kaplan Executive Function System (D-KEFS). The D-KEFS (Delis, Kaplan, & 
Kramer, 2001) is a nine-test battery that assesses key components of executive functions, 
including flexibility of thinking, inhibition, problem solving, planning, impulse control, concept 
formation, abstract thinking, and creativity in both verbal and spatial modalities. Several of the 
tests are similar to existing and well-established measures of executive functioning, and the 
standardization on a contemporary stratified national sample of ages 8 to 89 years is a practical 
and empirical advantage of this instrument. In addition, subtest scores have all been normed to a 
mean of 10 and standard deviation of 3 to enable comparison across measures.  Subtests used in 
this study include: Trail Making Test, Verbal Fluency Test, Color-Word Interference Test, and 
Tower Test.  Test-retest reliability coefficients range across tests from low to high, suggesting 
38
that the constructs most tasks are measuring remain consistent over time (Delis et al., 2001). 
Evidence of sufficient convergent and discriminant validity was described previously and is 
provided in the D-KEFS manual, and scores across the subtests are correlated appropriately 
based on similarity in construct being assessed (Table 3).  Primary measure scores across all 
components of the subtests will be generated by the D-KEFS computer scoring assistant.   
Wechsler Abbreviated Scale of Intelligence (WASI). The WASI (The Psychological Cor-
poration, 1999) is a brief, individually administered test of intelligence designed for individuals 
from age 6 to 89. Standardized on a stratified national sample according to 1997 U.S. Census fig-
ures, the WASI yields a Verbal Scale IQ (VIQ), Performance Scale IQ (PIQ), and a Full Scale IQ 
(FSIQ). Excellent internal consistency and test-retest reliability have been found, and scores on 
the WASI correlate highly with other ability tests (Sattler, 2001; The Psychological Corporation, 
1999). 
Independent Measures
Between-Group.  For both the CCPT-II and D-KEFS, diagnostic group (ADHD versus 
control group) serves as a between-group independent measure.
Dependent Measures
CCPT-II.  Based on prior research with the CCPT-II, several scores on the instrument 
have been found to be associated with ADHD symptoms.  For the current study, analyses focus 
on previously established diagnostically significant variables, which are also identified as the 
main performance by the authors of the task (Conners et al., 2003). 
D-KEFS.  All primary measure scores will be included in the D-KEFS analyses for the 
subtests that were administered.  This includes individual scores for each task within a subtest. 
Contrast and process scores were not included in the analyses because of the limited reliability of 
39
these types of scores as found in Wodka et al.?s (2008) research.  Refer to Table 4 for a list of the 
D-KEFS variables analyzed.  Ingredient scores refer to the tasks being basic skills that are 
combined with others in the more ?executive? condition of the task.
Table 4. D-KEFS Variables
D-KEFS test/variable Type of Measure
Trail Making Test
Primary Measures: Completion Times
Condition 1: Visual Scanning Total ? Ingredient
Condition 2: Number Sequencing Total ? Ingredient
Condition 3: Letter Sequencing Total ? Ingredient
Condition 4: Number-Letter Switching Total - Reconstitution/Inhibition
Condition 5: Motor Speed Total ? Ingredient
Verbal Fluency
Primary Measures
Letter Fluency Total - Reconstitution/Working Memory
Category Fluency Total ? Reconstitution/Working Memory
Category Switching Total - Reconstitution/Inhibition
Category Switching: Total switching accuracy Total - Reconstitution/Inhibition
Color Word Interference
Primary Measures: Completion Times
Condition 1: Color Naming Total - Ingredient
Condition 2: Word Reading Total - Ingredient
Condition 3: Inhibition Total - Inhibition
Condition 4: Inhibition/Switching Total - Reconstitution/Inhibition
Tower Test
Primary Measure
Total Achievement Score Total - Reconstitution/Inhibition
Procedure
Prior to participation, parents/guardians (preferably maternal) completed a screening 
packet that included a consent form, demographic questionnaire, and CPRS-R:L.  Children 
signed the consent form as well, giving their assent to participate.  After the forms were received, 
they were scored by a research assistant to determine participant eligibility.  Children for the 
control group were selected with age and gender corresponding as closely as possible to the 
ADHD participant characteristics.  Parents/guardians of children deemed eligible were contacted 
to schedule a laboratory research session.
40
The D-KEFS, WASI, and CCPT-II were administered in a counterbalanced order with 
two other computerized tasks not analyzed in the current study.  Specifically, either the two paper 
and pencil tasks or the three computer tasks were administered together first, so children did not 
switch back and forth between mode of testing.  Within each mode of testing (computer or paper 
and pencil), measures were completed in a counterbalanced order.  The one-time research session 
was approximately 2 ? hours in duration, and 5-minute breaks between each task were 
implemented.  Testing took place in a well-lighted room, free from distractions.  A graduate 
student or research assistant remained in the room with the child at all times during testing, and 
parents were not allowed to be in the room or to observe the research session to maintain control 
of surroundings.  For children currently taking psychostimulant medication for ADHD, parents 
consented for them to abstain from their medication on the day of the research session.  Prior to 
beginning the research session, parents and children verbally confirmed adherence to this 
procedure.
For the CCPT-II, a practice administration was administered prior to the test 
administration.  On-screen instructions were read out-loud to the participant, instructing him/her 
to press the spacebar for every letter presented except the letter X.  The participant was then 
required to paraphrase the instructions prior to beginning the task.  During the practice phase, for 
each commission error the participant made (pressed the spacebar for the X), the following 
prompt was given: ?Remember, don?t press for the X.?  This prompt was not given during the 
test phase.
The WASI and the four selected subtests from the D-KEFS were administered at a table, 
with the examiner seated directly across from the participant.  Standardized administration 
procedures as outlined in each test?s manual were followed.  
41
Upon completion of the laboratory phase, parents of participants were compensated $50. 
In addition, children received gift certificates (worth $10 in total) for free pizza, ice cream, 
movies, kids? meals, or bowling.  No compensation was given for solely completing and 
returning the research packet, and this was indicated clearly on the consent form.
42
Results
Statistical Considerations
A review of literature regarding appropriate statistical analyses in neuropsychological 
research revealed several key points.  Specifically, Tupper and Rosenblood (1984) discuss the 
issues of using attribute variables (sex, age, education, socioeconomic status, diagnosis) in 
research.  Attribute variables are not randomly assigned, as subjects are likely self-selected.  In 
addition, attribute variables are characteristics that the participant brings to the study, and these 
variables cannot be manipulated.  Therefore, statistical procedures used in true experimental 
design may not be appropriate for studies that do not manipulate a condition experimentally and 
have random assignment of participants to groups.  Most attribute variables are conceptually 
related in the real world, which makes it difficult to make clear or succinct conclusions regarding 
group differences due to possible confounding variables.  Attempts to match groups post-hoc on 
some variables are undesirable because it risks un-matching the groups on a potential 
confounding or unconsidered variable, may create a sub-sample that differs from the population 
of interest, and creates confusion in the ability to conclude causation (Meehl, 1970).  
Next, use of ANOVAs, despite their prolific use in neuropsychological research in peer-
reviewed journals (e.g. Epstein et al., 2003; Wodka et al., 2008a; Wodka et al., 2008b), is 
discouraged because of the strong relationship between attribute variables and 
neuropsychological status, which violates one of the basic assumptions of the ANOVA model 
(Miller & Chapman, 2001; Tupper & Rosenblood, 1984).  Analysis of covariance (ANCOVA) is 
often used to ?statistically control? for effects of a nuisance variable.  However, ANCOVA 
43
should only be used when the nuisance variable is statistically independent of the 
outcome variable (Tabachnik & Fidell, 2007).  In the current study, this rationale was 
used to determine that analyses which likely included impact from nuisance variables 
would not be appropriate for ANCOVA.  However, analyses where no nuisance variables 
were evident reverted to using traditional ANOVA.  
 Relationship between attribute variables is common.  For example, males are 
generally more likely than females to meet diagnostic criteria for ADHD, with a ratio of 
3:1 in community samples and a range of 5:1 to 9:1 in clinical samples (American 
Psychiatric Association, 2000).  When individuals self-select to participate in a study, the 
number of males versus females may not match the naturally occurring ratio. 
Furthermore, cell sizes in analyses are likely to be unequal, which can lead to low 
statistical power.  Some researchers attempt to solve this problem by over-sampling the 
less prevalent individuals to create equal cell sizes for analysis.  However, this method is 
discouraged because it ignores the conceptual problem that the independent variables are 
related (Tupper & Rosenblood, 1984).  Since MANOVA and MANCOVA are based on 
comparison between groups (cells), these methods are discouraged when using attribute 
variables.  
After review of the relevant literature, several options were explored for the most 
appropriate statistical analyses in the present study.  Due to race and diagnostic group 
differences, consideration was given to using IQ as a covariate in a MANCOVA. 
However, the high correlation between IQ and D-KEFS variable scores violates an 
assumption regarding this type of analysis ? that the covariate should not be moderately 
or highly correlated with the independent variables (Tabachnik & Fidell, 2007).  Race 
44
was also considered as a covariate, but no theoretical rationale linking race and EF could 
be established, thus precluding a decision to covary race solely based on the diagnostic 
group difference in proportion of race.  Another option was to re-match the groups so that 
each had an equal proportion of African American and Caucasian participants.  This 
would produce a smaller sample size but may eliminate the potential of the race/IQ 
interaction confounding subsequent analyses.  However, as noted above, post-hoc 
matching can result in unmatching on other variables; thus, this option was not chosen.
Finally, the literature suggests that the most appropriate analyses in 
neuropsychological research are those of correlations and regressions (Tupper & 
Rosenblood, 1984).  Given that these statistical models emphasize the relationship 
between variables, it is not required for the variables to be independent of each other or to 
be ?controlled for.?  Therefore, the approach taken in this study was based on what 
seemed most appropriate in this line of research.  IQ was used as a predictor, rather than a 
covariate, which eliminated the problem of moderate to high correlations with the D-
KEFS variables. 
Diagnostic Group Performance
D-KEFS.  Data were evaluated for outliers by performing frequency counts and 
viewing histogram plots of all participants.  Outliers would be those scores more than 3 
standard deviations from the mean, and no outliers were noted in any of the D-KEFS 
variables analyzed.  Scaled scores for all dependent variables, divided by diagnostic 
group, are presented in Table 5.  For the ADHD group, variables that had means below 
the average range were Trailmaking Letter-Number Switching and Tower total 
achievement.  All control group variables were within the average range of performance. 
45
Given statistical considerations previously outlined, a MANCOVA was not performed to 
discern diagnostic group differences.  Point biserial correlations were calculated between 
diagnostic group and D-KEFS variables, with the diagnostic variable being coded 1 for 
ADHD and 0 for control (Table 6).  Results indicate low to moderate significant 
relationship between diagnosis and all D-KEFS variables except Trailmaking Visual 
Scan, Verbal Fluency-Letter Fluency total correct, and Color Word Interference-Word 
Reading.  Specifically, as negative correlations suggest, better performance tended to be 
associated with control group status. 
Chi square analyses were performed to determine the difference between ADHD 
and control performance on D-KEFS scores, with scaled scores of 6 or lower being below 
average (8th percentile) and, for consistency in comparisons, matching the percentile of 
clinically significant T-scores from the CCPT.  Scores of 7 and up on the D-KEFS are 
considered average and above.  Scaled score values were divided into two groups: below 
average (6 and under) and average or above (7 and up).  Results indicated significant 
group differences on Trailmaking Letter Sequencing, ?2(1, N = 69) = 6.01, p < .05, Verbal 
Fluency Category Switching Total Switching Accuracy, ?2(1, N = 69) = 4.85, p < .05, and 
Tower total achievement, ?2(1, N = 69) = 9.42, p < .01.  ADHD group scores were more 
likely to be in the below average range than were those for controls.
46
       Table 5.  Group Performance on D-KEFS 
D-KEFS Variable ADHD Group
Mean Scaled Score 
(SD)
Control Group
Mean Scaled Score 
(SD)
Trailmaking Visual Scan 9.7 (3.2) 10.2 (2.9)
Trailmaking Number Sequencing 9.7 (3.0) 11.2 (2.6)
Trailmaking Letter Sequencing 8.1 (4.2) 10.3 (2.6)
Trailmaking Letter-Number Switching 7.0 (3.9) 10.0 (3.8)
Trailmaking Motor Speed 9.5 (2.7) 11.3 (1.7)
Verbal ? Letter Fluency total correct 9.6 (2.8) 10.3 (3.1)
Verbal ? Category Fluency total correct 9.8 (2.4) 12.2 (2.7)
Verbal ? Category Switching total correct 8.3 (2.8) 10.8 (3.0)
Verbal ? Category Switching Accuracy 8.5 (2.9) 11.0 (2.7)
Color Word ? Color Naming 8.9 (2.5) 10.4 (3.1)
Color Word ? Word Reading 10.5 (2.2) 11.2 (2.2)
Color Word ? Inhibition 8.2 (3.2) 10.8 (3.0)
Color Word ? Inhibition/Switching 9.3 (3.0) 10.9 (2.7)
Tower ? Total Achievement 7.5 (3.2) 10.2 (2.4)
       Table 6.  Point Biserial Correlations between Diagnostic Group and D-KEFS 
Variables
D-KEFS Variable  Correlation Coefficient
Trailmaking Visual Scan -.098
Trailmaking Number Seq.  -.263*
Trailmaking Letter Seq.  -.306*
Trailmaking L-N Switching    -.363**
Trailmaking Motor Speed   -.375**
Verbal ? Letter Fluency total correct                 -.120
Verbal ? Category Fluency total correct     -.427***
Verbal ? Category Switching total correct   -.407**
Verbal ? Category Switching Accuracy    -.410***
Color Word ? Color Naming                -.251*
Color Word ? Word Reading                -.164
Color Word ? Inhibition  -.384**
Color Word ? Inhibition/Switching -.289*
Tower ? Total Achievement    -.438***
          (Control = 0, ADHD = 1)                                   * p <.05      ** p <.01     *** p <.001
Correlations among D-KEFS variables are presented in Table 7, which shows low 
to moderate correlations across numerous variables.  Specifically, correlations among all 
tasks within each subtest were all significant, with magnitude of low to moderate.  Table 
A15 presents intercorrelations from the D-KEFS manual.  Magnitude of intercorrelations 
within each subtest were generally consistent with those reported in the manual. 
47
Correlations across subtests in the current study revealed low to moderate relationships 
between most variables.  Compared to the magnitudes reported in the D-KEFS manual, 
those in the current study revealed higher degrees of relationship between subtests.  Most 
notably, the manual indicates Tower total achievement score as a low negative correlation 
with most other measures (range of -.03 to -.13).  The current study found significant 
positive correlations between Tower and most other measures in the low to moderate 
magnitude (range of .24 to .54).  In addition, relationships between Trailmaking and the 
Color-Word Interference and Verbal Fluency tasks were higher than those in the manual. 
It is possible that the current study produced higher intercorrelations because it is a more 
homogeneous sample than that used in the D-KEFS normative research.  Specifically, a 
relative restriction of range in scores in the current study differs from the range of scores 
in the normative sample, which included individuals with severe EF deficits as well as 
those in the normal range of functioning.  
A high correlation (r = .92, p < .001) was noted between Verbal Fluency Category 
Switching total correct responses and Verbal Fluency Category Switching total switching 
accuracy.  Since these variables were so highly related, a decision was made to include 
only the total switching accuracy variable in subsequent analyses on inhibition/switching 
variables to hopefully eliminate a problem of multicolinearity.  This variable was chosen 
given that it is a measure of the ability of a participant to change between categories; 
whereas, the total correct responses does not take into account switching responses. 
Since the current study goal is to examine the executive skills, the switching accuracy 
score is a better estimation of the more complex skill.   
48
Table 7.  Intercorrelations among D-KEFS Variables and IQ
TM=Trailmaking; VF=Verbal Fluency; CW=Color Word Interference                            * p <.05      ** p <.01     *** p <.001
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.
1. Full Scale IQ    -
2. TM Visual Scan .124     -
3. TM Number Seq. .309** .305*     -
4. TM Letter Seq. .521*** .354** .434***     -
5. TM L-N 
Switching
.607*** .449*** .496*** .587***       -
6. TM Motor Speed .297* .321** .312** .351** .407**     -
7. VF ? Letter 
Fluency total correct .420*** .247* .109 .294* .388** .111 -
8. VF ? Category 
Fluency total correct .464*** .233 .263* .169 .304* .020 .400** -
9. VF ? Category 
Switch total correct .658*** .250* .331** .497*** .477*** .379** .413** .432*** -
10. VF ? Category 
Switching Accuracy .689*** .205 .336** .527*** .493*** .377** .389** .410*** .921***     -
11. CW ?Color 
Name
.254* .165 .242* .322** .448*** .163 .284* .054 .282* .228     -
12. CW ? Word 
Read
.224 .352** .202 .277* .376** .175 .265* .186 .316** .243* .679***      -
13. CW ? Inhibition .543*** .236 .443*** .595*** .670*** .222 .287* .252* .515*** .505*** .685*** .508***     -
14. CW ? 
Inhibition/Switching .340** .136 .283* .372** .461*** .250* .225 .181 .247* .223 .558*** .406** .650***    -
15. Tower ? Total 
Achievement .546*** .148 .220 .417*** .442*** .289* .328** .229 .320** .377** .129 .125 .288* .124
49
CCPT. Across all CCPT variables analyzed, three outliers were identified.  Two of these were for 
omissions T-score (1 ADHD and 1 control participant), and one was for detectability (control 
participant). To address these outliers, mean T-score for the diagnostic group was calculated for 
each variable, and the outlier results were recoded as the T-score three standard deviations from 
the mean of the corresponding dependent variable.  T-scores for all CCPT dependent variables 
are presented in Table 8.  Correlations across CCPT variables indicate modest to moderate 
relationship between measures, which is expected given that some of the scores are derived from 
performance on the other variables.  Correlations between CCPT performance and IQ revealed 
significant modest results only for Hit Reaction Time Standard Error (r = -.328, p <.01) and 
Omissions (r =  -.284, p <.05), as shown in Table 9. Therefore, unlike the significant relationship 
between IQ and D-KEFS variables, CCPT performance was largely unrelated to IQ.  No sex 
differences were evident in CCPT scores, F(6,60) = .72, p = .64.  
Table 8.  Group Performance on CCPT Variables
CCPT Variable ADHD Group
T-score Mean (SD)
Control Group
T-score Mean (SD)
F (1,66)
Omissions 55.6 (13.9) 48.9 (10.6) 5.01*
Commissions 49.6 (12.1) 44.5 (12.2) 2.27
Hit RT 54.8 (12.7) 49.9 (10.6) 3.60
Hit RT SE 57.3 (9.9)          49.4 (7.8)   13.48***
Detectability (d?) 52.9 (12.3) 46.2 (12.0)  4.76*
Response Style (B) 58.2 (15.4) 52.7 (14.7) 2.26
* p<.05   **p<.01   ***p<.001
Table 9.  CCPT and IQ Intercorrelations 
Hit RT= Hit Reaction Time; Hit RT SE=Hit Reaction Time Standard Error                 * p<.05   **p<.01   ***p<.001
50
CPT T-score Full Scale
IQ (FSIQ)
Omissions Commissions Hit RT Hit RT SE Detectability
Omissions -.284*
Commissions -.184 .171
Hit RT -.109 .176 -.600***
Hit RT SE -.334** .729*** .124 .531***
Detectability (d?) -.210 .254* .868*** -.465*** .154
Response Style (B) -.200 .306* -.088 .373** .349** .025
A MANOVA was conducted to test mean group differences across CCPT variables (Table 
8).  There was a significant main effect of diagnostic group, F(6,61) = 3.19, p < .01.  Specifically, 
the ADHD group produced more omission errors, with F(1,66) = 5.01, p < .05.  In addition, the 
ADHD group also produced more variability in their hit reaction time, as evidenced by Hit 
Reaction Time Standard Error scores, F(1,66) = 13.48, p < .001.  Finally, the ADHD group had 
more difficulty with Detectability (d?), F(1,66) = 4.76, p < .05.  However, mean T-scores for both 
the ADHD and control group across all variables were in the non-clinical range.  
Chi square analyses were performed to determine if CCPT scores were clinically 
significant (T-score ? 65) for more ADHD than control participants.  Results indicated 
significant group differences, with ADHD participants producing a higher number of clinically 
significant scores than controls on Hit RT, ?2(1, N = 68) = 7.04, p < .01, and Hit RT SE, ?2(1, N = 
68) = 4.15, p < .05.  No significant group differences were found for Omissions, ?2(1, N = 68) = 
2.82, p > .05, Commissions, ?2(1, N = 68) = .56, p > .05, Detectability, ?2(1, N = 68) = 1.48, p > 
.05, and Response Style, ?2(1, N = 68) = 3.62,  p > .05.
Correlations between D-KEFS, CCPT, and CPRS Scores
Bivariate correlations between the D-KEFS and CCPT dependent variables were 
calculated (Table 10).  Given the large number of correlations calculated, discussion in this 
section will focus on those with a coefficient of .01 or better.  CCPT variables that measure 
accuracy were related to several D-KEFS variables.  CCPT omission errors were significantly 
correlated with D-KEFS Verbal Fluency ? Category Switching total switching accuracy. CCPT 
commission errors were correlated with Verbal Fluency ? Category Switching total switching 
accuracy.  CCPT Detectability was significantly related to Trailmaking Number Sequencing. 
51
CCPT Response style was not significantly correlated at the .01 level with any D-KEFS 
variables.
The CCPT score that is a measure of variability in response times, Hit Reaction Time SE, 
was significantly correlated with the most D-KEFS variables: Trailmaking Letter Sequencing 
and Letter-Number Switching, Color Word Interference Color Naming, and Inhibition.  CCPT 
Hit Reaction Time was not significantly correlated at the .01 level with any D-KEFS variables.  
D-KEFS variables that were unrelated to any CCPT variables at the .05 level or better 
include Trailmaking Visual Scan and Motor Speed, Verbal Fluency ? Letter and Category 
Fluency total correct, and Color Word Interference Word Reading.  These results were somewhat 
surprising, given that the CCPT is a visual attention (visual scan) task that requires speed (motor 
speed) in responding to letters. However, due to several of the D-KEFS inhibition/switching 
variables being significantly correlated with multiple CCPT variables, it appears as though both 
measures may be tapping into a more complex construct than simple motor speed, visual 
scanning, and letter/word identification. 
52
Table 10. Correlations between CCPT and D-KEFS Variables    
* p<.05   **p<.01   ***p<.001
Bivariate correlations between D-KEFS and CCPT dependent variables and the CPRS 
Inattentive Symptoms, Hyperactive/Impulsive Symptoms, and Total ADHD Symptoms scales 
were calculated (Table 11).  Virtually all D-KEFS scores were modestly to moderately correlated 
with all three CPRS scores.  The only D-KEFS variables that were not correlated with any of the 
CPRS scores were Trailmaking Visual Scan, Verbal Fluency-Letter Fluency total correct, and 
Color Word Interference Word Reading.  Trailmaking Number Sequencing was only significantly 
correlated with CPRS Inattentive and Total Symptoms scores, and not with CPRS Hyperactive-
Impulsive, suggesting that symptoms of hyperactivity-impulsivity may not affect performance on 
this task.
CCPT variables had a much lower incidence of significant correlations.  Hit Reaction 
Time and Hit Reaction Time Standard Error produced low to moderate correlations with all three 
53
CCPT Omissions Commissions Hit RT Hit RT SE Detectability Response Style
D-KEFS 
TM Visual Scan .125 -.018 -.031 -.023 .046 .066
TM Number Seq. -.155 -.285* .030 -.202 -.382** -.105
TM Letter Seq. -.256* -.173 -.219 -.394** -.180 -.188
TM L-N Switching -.194 -.257* -.157 -.323** -.301* -.267*
TM Motor Speed -.103 -.219 .010 -.185 -.204 .006
VF ? Letter Fluency  
  total correct -.064 -.209 .036 -.132 -.075 .026
VF ? Category 
  Fluency total correct -.192 .002 -.119 -.200 .006 -.196
VF ? Category 
  Switch total correct -.294* -.311* .031 -.271* -.247* -.175
VF ? Category  
  Switching Accuracy -.330** -.316** -.039 -.301* -.303* -.180
CW ?Color Name -.085 -.309* -.137 -.362** -.233 -.150
CW ? Word Read .047 -.104 -.177 -.231 .005 -.080
CW ? Inhibition -.201 -.223 -.312* -.446*** -.238 -.310*
CW ? 
  Inhibition/Switching -.288* -.143 -.133 -.280* -.107 -.202
Tower ? Total 
    Achievement -.128 -.197 -.145 -.281* -.243* -.028
CPRS scores.  This suggests that the average speed of all target responses for the entire test (Hit 
RT) and the variability of response time across the 18 time blocks (Hit RT SE) are affected by 
both symptoms of hyperactivity/impulsivity and inattention.  Detectability (d?) and Response 
Style (B) had a significant low correlation with only the CPRS Hyperactive/Impulsive score, 
indicating that accuracy, or the power to differentiate between signal and noise (d?) and response 
tendency to be overly or less concerned about mistakenly responding to non-targets (B), is 
affected by symptoms of hyperactivity/impulsivity but not inattention.  Overall, it appears that 
CCPT performance is not highly related to symptoms of ADHD, as measured by the CPRS.
Taken together, these results indicate that, when either a D-KEFS or CCPT measure is 
significantly correlated with CPRS scores, it is typically related to both symptom domains of 
ADHD as measured by the CPRS.  Only a few subscales (2 on each measure) were related to 
only hyperactivity/impulsivity or inattention, indicating that the D-KEFS and CCPT do not have 
many aspects that are differentially related to only one type of symptom cluster.
54
Table 11. Correlations between D-KEFS, CCPT, and CPRS Scores
CPRS 
Inattentive
T-score
CPRS 
Hyperactive-
Impulsive T-score
CPRS Total 
ADHD
T-score
D-KEFS 
Trailmaking Visual Scan -.127 -.019 -.069
Trailmaking Number Seq. -.351** -.237 -.300**
Trailmaking Letter Seq. -.374** -.284* -.332**
Trailmaking L-N Switching -.487*** -.351** -.434***
Trailmaking Motor Speed -.412*** -.310** -.375**
Verbal ? Letter Fluency total correct -.108 -.054 -.082
Verbal ? Category Fluency total correct -.420*** -.434*** -.443***
Verbal ? Category Switching total correct -.350** -.390** -.383**
Verbal ? Category Switching Accuracy -.381** -.456*** -.433**
Color Word ? Color Naming -.297* -.263* -.290*
Color Word ? Word Reading -.185 -.204 -.200
Color Word ? Inhibition -.443*** -.403** -.437***
Color Word ? Inhibition/Switching -.375** -.292* -.348**
Tower ? Total Achievement -.423** -.435*** -.441***
CCPT
Omissions .250* .226 .242*
Commissions .168 .187 .187
Hit Reaction Time (RT) .264* .269* .266*
Hit Reaction Time Standard Error (RT SE) .424*** .398** .415***
Detectability (d?) .227 .257* .255*
Response Style (B) .203 .267* .238
* p<.05   **p<.01   ***p<.001
Predictive Utility
D-KEFS.  To test the hypothesis that the planning/inhibition/switching variables of the D-
KEFS would be predictive of diagnostic group membership, a logistic regression was performed 
(Table 12).  Sex and IQ were entered in the first block of the analysis, and then the 
planning/inhibition/switching D-KEFS variables (Trailmaking Letter Number Switching, Verbal 
Fluency-Category Switching total switching accuracy, Color Word Interference Inhibition and 
Inhibition/Switching, and Tower total achievement) were entered in the second block using a 
forward stepwise LR method.  An alpha level of .05 was used as the basis of inclusion of 
dependent variables in the final model.  The likelihood ratio test was significant, meaning the 
overall model can be interpreted, and the Hosmer & Lemeshow goodness-of-fit test (with a 
55
nonsignificant chi-square) indicated that the data fit the model well.  Tower total achievement 
score was the only significant predictor in the model, and the model predicted 22.9% of the total 
variance (Cox and Snell R2). Full scale IQ and sex were not significant in the final model
The resulting classification matrix correctly identified 21 of 32 members of the ADHD 
group, producing a sensitivity coefficient of 65.6%; and 28 of 36 members of the control group, 
producing a specificity coefficient of 77.8%.  The overall correct classification rate was 72.1%. 
Positive and negative predictive power was 72.4% and 71.8%, respectively.  
Table 12. Logistic regression for D-KEFS planning/inhibition/switching variables
Predictor ? SE ? Wald?s ?2 df p Odds ratio
Constant 6.866 2.30 8.325 1 .004 959.33
Sex -.796 .771 1.067 1 .302 .451
Full Scale IQ -.035 .020 3.169 1 .075 .965
Tower Total Achievement -.260 .114 5.151 1 .023 .771
Test ?2 df p
Overall Model Evaluation
    Likelihood ratio test 17.665 3 .001
Goodness-of-fit test ?2 df p
    Hosmer & Lemeshow 8.298 8 .405
Cox and Snell R2 = .229. Nagelkerke R2 = .305
An exploratory logistic regression analysis was conducted using all of the D-KEFS 
variables to determine if prediction of diagnostic group improved when allowing variables of 
both the basic tasks and the planning/inhibition/switching variables to enter into the statistical 
model (Table 13).  Sex and IQ were entered in the first block of the analysis, then all D-KEFS 
variables were entered into second block using a forward stepwise LR method.  An alpha level of 
.05 was used as the basis of inclusion of dependent variables in the final model.  The likelihood 
ratio test was significant, meaning the overall model can be interpreted, and the Hosmer & 
Lemeshow goodness-of-fit test (with a nonsignificant chi-square) indicated that the data fit the 
model well.  The Tower total achievement, Trailmaking Motor Speed, and Verbal Fluency- 
Category Fluency total correct scores were the only significant predictors in the final model, and 
56
the model predicted 39.0% of the total variance (Cox and Snell R2).  Full scale IQ and sex were 
not significant in the final model.
The resulting classification matrix correctly identified 24 of 32 members of the ADHD 
group, producing a sensitivity coefficient of 75.0%; and 30 of 36 members of the control group, 
producing a specificity coefficient of 83.3%.  The overall correct classification rate was 79.4%. 
Positive and negative predictive power was 80.0% and 78.9%, respectively.  
Table 13. Logistic regression for all D-KEFS variables
Predictor ? SE ? Wald?s ?2 df p Odds ratio
Constant 13.845 3.990 12.043 1 .001 1.03E6
Sex -.530 .918 .334 1 .563 .588
Full Scale IQ -.005 .025 .040 1 .841 .995
Trailmaking Motor Speed -.509 .196 6.763 1 .009 .601
Verbal Fluency-Category 
    Fluency
-.401 .143 7.877 1 .005 .670
Tower Total Achievment -.341 .149 5.267 1 .022 .711
Test
Overall Model Evaluation ?2 df p
    Likelihood ratio test 33.586 5 .000
Goodness-of-fit test ?2 df p
    Hosmer & Lemeshow 8.159 8 .418
Cox and Snell R2 = .390. Nagelkerke R2 = .520
CCPT.  To examine the diagnostic utility of the CCPT, variables that demonstrated 
significant between-group differences (Omissions, Hit RT SE, and detectability) were entered in 
a forward stepwise logistic regression (Table 14).  An alpha level of .05 was used as the basis of 
inclusion of dependent variables in the final model.  The likelihood ratio test was significant, 
meaning the overall model can be interpreted, and the Hosmer & Lemeshow goodness-of-fit test 
(with a nonsignificant chi-square) indicated that the data fit the model well.  Hit RT SE remained 
in the final model, and the model predicted 16.7% of the total variance (Cox and Snell R2).  The 
resulting classification matrix correctly identified 19 of 31 members of the ADHD group, 
producing a sensitivity coefficient of 61.3%; and 31 of 37 members of the control group, 
producing a specificity coefficient of 83.8%.  The overall correct classification rate was 73.5%. 
57
Positive and negative predictive power was 76.0% and 72.1%, respectively.
Table 14. Logistic regression for CCPT variables
Predictor ? SE ? Wald?s ?2 df p Odds ratio
Constant -5.583 1.751 10.171 1 .001 .004
Hit RT SE .102 .033 9.723 1 .002 1.107
Test
Overall Model Evaluation ?2 df p
    Likelihood ratio test 12.428 1 .000
Goodness-of-fit test ?2 df p
     Hosmer & Lemeshow 5.771 8 .673
Cox and Snell R2 = .167. Nagelkerke R2 = .223
Combined Tasks.  To determine the predictive utility of diagnostic group by both the D-
KEFS and CCPT, variables that were significant in the logistic regressions described above were 
included in analysis (Table 15).  Specifically, D-KEFS Tower total achievement was the only 
significant predictor in the analyses of planning/inhibition/switching variables.  Results indicated 
only Hit RT SE as a significant predictor in the CCPT.  Sex and full scale IQ were entered in the 
first block of the equation.  The CCPT and D-KEFS variables were entered in a forward stepwise 
logistic regression.  The only variables that remained in the final model were CCPT Hit RT SE 
and Full Scale IQ, and the model predicted 25.4% of the total variance (Cox and Snell R2).  Sex 
was not significant in the final model.  The resulting classification matrix correctly identified 19 
of 31 members of the ADHD group, producing a sensitivity coefficient of 61.3%; and 28 of 37 
members of the control group, producing a specificity coefficient of 75.7%.  The overall correct 
classification rate was 69.1%.  Positive and negative predictive power was 67.9% and 70.0%, 
respectively.  
58
Table 15. Logistic regression for combined D-KEFS planning/inhibition/switching and CCPT   
                 variables
Predictor ? SE ? Wald?s ?2 df p Odds ratio
Constant 1.112 3.023 .135 1 .713 3.041
Sex -1.061 .789 1.807 1 .179 .346
Full Scale IQ -.043 .019 5.235 1 .022 .958
Hit RT SE .085 .034 6.230 1 .013 1.089
Test
Overall Model Evaluation ?2 df p
    Likelihood ratio test 19.965 3 .000
Goodness-of-fit test ?2 df p
     Hosmer & Lemeshow 4.655 8 .794
Cox and Snell R2 = .254. Nagelkerke R2 = .340
Given that only the CCPT Hit RT SE variable entered into the final model that combined 
D-KEFS planning/inhibition/switching variables and CCPT variables, another logistic regression 
was run using all of the D-KEFS variables that were significant in the exploratory analysis 
(Tower total achievement, Trailmaking Motor Speed, and Verbal Fluency-Category Fluency total 
correct) and CCPT Hit RT SE (Table 16).  This was to determine whether the additional 
information across more basic EF tasks would improve the diagnostic utility of combining the D-
KEFS and CCPT measures.  Sex and full scale IQ were entered in the first block of the equation. 
The CCPT and D-KEFS variables were entered in a forward stepwise logistic regression.  The 
variables that remained in the final model were Trailmaking Motor Speed, and Verbal Fluency-
Category Fluency total correct, and the model predicted 38.1% of the total variance (Cox and 
Snell R2).  Sex and full scale IQ were not significant in the final model.  The resulting 
classification matrix correctly identified 23 of 31 members of the ADHD group, producing a 
sensitivity coefficient of 74.2%; and 29 of 37 members of the control group, producing a 
specificity coefficient of 78.4%.  The overall correct classification rate was 76.5%.  Positive and 
negative predictive power was 74.2% and 78.4%, respectively. 
59
Table 16. Logistic regression for combined CCPT and exploratory D-KEFS variables
Predictor ? SE ? Wald?s ?2 df p Odds ratio
Constant 14.644 4.107 12.710 1 .000 2.29E6
Sex -1.378 1.002 1.890 1 .169 .252
Full Scale IQ -.018 .002 .651 1 .420 .982
Trailmaking Motor Speed -.584 .200 8.523 1 .004 .558
Verbal Fluency-Category 
    Fluency
-.481 .154 9.812 1 .002 .618
Test ?2 df p
Overall Model Evaluation
    Likelihood ratio test 32.630 4 .000
Goodness-of-fit test ?2 df p
    Hosmer & Lemeshow 1.894 8 .984
Cox and Snell R2 = .381. Nagelkerke R2 = .509
D-KEFS Prediction of ADHD Symptoms by Age Group
To determine whether specific D-KEFS variables were more predictive of ADHD 
diagnostic status and ADHD symptoms at different ages, two age groups were formed based on 
cognitive developmental theory of Piaget (1963), which divides children 7 to 11 years into the 
concrete operational period and those 11 years and up in the formal operational period.  In 
addition, children in the city where the study was conducted are typically entering middle school 
(6th grade) while they are still 11 years old and remain in middle school through age 14. 
Research suggests that ADHD symptoms decline over time; however, the transition to middle 
school disrupts this decline with an exacerbation of symptoms (Langberg, Epstein, Altaye, 
Molina, Arnold, et al., 2008).  Therefore, the cut point of 11 years was decided upon based on 
cognitive theory and research regarding school transitions and possible changes in ADHD 
symptomatology.  Children 8 years to 10 years 11 months were included in the younger age 
group, and children 11 years to 14 years were included in the older age group.  
Other methods, such as a median split based on the data or EF developmental models 
were considered; however, potential problems are found in all options for creating age groups. 
60
Median split does not give a theoretical rationale to age group division, as it focuses on creating 
equal group sizes for analyses.  Brocki & Bohlin (2006) appear to have utilized a median split 
method in their study, as the age groups have the same number in participants in each and age 
cutoffs were not explained as theoretically driven.  For the current study, median split was not 
chosen as an appropriate method, as it does not take into account what might be occurring 
developmentally at different ages.  EF developmental models are not yet fully elucidated, as 
there appear to be different trajectories for development of each of the different facets of EF, 
which makes it difficult to choose age cutoffs when several tasks of EF are being studied. 
Recognizing the potential pitfalls to using any age cutoff rationale, the following analyses were 
conducted for exploratory purposes.  In the future, other analyses based on different age cutoffs 
could be conducted.  
The logistic regressions for the D-KEFS described above were re-run separately for each 
age group.  Sex and full scale IQ were entered in the first step, and D-KEFS 
planning/inhibition/switching variables were entered in the second step in a forward stepwise 
method.  
For the younger age group, the only significant predictor in the final model was Full 
Scale IQ (Table 17).  When D-KEFS variables were entered in the second block in a stepwise 
manner, Tower total achievement was the only variable that entered the step, with the overall 
model being significant.  However, the Tower coefficient was non-significant, with p > .05. 
Therefore, the previous step that only included sex and IQ was deemed more appropriate, and the 
model predicted 41.1% of the total variance (Cox and Snell R2).  Sex was not significant in the 
final model.  The resulting classification matrix correctly identified 14 of 17 members of the 
ADHD group, producing a sensitivity coefficient of 82.4%; and 15 of 17 members of the control 
61
group, producing a specificity coefficient of 88.2%.  The overall correct classification rate was 
85.3%.  Positive and negative predictive power was 87.5% and 83.3%, respectively.  
Table 17. D-KEFS logistic regression for younger age group
Predictor ? SE ? Wald?s ?2 df p Odds ratio
Constant 10.644 3.866 7.582 1 .006 4.19E4
Full Scale IQ -.093 .032 8.417 1 .004 .911
Sex -.497 1.021 .237 1 .626 .608
Test
Overall Model Evaluation ?2 df p
     n/a (block 1)
Goodness-of-fit test ?2 df p
    Hosmer & Lemeshow 7.516 8 .482
Cox and Snell R2 = .411. Nagelkerke R2 = .548
For the older age group, no models were found to be predictive of diagnostic group.  For 
the first step, in which sex and Full Scale IQ were entered, the model was non-significant, as 
indicated by the likelihood ratio test (p > .05).  In the second step, Color Word Interference 
Inhibition entered into the model, with the likelihood ratio test being significant (p < .05), but the 
Homer and Lemeshow test revealed that the model was not a good fit for the data (p < .05). 
Therefore, none of the variables were significant predictors of diagnostic category in the older 
age group.
Using the dimensional CPRS variables, stepwise hierarchical regressions were run for 
each age group.  Sex and IQ entered in the first step, followed by all D-KEFS variables entered 
in a forward stepwise manner.  Two regressions were run for each age group, with the dependent 
variable being the CPRS Inattentive scale in the first analysis and the Hyperactive/Impulsive 
scale in the second.  
For the younger group, the significant predictors of CPRS Inattentive symptoms in the 
final model were Verbal Fluency ? Category Fluency total correct (? = -.606, t(27) = -4.54, p < 
.001), Tower total achievement (? = -.454, t(27) = -3.13, p < .01), Verbal Fluency ? Letter 
Fluency total correct (? = .300, t(27) = 2.35, p < .05), and Trailmaking Motor Speed (? = -.255, 
62
t(27) = -2.28, p < .05) .  These variables explained a significant proportion of variance in 
Inattentive scores, R2 = .71, F(6, 27) = 11.41, p < .001.  Sex and full scale IQ were non-
significant in the model.
The only significant predictors of CPRS Hyperactive/Impulsive symptoms in the final 
model for the younger age group were Verbal Fluency ? Category Fluency total correct (? = -.34, 
t(30) = -2.16, p < .05)  and full scale IQ (? = -.45, t(30) = -2.78, p < .01).  These variables also 
explained a significant proportion of variance in Hyperactivity scores, R2 = .49, F(3, 30) = 9.55, 
p < .001.  
For the older group, the only significant predictor of CPRS Inattentive symptoms in the 
final model was Color Word Interference Inhibition (? = -.58, t(30) = -2.98, p < .01).  Sex and 
full scale IQ were not significant in the final model.  Color Word Interference Inhibition also 
explained a significant proportion of variance in Inattentive scores, R2 = .29, F(3, 30) = 4.06, p < 
.05.  
The final model for CPRS Hyperactive/Impulsive symptoms was non-significant, F(3,30) 
= 2.64, p > .05.  Therefore no D-KEFS scores, nor sex or full scale IQ were significant predictors 
of Hyperactivity symptoms in the older group.
63
64 
 
Discussion 
 Prominent current theories of ADHD implicate EF deficits in individuals with ADHD 
compared to their normally developing peers.  Extensive research has found significant 
performance deficits by ADHD participants on classic measures of EF (Frazier et al., 2004; 
Willcutt et al., 2005).  The D-KEFS is a relatively new neuropsychological measure of EF and 
has been minimally studied with an ADHD population, with only some diagnostic group 
differences noted.  Furthermore, research examining D-KEFS performance in other psychiatric 
populations is limited, which suggests that relatively little is known about its clinical utility.  In 
addition, deficient performance by ADHD groups on CPTs has been noted (Conners et al., 2003; 
Frazier et al., 2004; Willcutt et al., 2005).  The primary aim of this study was to determine 
whether performance on the D-KEFS and CCPT-II differentiated ADHD children from non-
ADHD children.  In addition, concurrent and incremental validity between the measures was 
examined. 
D-KEFS Between-Group Performance 
Several analytic strategies revealed ADHD diagnosis being associated with below 
average performance on several D-KEFS tasks that measure inhibition/switching and planning 
(Trailmaking Letter-Number Switching and Tower total achievement), as well as more deficient 
performance than controls on most of the D-KEFS measures (including all EF measures as well 
as some ingredient skill subtests).  These results are consistent with previous literature, in that 
individuals with ADHD have significant difficulty (below average performance) with traditional 
trailmaking, verbal fluency, and tower tasks (Frazier et al., 2004; Willcutt et al., 2005).  These 
65 
 
tasks all include components of reconstitution and planning, inhibition of more automatic 
responding, working memory, and fine motor control and speed.  Therefore, these results fit with 
aspects of both Barkley?s (2006) and Nigg?s (2006) theories of ADHD and EF, as they both 
articulate inhibition and planning/reconstitution (part of Nigg?s working memory component) as 
deficits in individuals with ADHD. 
Using the D-KEFS, Wodka et al. (2008a) found significant differences between ADHD 
and control performance on all tasks of the Color-Word Interference subtest and on Tower Total 
Achievement scores.  The only other study using the D-KEFS found no significant differences 
between ADHD and control participants on the two D-KEFS subtests examined (Color-Word 
Interference and Verbal Fluency subtests), and ADHD participants? scores on these subtests were 
all within average range (Corbett et al., 2009).  Therefore, previous findings of ADHD 
performance on the D-KEFS are mixed, some indicating deficits in individuals with ADHD on 
planning/reconstitution and inhibition tasks, while others not finding these differences. 
It should be noted that the ADHD group means on the D-KEFS in Wodka et al.?s (2008a) 
study were all within the average range, even though there was a significant difference compared 
to controls.  Discrepancy between their results and those of the current study might be due to the 
effects of a more heterogeneous sample and larger range of IQ scores in our study, as our 
standard deviations for IQ scores were larger than those of Wodka et al.?s (SDs all 14 and 
below).  The authors indicate that measures of EF tend to be less sensitive in children with above 
average IQ (Mahone, Hagelthorn, Cutting, Schuerholz, Pelletier, Rawlins, et al., 2002).  
Therefore, our sample characteristics likely allowed for the more distinct between-group 
performance on the D-KEFS. 
D-KEFS subtest performance was found to be associated with both inattention and 
66 
 
hyperactivity/impulsivity, as rated by parents on the CPRS, for all executive measures and many 
ingredient skills.  The only variables that were not significant in these analyses are associated 
with more basic tasks of the subtests, rather than the executive components.  Therefore, it 
appears that visual scanning skills, phonological fluency, and reading skills likely do not differ 
between diagnostic groups in the current study.  Verbal Fluency Letter Fluency has been found to 
be deficient in children with ADHD by some researchers (e.g. Grodzinsky & Diamond, 1992) 
but not others (e.g. Carte, Nigg, & Hinshaw, 1996).  The weighted mean effect size from a meta-
analysis examining reading difficulties in children with ADHD was found to be 0.64, suggesting 
moderate reading deficits associated with the diagnosis of ADHD (Frazier et al., 2004).  Though 
previous research suggests a relationship between phonological fluency and reading and ADHD 
diagnosis, our results did not support this.  Current sample characteristics (e.g. IQ, 
demographics) may be influencing these findings. 
Several other ingredient skills (number sequencing, letter sequencing, motor speed, and 
color naming) produced significant correlations with diagnosis and 
inattention/hyperactivity/impulsivity, though they were all lower in magnitude compared to the 
executive tasks.  It is possible that these skills are more critical to overall performance in the 
inhibition/switching conditions of the D-KEFS subtests, thus accounting for the relationship with 
diagnostic group.   
Some more specific patterns in the relationship between D-KEFS and CPRS scores were 
noted.  Scores on the Inattentive Symptoms scale produced roughly 0.1 higher magnitude of 
correlations with the Trailmaking and Color Word Interference tasks than did the 
Hyperactive/Impulsive Symptoms scale.  Wodka et al. (2008a) found that children with ADHD-
PI performed significantly worse than ADHD-C on Trailmaking Number Sequencing and Letter 
67 
 
Sequencing.  Better performance has also been noted on traditional trailmaking tasks for ADHD-
C compared to the ADHD-PI group (Chhabildas, Pennington, & Wilcutt, 2001; Nigg, Blaskey, 
Huang-Pollock, & Rappley, 2002).  However, symptom ratings of inattention and 
hyperactivity/impulsivity predicted total time on Trailmaking equally as well (Chhabildas et al., 
2001).  Though our sample did not categorize participants by ADHD subtype, results are 
comparable that symptoms of inattention are more related to Trailmaking tasks than are 
symptoms of hyperactivity/impulsivity.  On the Stroop Color-Word Interference test, 
performance between ADHD-C and ADHD-PI subtypes did not differ significantly from each 
other but were both lower than that of controls (Nigg et al., 2002).  Our study found symptoms of 
inattention to have a higher correlation with Color Word Interference performance. 
Both the Trailmaking and Color-Word Interference subtests require persistent attention on 
a timed test, while following along visually on a sheet of paper (reading and/or writing).  Any 
inattention or stoppage in the task would require significant time to reorient to the task.  
Therefore, it follows that inattention might be more detrimental to task performance than 
possible fidgeting, impulsivity, or general hyperactivity.  However, both symptom domains were 
significantly associated with these tasks, so it is possible that inattention might be more critical to 
performance, as individuals with ADHD-C also exhibit symptoms of inattention.  Parsing out the 
inattentive component by comparing individuals with ADHD-HI type to those with ADHD-PI 
would help clarify this relationship. 
CCPT Between-Group Performance 
 Performance on the CCPT revealed diagnostic group differences on Omissions, Hit RT 
SE, and Detectability (d?).  Since mean scores on these measures were within the average range 
for both diagnostic groups, further analysis indicated only Hit RT and Hit RT SE scores in the 
68 
 
clinical range for ADHD participants.  Conners et al.?s (2003) research on the CCPT revealed 
significantly worse performance by the ADHD group on Hit RT SE, errors of commission, errors 
of omission, and Detectability (d?).  Meta-analysis of CPTs indicated moderate but 
heterogeneous effect sizes across CPT variables, with CPT hits (number of hits recorded) 
producing the largest effect sizes of all EF measures examined (Frazier et al., 2004).  This points 
to results being somewhat inconsistent across studies; however, our results are commensurate 
with some of them. 
Current study results indicate that response time and variability of response times are 
more deficient in our sample of ADHD participants compared to their control counterparts.  
Sergeant?s (2005) theory suggests that activation and arousal are problematic for children with 
ADHD, and thus gives rise to more variable reaction times (both faster and slower).  Conners et 
al.?s (2003) research on the CCPT revealed significantly worse performance by the ADHD group 
on Hit RT SE.  Wodka et al. (2007) found ADHD children with greater reaction time variability 
than controls on a basic go/no-go task.  Reaction time variability has been found to differentiate 
between children with ADHD and controls on the go/no-go continuous performance test (Rubia, 
Russell, Overmeyer, Brammer, Bullmore, et al., 2001; Van der Meere, Marzocchi, & De Meo, 
2005; Wahlstedt, Thorell, & Bohlin, 2009; Wodka et al, 2007).    
In addition, ability to detect targets versus non-targets (d?) and not responding to target 
stimuli (omissions) distinguished between ADHD and control groups in our study, though at a 
lower magnitude than the reaction time variables.  Conners et al. (2003) also found group 
differences research on CCPT omissions and Detectability (d?).  Meta-analyses revealed mean 
effect sizes in the moderate range (.64 to .66) for CPT omissions (Frazier et al., 2004; Wilcutt et 
al., 2005).  Thus, our results are consistent with prior literature that found individuals with 
69 
 
ADHD to have difficulty with accuracy on CPT tasks.  Specifically, detectability and omissions 
are conceptually related, as they both assess the ability of the individual to discriminate targets 
from non-targets and then respond appropriately. 
 Comparison of CCPT variables and CPRS Inattentive and Hyperactive/Impulsive scales 
found significant correlations with the reaction time variables.  Specifically, Hit RT SE was 
moderately correlated with both CPRS Inattentive and Hyperactive/Impulsive subscales, while 
Hit RT was modestly correlated with both subscales.  This is consistent with most theories of 
ADHD, that control of motor behavior (and consequently time taken to perform motor behavior) 
is associated with the disorder (e.g. Barkley, 1997; Nigg, 2001; Sergeant, 2005; Sonuga-Barke, 
2002).   
In regards to CCPT accuracy variables, omissions were modestly correlated with CPRS 
Inattention, while commission errors showed no significant correlations.  A previous study found 
no significant correlations between CCPT omissions and neither parent nor teacher ratings of 
inattentive and hyperactive-impulsive behaviors in children with ADHD, and a negative 
relationship between commissions and teacher reported hyperactive-impulsive behaviors was 
noted (Edwards, Gardner, Chelonis, Schulz, Flake, & Diaz, 2007).  Another study found 
commission errors related to both ADHD symptom domains (Brocki, Tillman, & Bohlin, 2010).  
Prior research has suggested that omission errors are suggestive of inattention because responses 
to target stimuli are not being generated (Barkley, 1991).  Since stimuli are being presented 
relatively frequently, omissions mean that the individual is likely off-task, bored, or not engaged 
in the task. Therefore, our results modestly support this hypothesis that inattention is related to 
omissions.   
Detectability (d?) and response style (B) were modestly correlated with CPRS 
70 
 
Hyperactivity/Impulsivity.  Significant weaknesses in CCPT detectability have been found for 
both the ADHD-C and ADHD-PI subtypes (Huang-Pollock, Nigg, & Halperin, 2006; Losier, 
McGrath, & Klein, 1996).  Huang-Pollock et al. (2006) also found children with ADHD-C to 
have more deficient response style (B) scores than those with ADHD-PI, suggesting that the 
hyperactive/impulsive component of ADHD-C is associated with difficulty in ?activation? on the 
CCPT, which is consistent with our results.  Detectability measures perceptual sensitivity to 
targets, and response style measures if a participant responds too little or too much based on the 
actual distribution of target stimuli.  More risky or impulsive (?too much?) responding would 
result in clinically significant scores.  Sergeant et al. (1999) posit that deficits in ADHD are 
associated with activation and alertness/arousal, both of which influence detectability and 
response style scores. 
Epstein et al. (2003) compared CCPT performance with symptom report on a parent 
diagnostic interview, and found Hit RT SE, Commissions, Omissions, Detectability, and 
Response Style demonstrating relationships across both ADHD symptom domains.  Other 
research has found no significant correlations between Conners? Parent and Teacher rating scale 
ADHD symptoms and any of the CCPT detectability, response style, hits, omission, and 
commission variables (Naglieri, Goldstein, Delauder, & Schwebach, 2005).  Therefore, our 
results differ somewhat in that we did not find consistent relationships across CCPT variables 
and both symptom domains.  A difference could lie in the way ADHD symptoms were measured, 
as Epstein?s study utilized a face-to-face parent interview, and while our study and Naglieri et al. 
used a rating scale.   
Predictive Utility of Diagnostic Group 
D-KEFS. Prediction of diagnostic group membership by the D-KEFS was evaluated with 
71 
 
D-KEFS planning/inhibition/switching variables, and only Tower Total Achievement score 
emerged as a significant predictor of diagnostic group (overall correct classification rate 72.1%).  
An exploratory analysis was conducted with all D-KEFS variables entered, and Tower Total 
Achievement, Trailmaking Motor Speed, and Verbal Fluency- Category Fluency Total Correct 
were the only significant predictors in the model (overall correct classification rate 79.4%).  
Thus, including more basic tasks in the equation allowed for more accurate classification than 
provided by the Tower Total Achievement score alone.  
Prior research found Tower Total Achievement and Trailmaking (Switching vs. Number 
Sequencing contrast score) to predict diagnostic status in girls, with overall correct classification 
being 67.3% (Wodka et al., 2007).  For boys, the Color Word Interference combined Color 
Naming+Word Reading optional score was the only significant predictor, with overall correct 
classification being 58.5%.  Therefore, the current study?s discriminative utility was slightly 
better than that of Wodka and colleagues, and also included Tower Total Achievement as a 
significant predictor.  However, Wodka et al. only found Tower to be a significant predictor for 
females.  As our sample included very few females, and sex was not significant in any analysis, it 
appears that the Tower test may, in fact, differentiate males with ADHD as well.  Weighted mean 
effect sizes found traditional tower tests to have moderate effects between diagnostic group (.51 
to .69) (Wilcutt et al., 2005).  This is compared to mean effect sizes of trailmaking, Stroop-like 
color-word, and verbal fluency tests averaging .40 to .56.  Thus, there is evidence from prior 
research that tower tests are slightly better at discriminating between groups than other EF tests. 
Two additional variables were significant in the current study ? Trailmaking Motor Speed 
and Verbal Fluency ? Category Fluency.  Though Wodka et al. found a Trailmaking variable to 
be predictive of diagnostic group in females, it was not the motor speed condition.   Traditional 
72 
 
trailmaking tests have been found to have a total predictive power of 54% (Grodzinsky et al., 
1999).  As our final model combined the Trailmaking variable with others, it is not surprising 
that the discriminative utility was improved.  It is possible that motor speed might be what is 
deficient in individuals with ADHD, rather than the more complex tasks of sequencing and 
switching.  Sergeant?s (2005) cognitive-energetic model posits that failure of activation of 
response is a defining factor of ADHD.  Greater sluggishness and variability in motor 
preparation, as well as difficulty in execution of complex motor movements have been found to 
be deficient in children with ADHD (McMahon & Greenberg, 1977; Oosterlaan & Sergeant, 
1995).  Our results point to this as more critical to predicting diagnostic status than the more 
complex aspects of the Trailmaking subtest.   
 Traditional category fluency tests have a mean effect size of .41 (Frazier et al., 2004), 
with mixed results indicating that individuals with ADHD have difficulty with both the letter 
fluency and category fluency tests (e.g. Grodzinsky & Barkley, 1999; Grodzinksy & Diamond, 
1992), and others indicating only difficulty with letter fluency (Sergeant, Geurtz, & Oosterlaan, 
2002).  Our study found only Category Fluency to be a significant predictor of diagnostic group, 
and the more complex task of Category Switching did not predict diagnostic group status.  It is 
unclear why only Category Fluency and neither Category Switching nor Letter Fluency were 
predictive of diagnostic group.  Given the results that showed ADHD group performance was not 
below average on either Category Fluency or Letter Fluency, it is possible that the control group 
just performed particularly well, resulting in the significant prediction of diagnostic group status.   
 CCPT.  Analyses revealed Hit RT SE being the only predictor of diagnostic status (overall 
correct classification 73.5%).  Meta-analyses of CPTs have reported low sensitivity (15-38%), 
good specificity (91%), moderate positive predictive power (68-87%), and low negative 
73 
 
predictive power (47-61%) (Grodzinsky et al., 1999; Doyle et al., 2000).  Overall correct 
classification ranged 50-67%.  These numbers were based on measures of number correct, 
omissions, and commissions.   
As the current study?s discriminative utility was based on only one significant predictor, 
Hit Reaction Time SE (variability), direct comparison with previous research is limited.  
However, some trends can be noted.  Sensitivity was significantly higher than that reported in 
previous studies, suggesting that ADHD children are more likely to have abnormal scores on the 
test than controls.  In addition, positive predictive power (76%), which shows moderate 
likelihood that an abnormal test result predicts ADHD status, was similar to that in the meta-
analyses.  Specificity was slightly lower than previous research, indicating that control 
participants were less likely to have normal scores in our study than in previous research.  
However, this rate was still 83% in the current study, which was the highest out of all predictive 
rates.  Finally, negative predictive power (72.1%) in the current study was quite a bit higher than 
previous research, indicating that individuals with normal test scores were more likely to be 
correctly classified as non-ADHD.   
It is possible that Hit RT SE is a better predictor of ADHD group status than the CPT 
variables used in the meta-analyses, as research on motor responding has found both greater 
sluggishness and greater variability in motor preparation in individuals with ADHD (Oosterlaan 
& Sergeant, 1995). Our overall results point to the CPT as being better able to classify non-
ADHD individuals than those with ADHD, which is consistent with the literature. 
D-KEFS and CCPT.  Using the planning/inhibition/switching D-KEFS variable (Tower 
Total Achievement) with the CCPT variable (Hit RT SE), the final model only retained CCPT Hit 
RT SE and Full Scale IQ as significant predictors.  Therefore, incremental utility of diagnostic 
74 
 
prediction by the tasks was not found.  Sensitivity of this model was 61.3%, specificity was 
75.7%, and the overall correct classification rate was 69.1%.  Positive and negative predictive 
power was 67.9% and 70.0%, respectively.  All of these values were lower than the model 
described above that used only CCPT variables as predictors (and not IQ).  However, the 
combined model included IQ as a predictor given the relation to D-KEFS scores.  Though IQ 
was a significant predictor in the combined analysis, it actually reduced the predictive utility of 
the CCPT.  These results also suggest that Hit Reaction Time SE on the CCPT was more 
important to predicting diagnostic group than the Tower Total Achievement score, pointing to a 
possible more pronounced deficit in motor speed variability (?erraticness?) than planning in 
individuals with ADHD. 
 When entering the CCPT Hit RT SE and D-KEFS variables that were significant in the 
exploratory analysis that included all D-KEFS variables, the final model of combined predictive 
utility only included Trailmaking Motor Speed and Verbal Fluency ? Category Fluency.  Again, 
the hypothesis of combined predictive utility of the tasks was not supported, as only D-KEFS 
variables entered into the final model, while CCPT variables did not.  Interestingly, no D-KEFS 
planning/inhibition/switching variables were significant predictors when the model attempted to 
include the CCPT Hit RT SE score.  The addition of CCPT Hit RT SE as a possible predictor 
?forced? the Tower variable out of the model.  Predictive utility of the D-KEFS-only model was 
better than that of the current model that attempted to combine CCPT and D-KEFS, suggesting 
that combining tasks is actually detrimental to prediction of diagnostic group.   
Concurrent Validity Between the D-KEFS and CCPT 
 Patterns of intercorrelations between the D-KEFS and CCPT were examined with the 
hypothesis that the D-KEFS inhibition/switching variables would be related to CCPT scores, as 
75 
 
the CCPT measures relatively basic inhibition skills.  Interpretation was based on those at the p < 
.01 level or better.  CCPT accuracy (omission and commission errors) were both significantly 
correlated with D-KEFS Verbal Fluency ? Category Switching total switching accuracy, which 
assesses  the ability to keep two concepts in mind while responding to both.  Though the CCPT 
requires motor responses and category switching requires verbal responses, both tasks include an 
aspect of keeping two rules in mind at the same time, which taps into the EFs of working 
memory and reconstitution.   
CCPT Detectability (discrimination between X and non-X) was significantly related to 
Trailmaking Number Sequencing, which is somewhat surprising, given that the CCPT requires 
the participant to discriminate between letters, not numbers.  It is possible that this result is due 
to the Number Sequencing being the first task presented on the D-KEFS Trailmaking test that 
requires the individual to view an array of letters and numbers and only pay attention to one 
modality (numbers), and practice effects might be in place when completing Letter Sequencing.  
However, both ADHD and control group means for Letter Sequencing were actually lower than 
for Number Sequencing.  Therefore, some other variable might be affecting the non-significant 
relationship between CCPT detectability and Trailmaking Leter Sequencing. 
Variability in reaction time on the CCPT was related to two D-KEFS inhibition/switching 
measures (Letter-Number Switching and with Color Word Interference Inhibition, which 
partially supports our hypothesis.  In addition, D-KEFS Letter Sequencing and Color Word 
Interference Color Naming, two more basic tasks, were also significantly correlated with CCPT 
Hit RT SE.  Though it would follow more naturally for the Trailmaking subtest to be related to an 
accuracy measure of the CCPT (because of discrimination between letters), it was related to a 
measure of reaction time variability.  Both Trailmaking and Color Word Interference subtests are 
76 
 
time-sensitive, with better scores being associated with faster response times.  Thus, more 
variable reaction time in responding would predict difficulty with any task that requires speed of 
responding combined with accuracy. 
No previous research has compared performance between these two measures; therefore, 
this study provides exploratory findings.  Overall, a clear pattern of significance between 
inhibition/switching D-KEFS variables and CCPT variables was not noted.  The tasks of each 
measure are inherently very different, with the CCPT likely requiring fewer complex cognitive 
and motor processes than some of the tasks of the D-KEFS.  However, since several of the 
variables were related to each other, it is also possible that the CCPT is actually tapping into 
more complex EFs.  Some researchers found the CCPT unable to distinguish between children 
with ADHD and those with reading disorders (McGee, Clark, & Symons, 2000).  They suggest 
that the CCPT?s use of letters as stimuli confounds phonological skills with inhibition, and this 
may confound the measurement of more basic continuous responding and behavioral inhibition. 
Given the lack of significant correlations across all measures of both tasks, it may be that the 
CCPT and the D-KEFS offer potentially useful clinical information on different aspects of EF. 
Predictive Utility of the D-KEFS by Age Group 
 Two age groups (8-10.11 years and 11.0 to 14 years) were formed, based on cognitive 
and developmental theory, to assess whether ADHD diagnosis or symptoms were predicted by 
different D-KEFS tasks at different ages.  For the younger age group, Full Scale IQ was the only 
significant predictor of diagnostic status.  The older group produced no significant models as 
predictors of diagnostic group.  Possible non-significant results could be due to statistical power 
issues, since the total sample was divided into two groups.  Full scale IQ was not significantly 
different between age group.  However, diagnosis by age group analyses produced a significant 
77 
 
interaction, with the younger control group participants having significantly higher IQ (M=120) 
than the older control group participants (M=105) and both older (M=101) and younger (M=97) 
ADHD participants.  This interaction might have affected the predictive utility of the D-KEFS, 
given the strong relationship between D-KEFS performance and IQ in our sample.    
 CPRS Inattentive symptoms in the younger age group were predicted by Verbal Fluency 
? Category Fluency and Letter Fluency total correct, Tower Total Achievement, and Trailmaking 
Motor Speed, with 71% of variance explained by these variables.  Hyperactive/Impulsive 
symptoms for the younger group were predicted by Verbal Fluency ? Category Fluency and Full 
Scale IQ, with 49% of variance explained by these variables.  For the older age group, 
Inattentive symptoms were predicted only by Color Word Interference Inhibition, accounting for 
29% of the variance.  No variables were found to be significant predictors of 
Hyperactive/Impulsive scores for this group.   
Brocki and Bohlin (2006) found that younger children?s (ages 6.0 to 9.7) ADHD 
Inattentive and Hyperactive/Impulsive symptoms were related to their performance on tasks of 
disinhibition, while the older age group?s (9.8 to 13.0 years) ADHD Inattentive symptoms were 
significantly related to verbal working memory and fluency and speed/arousal.  No relationship 
with Hyperactive/Impulsive symptoms in the older age group was noted.  Though age range in 
the groups was different than in the current study (due to Brocki and Bohlin?s median split 
method of creating groups), current results are not commensurate with their findings for the 
younger age group, as none of the tasks that were significant predictors of inattention or 
hyperactivity/impulsivity in our study were direct measures of inhibition.  The older group in the 
current study did replicate a portion of Brocki and Bohlin?s findings, as Color Word Interference 
Inhibition was predictive of inattentive symptoms.  Additionally, no measures predicted 
78 
 
hyperactive/impulsive symptoms in the older age group in either study. 
Regarding the developmental progression of EFs, research is not consistent across age 
groups and measures, thus making it difficult to establish an accurate picture of the relationship 
between ADHD symptoms and EF at different ages.  EEG studies show central nervous system 
changed through childhood, with growth periods appearing between birth and 2 years, another 
from 7 to 9 years, and a final spurt at 16 to 19 years (Thatcher, 1991).  Hudspeth and Pribram 
(1990) found frontal regions of the brain to have accelerated development from 7 to 10 years.  
Given these biologically-based changes, it would be expected that neuropsychologically 
measured cognitive changes would be evident around these periods of time.  Several researchers 
have examined the development of EF in normally developing children and have found a steady 
stage-like progression of skills from 6 to 12 years of age (Chelune & Baer, 1986; Levin, 
Culhane, Hartmann, Evankovich, Mattson, Harward, et al., 1991; Passler, Isaac, & Hynd, 1985).   
Results from the current study found Verbal Fluency, Trailmaking, and Tower Test 
predictive of ADHD symptoms in children under age 11.  Both Tower Test and Verbal Fluency 
task performance have been found to continue development past age 12 (Welsh, Pennington, & 
Groisser, 1991), which may indicate that skills are much more variable at younger ages.  Thus, 
these tests may be more predictive of ADHD symptoms in the younger age group because they 
are still in flux, and children with ADHD may be further behind in development than their same-
age peers.  In regards to Trailmaking, prior research found possible stabilization around age 12 
years (Anderson, 1998). Since the current study?s age cut point was 11 years, it is difficult to 
assess the consistency of these results.  However, Trailmaking skills are reportedly continuing to 
develop past age 11.  Our youngest age group?s ADHD symptoms were predicted by a basic 
motor speed aspect of the task, which suggests the possibility that, though task skills in general 
79 
 
are continuing to develop, children with ADHD might show difficulty with Trailmaking due to a 
motor speed deficit at younger ages. 
Limitations 
Several limitations should be noted regarding the current study.  First, inclusion criteria 
for the study did not rule out comorbid psychological disorders; therefore, the sample may have 
been more heterogeneous than those of other studies of children with ADHD, which often have 
more stringent inclusion criteria that rules out comorbidities.  Additionally, ADHD diagnosis was 
confirmed by rating scale scores and prior diagnosis by a professional.  This method was used 
given Pelham et al.?s (2005) findings that rating scales are sufficient to document ADHD 
symptomatology in research studies.  However, more strict inclusion criteria might have 
produced different outcomes.  In addition, lack of a comprehensive ADHD evaluation precluded 
the use of ADHD subtypes in analyses, as all children with ADHD were combined into one 
group.  Though the CPRS symptom scales do not necessarily classify subtypes, use of this 
measure would have resulted in most participants being labeled the Combined subtype.  
Therefore, differences in performance between subtypes could not be assessed.   
Recruitment resulted in a small number of females in the ADHD group, which is 
consistent with national prevalence rates.  Some studies have oversampled females in order to 
conduct gender comparisons (e.g. Wodka et al., 2008a); however, this method was not used for 
the current study.  Wodka and colleagues noted sex differences on a few subscales of the D-
KEFS, but our sample did not allow for specific sex comparisons.  It should be noted that sex 
was used as a predictor in all D-KEFS analyses in the current study, and it was a non-significant 
factor in all analyses. 
Sample size was relatively small for the number of variables studied, which could have 
80 
 
influenced the detectability of performance differences in the measures to be detected.  
Recruitment difficulties in a small town likely influenced sample size, as data collection spanned 
the course of almost one year.  Furthermore, many research packets were completed and returned 
for children who did not have a prior diagnosis of ADHD, but scores on the CPRS were 
clinically significant.  These potential participants did not meet criteria for either the ADHD 
group or control group.  It is possible that parents thought the study would assess their children 
for ADHD, even though it was clearly stated in the packet that this was not the case.  Regardless, 
many fewer packets were returned in which the child did not have an ADHD diagnosis and had 
non-significant CPRS scores.   
This response pattern resulted in difficulty with recruitment of control participants, 
specifically since each ADHD participant was matched to a control participant by age and sex.  
Midway through the data collection period, recruitment focused solely on control participants in 
order to obtain relatively equal numbers in each group.  Flyers continued to be posted at schools, 
daycares, doctor?s offices, and psychological clinics.  However, the snowball effect was the most 
successful method for recruiting control participants.  This could have influenced the sample 
characteristics, though, because families that were affiliated with the university (professors, 
friends of professors, etc.) were most likely to respond to a call for research participation.  
Demographics of the university faculty indicate a majority of Caucasian individuals, which 
resulted in a large number of Caucasian participants in the control group.  Racial makeup of the 
control group might have differed if all participants had been from the initial recruitment phase. 
Recruitment difficulties likely affected the confounding issues of IQ differences by race 
(African American participants had lower mean IQ than Caucasian participants).  There was a 
non-significant interaction between race and diagnostic group on IQ.  However, the 
81 
 
disproportionate number of African American participants in the ADHD group compared to the 
control group make conclusions from the results complicated in regards to race.  An attempt to 
statistically address this issue was made by including IQ as a predictor in analyses.  
Nevertheless, results would have been clearer if groups had been matched by race (perhaps 
thereby controlling for IQ in this sample) as well as sex and age. 
 Evaluation of the developmental relationship between EF and ADHD was difficult to 
complete with any certainty.  The lack of consistent literature giving age-specific cutpoints in 
development of EFs complicated the matter of selecting age groups for analyses.  The choice of 
using Piaget?s cognitive developmental theory and theories related to progression of ADHD 
symptoms upon entering middle school was made for exploratory purposes.  However, future 
analyses could be done with each D-KEFS subtest using age cut points indicated in the literature 
as reaching adult levels for each task that is similar to its D-KEFS counterpart. 
Conclusions and Future Directions 
The predictive utility of ADHD diagnostic status by the D-KEFS was found to be 
moderate and largely unrelated to the more EF task components of the test battery, with the 
exception of the Tower test.  Combined predictive utility with a more basic measure of 
behavioral inhibition, the CCPT, did not result in better diagnostic group classification rates.  In 
fact, statistical attempts at combining the tasks actually resulted in worse prediction rates.  This 
could be due to sample size and the number of variables used in the analyses, or to true lack of 
utility in combination of tasks.  Some diagnostic group performance trends were noted, with 
ADHD individuals? performance on the planning/inhibition/switching tasks of Trailmaking, 
Verbal Fluency, Color Word Interference, and Tower tests more likely to be below average than 
performance by control participants.  In addition, CPRS ratings of inattention and 
82 
 
hyperactivity/impulsivity were significantly related to almost all of the D-KEFS variables.  These 
results point to the possible clinical utility of the D-KEFS in ADHD diagnosis.  However, 
clinicians should be cautious in using either of these measures when evaluating individuals for 
ADHD, and should, rather, combine information from more clinically established sources of 
information.  Current assessment practices point to diagnostic interviews and rating scales as the 
most valid forms of assessment for ADHD (Barkley, 2006), given the unreliability of many 
neuropsychological tests to accurately predict diagnosis.  However, there is no gold standard for 
ADHD evaluation, and information from neuropsychological tests can be helpful collateral 
information in some circumstances.  Future research needs to focus on the ecological validity of 
laboratory measures, and the creation of tests that can be administered or evaluated in a ?real 
world? setting could prove better at discriminating between individuals with ADHD and those 
without the disorder.    
In addition, findings in the current study are consistent with previous research that not all 
individuals with ADHD will demonstrate deficits in EF on laboratory measures (Bernstein & 
Waber, 1990).  Researchers posit that some individuals with ADHD might recruit other 
?compensatory? cognitive resources to perform in lab settings but still have difficulties when 
subject to more distractions in the real world (Doyle et al., 200).  It remains unclear why EF 
deficits are found in some individuals with ADHD but not others, but dual-pathway models such 
as Sonuga-Barke?s (2002) might explain this.  As suggested, there might be two different 
etiologies toward ADHD, with different neuropsychological deficits as a result.  The only ways 
to clarify this relationship would be to either differentiate ADHD groups by symptom 
presentation instead of grouping them all together into one combined ADHD group, or determine 
a way to differentiate individuals by etiology based on Sonuga-Barke?s theory. 
83 
 
 Developmentally, results did not reveal any clear patterns regarding the predictive utility 
of the D-KEFS at different age groups.  However, some age group patterns were noted with the 
relationship between inattentive and hyperactive/impulsive symptoms and D-KEFS performance.  
As these analyses were exploratory, and choice of age group cut points was subject to the 
inherent lack of consensus in the literature, further research is needed to clarify age patterns.   
 Future research is necessary to evaluate the usefulness of the D-KEFS in ADHD 
diagnostic assessments, given that only two published research groups have used the measure 
with ADHD populations, with mixed results.  Critical in the process of establishing reliability 
and validity is evaluating a wide range of well-defined clinical samples.  Future studies should be 
very specific in looking at and describing their sample?s demographics and comorbid 
characteristics, as these factors could affect group performance.  In addition, since IQ scores 
were very significantly related to most D-KEFS variables, research clarifying the relationship or 
distinction between the constructs of intelligence and EF (as measured by the D-KEFS) are 
critical.  Comparison between ADHD subtypes could also elucidate possible different EF profiles 
for each subtype, as it is often the case that all ADHD subtypes are combined into one group for 
analyses.  Along similar lines, further study of possible sex differences in D-KEFS performance 
is important, given that research has pointed to differences in ADHD symptom presentation in 
girls versus boys.  Longitudinal studies with retests on the D-KEFS every few years would 
provide useful information regarding the trajectory of EF development in individuals with 
ADHD.  In addition, further clarification of age-related development of EFs would be useful to 
give researchers a comparison of how abnormal or delayed EF development is in ADHD 
children.  Previous research has pointed to the lack of reliability of the process and contrast 
scores generated by the D-KEFS scoring software.  Though it was beyond the scope of this study 
84 
 
to include analyses of these variables, further evaluation of these measures might be useful.  
Finally, the high number of intercorrelations between D-KEFS variables could have reduced the 
significance of findings in the regression analyses.  A large-scale study of D-KEFS performance 
by individuals with ADHD would allow for factor analysis, which may illuminate clusters of 
skills that may better explain possible EF difficulties. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
References
American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders 
(4th ed., rev.). Washington, DC: Author.
Anderson, V. (1998). Assessing executive functions in children: Biological, psychological, and 
developmental considerations. Neuropsychological Rehabilitation, 8, 319-349.
Anderson. V.,  Anderson, P., & Lajoie, G. (1996). The Tower of London Test: Validation and 
standardization for pediatric populations. The Clinical Neuropsychologist, 10, 54-65.
Barkley, R. A. (1991). The ecological validity of laboratory and analogue assessment methods of 
ADHD symptoms. Journal of Abnormal Child Psychology, 19, 149?178.
Barkley, R. A. (1997). Behavioral inhibition, sustained attention, and executive function: 
Constructing a unified theory of ADHD. Psychological Bulletin, 121, 65-94. 
Barkley, R. A. (2006). Attention-Deficit Hyperactivity Disorder : A handbook for diagnosis and 
treatment (3rd ed.). New York: Guilford Press.
Benton, A. L., & Hamsher, K. deS. (1976). Multilingual Aphasia Examination Manual (Rev. 
ed.). Iowa City: University of Iowa.
Bernstein, J., & Waber, D. (1990). Developmental neuropsychological assessment: The systemic 
approach. In A. Boulton, G. Baker, & M. Hiscock (Eds.), Neuromethods: 
Neuropsychology (pp. 311-371). Clifteon, NJ: Humana Press.
Bianchi, L. (1922). The mechanism of the brain and the function of the frontal lobes. Edinburgh: 
Livingston.
Bidwell, L. C., Willcutt, E. G., DeFries, J. C., & Pennington, B. F. (2007). Testing for 
85
neuropsychological endophenotypes in siblings discordant for Attention-
Deficit/Hyperactivity Disorder. Biological Psychiatry, 62, 991-998.
Blair, C., Zelazo, P. D., & Greenberg, M. T. (2005). The measurement of executive function in 
early childhood. Developmental Neuropsychology, 28, 561-571.
Borys, S. V., Spitz, H. H., & Dorans, B. A. (1982). Tower of Hanoi performance of retarded 
young adults and nonretarded children as a function of solution length and goal state. 
Journal of Experimental Child Psychology, 33, 87-110.
Brocki, K. C., & Bohlin, G. (2006). Developmental change in the relation between executive 
functions and symptoms of ADHD and co-occurring behaviour problems. Infant and 
Child Development, 15, 19-40.
Brown, T. E. (2006). Executive functions and attention deficit hyperactivity disorder: 
Implications of two conflicting views. International Journal of Disability, Development, 
and Education, 53, 35-46.
Carroll, J. B. (1993) Human cognitive abilities: A survey of factor analytic studies. New York: 
Cambridge University Press.
Carte, E. T., Nigg, J. T., & Hinshaw, S. P. (1996). Neuropsychological functioning, motor speed, 
and language processing in boys with and without ADHD.  Journal of Abnormal Child 
Psychology, 24, 481-498.
Chhabildas, N., Pennington, B. F., & Wilcutt, E. G. (2001). A comparison of neuropsychological 
profiles of the DSM-IV subtypes of ADHD. Journal of Abnormal Child Psychology, 29, 
529-540.
Chelune, G. J., & Baer, R. A. (1986). Developmental norms for the Wisconsin Card Sorting Test. 
Journal of Clinical and Experimental Neuropsychology, 8, 219-228.
86
Collett, B. R., Ohan, J. L., & Myers, K. M. (2003). Ten-year review of rating scales. VI: Scales 
assessing externalizing behaviors.  Journal of American Academy of Child and 
Adolescent Psychiatry, 42, 1143-1170.
Conners, C. K. (1995). Conners? Continuous Performance Test. North Tonawanda, NY: Multi-
Health Systems, Inc.
Conners, C. K. (2000).  Conners? Rating Scales ? Revised.  North Tonawanda, NY: Multi-Health 
Systems, Inc.
Conners, C. K., Epstein, J. N., Angold, A., & Klaric, J. (2003).  Continuous performance test 
performance in a normative epidemiological sample.  Journal of Abnormal Child 
Psychology, 31, 555-562.
Conte, R., & Regehr, S. M. (1991). Learning and transfer of inductive reasoning rules in 
overactive children. Cognitive Therapy and Research, 15, 129-139.
Corbett, B., Constantine, L., Hendren, R., Rocke, D., & Ozonoff, S. (2009). Examining executive 
functioning in children with autism spectrum disorder, attention deficit hyperactivity 
disorder, and typical development. Psychiatry Research, 166, 210-222.
Culbertson, W. C., & Zillmer, E. A. (1999). Tower of London: Examiner?s manual. North 
Towanda, NY: Multi-Health Systems.
Danforth, J. S., Barkley, R. A., & Stokes, T. F. (1991). Observations of parent-child interactions 
with hyperactive children: Research and clinical implications. Clinical Psychology 
Review, 11, 703-727.
Davidson, M. C., Amso, D., Anderson, L. C., & Diamond, A. (2006). Development of cognitive 
control and executive functions from 4 to 13 years: Evidence from manipulations of 
memory, inhibition, and task switching. Neuropsychologia, 44, 2037-2078.
87
Delis, D. C., Kaplan, E., & Kramer, J. H. (2001). Delis-Kaplan Executive Function System. San 
Antonio, Texas: The Psychological Corporation.
Delis, D. C., Lansing, A., Houston, W. S., Wetter, S., Han, S. D., et al. (2007). Creativity lost: 
The importance of testing higher-level executive functions in school-age children and 
adolescents. Journal of Psychoeducational Assessment, 25, 29-40.
Douglas, V. I. (1985). The response of ADD children to reinforcement. In L. M. Bloomingdale 
(Ed), Attention deficit disorder: diagnostic, cognitive, and therapeutic understanding (pp. 
147-162). New York: Spectrum.
Doyle, A. E., Biederman, J., Seidman, L. J., Weber, W., & Faraone, S. V. (2000). Diagnostic 
efficiency of neuropsychological test scores for discriminating boys with and without 
attention deficit-hyperactivity disorder. Journal of Consulting and Clinical Psychology, 
68, 477-488.
Drewe, E. A. (1975). Go-no go learning after frontal lobe lesions in humans. Cortex, 11, 8-16.
Dykman, R. A., Ackerman, M. A., & Oglesby, B. A. (1980). Correlates of problem solving in 
hyperactive, learning disabled, and control boys. Journal of Learning Disabilities, 13, 23-
32.
Edwards, M. C., Gardner, E. S., Chelonis, J. J., Schulz, E. G., Flake, R. A., & Diaz, P. F. (2007).  
Estimates of the validity and utility of the Conners? continuous performance test in the 
assessment of inattentive and/or hyperactive-impulsive behaviors in children. Journal of 
Abnormal Child Psychology, 35, 393-404.
Epstein, J. N., Alaatin, E., Conners, C. K., Klaric, J., Costello, J. E., et al. (2003). Relations 
between continuous performance test performance measures and ADHD behaviors. 
Journal of Abnormal Child Psychology, 31, 543-554.
88
Floyd, R. G., McCormack, A. C., Ingram, E. L., Davis, A. E., Bergeron, R., & Hamilton, G. 
(2006). Relations between the Woodcock-Johnson III clinical clusters and executive 
functions from the Delis-Kaplan Executive Function System. Journal of 
Psychoeducational Assessment, 24, 303-317. 
Frazier, T. W., Demaree, H. A., & Youngstrom, E. A. (2004). Meta-analysis of intellectual and 
neuropsychological test performance in attention-deficit/hyperactivity disorder. 
Neuropsychology, 18, 543-555.
Gogtay, N., Giedd, J. N., Lusk, L., Hayashi, K. M., Greenstein, D., et al. (2004). Dynamic 
mapping of human cortical development during childhood through early adulthood. 
Proceedings of the National Academy of Sciences of the United States of America, 101, 
8174-8179.
Golden, C. J. (1978). The Stroop color and word test. Chicago: Stoelting Company.
Grant, D. A., & Berg, E. A. (1948). A behavioral analysis of degree of reinforcement and ease of 
shifting to new response in a Weigl type card sorting problem. Journal of Experimental 
Psychology, 38, 404-411.
Grodzinsky, G. M., & Barkley, R. A. (1999). Predictive power of frontal lobe tests in the 
diagnosis of attention deficit hyperactivity disorder. The Clinical Neuropsychologist, 13, 
12-21.
Grodzinsky, G. M., & Diamond, R. (1992). Frontal lobe functioning in boys with attention-
deficit 
hyperactivity disorder. Developmental Neuropsychology, 8, 427-445.
Haenlein, M., & Caul, W. F. (1987). Attention deficit disorder with hyperactivity: A specific 
hypothesis of reward dysfunction. Journal of American Academy of Child and Adolescent 
89
Psychiatry, 26, 356-362.
Hinshaw, S. P., & Melnick, S. (1992). Self-management therapies and attention-deficit 
hyperactivity disorder: Reinforced self-evaluation and anger control interventions. 
Behavior Modification, 16, 253-273.
Hinshaw, S. P., Carte, E. T., Sami, N., Treuting, J. J., & Zupan, B. A. (2002). Preadolescent girls 
with attention-deficit/hyperactivity disorder: II, Neuropsychological performance in 
relation to subtypes and individual classification. Journal of Consulting and Clinical 
Psychology, 70, 1099-1111.
Houghton, S., Douglas, G., West, J., Whiting, K., Wall, M., et al. (1999). Differential 
patterns of executive function in children with attention-deficit hyperactivity disorder 
according to gender and subtype. Journal of Child Neurology, 14, 801-805.
Hudspeth, W., & Pribram, K. (1990). Stages of brain and cognitive maturation. Journal of 
Educational Psychology, 82, 881-884.
Kleinhans, N., Akshoomoff, N., & Delis, D. C. (2005). Executive functions in autism and 
Asperger?s disorder: Flexibility, fluency, and inhibition. Developmental 
Neuropsychology, 27, 379-401.
Klimkeit, E. I., Mattingley, J. B., Sheppard, D. M, Lee, P., & Bradshaw, J. L. (2005). Motor 
preparation, motor execution, attention, and executive functions in Attention 
Deficit/Hyperactivity Disorder (ADHD). Child Neuropsychology, 11, 153-173.
Kochanska, G., DeVet, K., & Goldman, M. (1994). Maternal reports of conscience development 
and temperament in young children. Child Development, 65, 852-868.
Lahey, B. B., Applegate, B., McBurnett, K., Biederman, J., Greenhill, L., Hynd, G.W., et al. 
(1994). DSM-IV field trials for attention deficit/hyperactivity disorder in children and 
90
adolescents.  Journal of the American Academy of Child and Adolescent Psychiatry, 151, 
1673-1685.
Levin, H. S., Culhane, K. A., Hartmann, J., Evankovich, K., Mattson, A. J., Harward, H., et al. 
(1991). Developmental chages in performance on tests of purported frontal lobe 
functioning.  Developmental Neuropsychology, 7, 377-395.
Lezak, M. D. (1985). Neuropsychological assessment. In J. A. M. Frederiks (Ed.), Handbook of 
Clinical Neurology: Vol. I. Clinical Neuropsychology (pp. 515-530). New York: Elsevier.
Logan, G. D., Schachar, R. J., & Tannock, R. (1997). Impulsivity and inhibitory control. 
Psychological Science, 8, 60-64.
Luria, A. (1966). Higher cortical functions in man. New York: Basic Books.
Mahone, E. M., Hagelthorn, K. M., Cutting, L. E., Schuerholz, L. J., Pelletier, S. F., Rawlins, C., 
et al. (2002). Effects of IQ on executive function measures in children with ADHD. 
Child Neuropsychology, 8, 52-65.
Mattson, S. N., Goodman, A. M., Caine, C., Delis, D. C., & Riley, E. P. (1999). Executive 
functioning in children with heavy prenatal alcohol exposure. Alcoholism: Clinical and 
Experimental Research, 23, 1808-1815.
McGee,R. A., Clark, S. E., & Symons, D. K. (2000). Does the Conners? continuous performance 
test aid in ADHD diagnosis? Journal of Abnormal Child Psychology, 28, 415-424.
McMahon, S. A., & Greenberg, L. M. (1977). Serial neurologic examination of hyperactive 
children. Pediatrics, 59, 584-587.
Meehl, P. E. (1970). Nuisance variables and the ex post facto design. In M. Radner & S. Winokur 
(Eds.), Minnesota studies in the philosophy of science: Analyses of theories and methods 
of physics and psychology (Vol. 4) (pp. 373-402). Minneapolis: University of Minnesota 
91
Press.
Miller, G. A. & Chapman, J. P. (2001). Misunderstanding analysis of covariance. Journal of 
Abnormal Psychology, 110, 40-48.
Naglieri, J. A., Goldstein, S., Delauder, B. Y., & Schwebach, A. (2005). Relationships between 
the WISC-III and the Cognitive Assessment System with Conners? rating scales and 
continuous performance tests. Archives of Clinical Neuropsychology, 20, 385-401.
Newcombe, F. (1969). Missile wounds of the brain. New York: Oxford University Press.
Nigg, J. T. (2001). Is ADHD a disinhibitory disorder? Psychological Bulletin, 127, 571-598.
Nigg, J.  T. (2006). What causes ADHD? Understanding what goes wrong and why. New York: 
Guilford Press.
Nigg, J. T., Blaskey, L. G., Huang-Pollock, M. A., & Rappley, M. D. (2002). Neuropsychological 
executive functions and DSM-IV ADHD subtypes. Journal of the American Academy of 
Child and Adolescent Psychiatry, 41, 59-66.
Oosterlaan, J., & Sergeant, J. A. (1995). Response choice and inhibition in ADHD, anxious, and 
aggressive children : The relationship between S-R compatibility and stop signal task. In 
J. A. Sergeant (Ed.), Eunethydis: European approaches to hyperkinetic disorder (pp. 225-
240). Amsterdam: University of Amsterdam.
Passler, M. A., Isaac, W., & Hynd, G. W. (1985). Neuropsychological development of behavior 
attributed to frontal lobe functioning in children. Developmental Neuropsychology, 1, 
349-370.
Pelham, W. E., Fabiano, G. A., & Massetti, G. A. (2005). Evidence-based assessment of 
attention deficit hyperactivity disorder in children and adolescents. Journal of Clinical 
92
Child and Adolescent Psychology, 34, 449-476.
Pennington, B. F. & Ozonoff, S. (1996). Executive functions and developmental 
psychopathology. Journal of Child Psychology and Psychiatry, 37, 51-87.
Piage, J. (1963). The origins of intelligence in children. New York: W. W. Norton.
Rasmussen, C., & Bisanz, J. (2009). Executive functioning in children with fetal alcohol 
spectrum disorders: Profiles and age-related differences. Child Neuropsychology, 15, 
201-215.
Reitan, R., & Wolfson, D. (1985). The Halstead-Reitan Neuropsychological Test Battery: Theory 
and clinical interpretation. Tucson: Neuropsychology Press.
Rowland, A. S., Lesene, C. A. & Abramowitz, A. J. (2002). The epidemiology of attention-
deficit/hyperactivity disorder (ADHD): A public health view. Mental Retardation and 
Developmental Disabilities: Research Reviews, 8, 162-170.
Sarkis, S. M., Sarkis, E. H., Marshall, D., & Archer, J. (2005). Self-regulation and inhibition in 
comorbid ADHD children: An evaluation of executive functions. Journal of Attention 
Disorders, 8, 96-108.
Sattler, J. M. (2001). Assessment of children: Cognitive applications. San Diego, CA: Jerome M. 
Sattler, Publisher, Inc.
Scheres, A., Oosterlaan, J., Geurts, H. M., Morein-Zamir, S., Meiran, N., et al. (2004). 
Executive functioning in boys with ADHD: Primarily an inhibition deficit? Archives of 
Clinical Neuropsychology, 19, 569-594.
Schonfeld, A. M., Mattson, S. N., Lang, A. R., Delis, D. C., & Riley, E. P. (2001). Verbal and 
non-verbal fluency in children with heavy prenatal alcohol exposure. Journal of Studies 
of Alcohol, 62, 239-246.
93
Schuck, S. E. B. & Crinella, F. M. (2005). Why children with ADHD do not have low IQs. 
Journal of Learning Disabilities, 38, 262-280.
Seguin, J. R., Phil, R. O., Harden, P. W., Tremblay, R. E., & Boulerice, B. (1995). Cognitive and 
neuropsychological characteristics of physically aggressive boys. Journal of Abnormal 
Psychology, 104, 614-624.
Seidman, L. J. (2006). Neuropsychological functioning in people with ADHD across the 
lifespan. Clinical Psychology Review, 26, 466-485.
Seidman, L. J., Biederman, J., & Faraone, S. V. (1995). Effects of family history and comorbidity 
on the neuropsychological performance of children with ADHD: Preliminary findings. 
Journal of the American Academy of Child & Adolescent Psychiatry, 34, 1015-1024.
Sergeant, J. A. (2005). Modeling Attention-Deficit/Hyperactivity Disorder: A critical appraisal 
of the cognitive-energetic model. Biological Psychiatry, 57, 1248-1255.
Sergeant, J. A., Geurts, H., & Oosterlaan, J. (2002). How specific is a deficit of executive 
function for attention-deficit/hyperactivity disorder? Behavioural Brain Research, 130, 
3?28.
Shallice, T. (1982). Specific impairments of planning. Philosophical Transactions of the Royal 
Society of London, Series B:  Biological Sciences (London), 298, 199-209.
Sonuga-Barke, E. J. S. (2002). Psychological heterogeneity in AD/HD ? a dual pathway model 
of behaviour and cognition.  Behavioural Brain Research, 130, 29-36.
Sonuga-Barke, E. J. S. (2005). Causal models of Attention-Deficit/Hyperactivity Disorder: From 
common simple deficits to multiple developmental pathways. Society of Biological 
Psychiatry, 57, 1231-1238.
Stroop. J. R. (1935). Studies on interference in serial verbal reactions. Journal of Experimental
94
Psychology. 6. 643-661.
Tabachnick, B. G. & Fidell, L. S. (2007). Using multivariate statistics. Boston: Pearson 
Education, Inc.
Thatcher, R. W. (1991). Maturation of the human frontal lobes. Physiological evidence for 
staging. Developmental Neuropsychology, 7, 397-419.
The Psychological Corporation. (1999). Manual for the Wechsler Abbreviated Scale of 
Intelligence. San Antonio, TX: Author.
Tupper, D. E. & Rosenblood, L. K. (1984). Methodological considerations in the use of attribute 
variables in neuropsychological research. Journal of Clinical Neuropsychology, 6, 441-
453.
Van der Meere, J., Marzocchi, G. M, & De Meo, T. (2005). Response inhibition and attention 
deficit hyperactivity disorer with and without oppositional defiant disorder screened from 
a community sample. Developmental Neuropsychology, 28, 459-472.
Van Mourik, R., Oosterlaan, J., & Sergeant, J. A. (2005). The Stroop revisited: A meta-analysis 
of interference control in AD/HD. Journal of Child Psychology and Psychiatry, 46, 150-
165.
Wahlstedt, C., Thorell, L. B., & Bohlin, G. (2009). Heterogeneity in ADHD: Neuropsychological 
pathways, comorbidity, and symptom domains. Journal of Abnormal Child Psychology, 
37, 551-564.
Waltz, J. A., Knowlton, B. J., & Holyoak, K. J. (1999). A system for relational reasoning in the 
human prefrontal cortex. Psychological Science, 10, 119-125.
Wechsler, D. (1992). Manual for the Wechsler Intelligence Scale for Children (3rd ed.). Toronto: 
Psychological Corporation.
95
Wechsler, D. (2003). Wechsler Intelligence Scale for Children, Fourth Edition technical and 
interpretive manual. San Antonio, TX: Psychological Corporation.
Welsh, M. C., Pennington, B. F., & Groisser, D. B. (1991). A normative-developmental study of 
executive function: A window on prefrontal function in children. Developmental 
Neuropsychology, 7, 131-149.
Willcutt, E. G., Doyle, A. E., Nigg, J. T., Faraone, S. V., & Pennington, B. F. (2005). Validity of 
the executive function theory of attention-deficit/hyperactivity disorder: A meta-analytic 
review. Biological Psychiatry, 57, 1336-1346.
Wodka, E. L., Loftis, C., Mostofsky, S. H., Prahme, C., Gidley Larson, J. C., et al. (2008a). 
Prediction of ADHD in boys and girls using the D-KEFS. Archives of Clinical 
Neuropsychology, 23, 283-293.
Wodka, E. L., Mahone, E. M., Blankner, J. G., Larson, J. C., Fotedar, S., Denckla, M.  B., & 
Mostofsky, S. H. (2007). Evidence that response inhibition is a primary deficit in ADHD. 
Journal of Clinical and Experimental Neuropsychology, 29, 345-356.
Wodka, E. L., Mostofsky, S. H., Prahme, C., Gidley Larson, J. C., Loftis, C., et al. (2008b). 
Process examination of executive function in ADHD: Sex and subtype effects. The 
Clinical Neuropsychologist, 22, 826-841.
Woodcock, R. W., McGrew, K. S., & Mather, N. (2001). Woodcock-Johnson III Tests of 
Cognitive Abilities. Itasca, IW: Riverside.
Zakay, D. (1992). The role of attention in children?s time reproduction. Journal of Experimental 
Child Psychology, 54, 355-371.
96
Appendix A: Tables
97
Table A1. Meta-analysis results for selected EF measures
Outcome Variable Weighted 
Mean Effect 
Size (+/- 
95% CI) 
Frazier et 
al., 2004
Number of 
studies 
included in 
analysis, 
Frazier et 
al., 2004
Weighted 
Mean Effect 
Size (+/- 95% 
CI) Willcutt 
et al., 2005
Number of 
studies that 
found sign. 
group 
difference 
(p<.05) 
Willcutt et al., 
2005
Behavioral Inhibition
Stop-Signal RT .61 (+/-.09) 22/27 (82%)
SST- Go RT .66 (+/- .20) 10
SST ? Stop RT .54 (+/-.14) 13
SST ? Probability of 
Inhibition .34 (+/-.16) 9
CPT Comm. Errors .55 (+/-.08) 40 .51 (+/- .08) 17/28 (61%)
CPT Omiss. Errors .66 (+/-.08) 33 .64 (+/-.09) 23/30 (77%)
CPT-Mean RT .39 (+/-.13) 17
CPT- Hits 1.00 (+/-.13) 19
Reconstitution/Planning
Trails A Time .40 (+/-.14) 13
Trails B Time .59 (+/-.13) 14 .55 (+/-.11) 8/14 (57%)
Tower of Hanoi .69 (+/-.26) 4/7 (57%)
Tower of London .51 (+/-.18) 3/6 (50%)
Verbal Working Memory
Letter Fluency .46 (+/-.13) 13
Category Fluency .41 (+/-.17) 9
Interference
Stroop Interference .56 (+/-.10) 20
Table A2. Predictive Validity of ADHD by selected EF tests
SE = sensitivity; SP = specificity; PPP = positive predictive power; NPP = negative 
predictive power; TPV = total predictive value
98
Measure
SE % 
(Grodzinsky 
et al./Doyle 
et al.)
SP %
(Grodzinsky 
et al./Doyle 
et al.)
PPP %
(Grodzinsky 
et al./Doyle 
et al.)
NPP %
(Grodzinsky 
et al./Doyle 
et al.)
TPV %
(Grodzinsky 
et al./Doyle 
et al.)
Stroop 
Interference 43/11 --/92 88/62 62/46 68/48
CPT
# correct 41/-- -- 87/-- 61/-- 67/--
Omissions --/15 --/91 --/68 --/47 --/50
Comissions 38/-- -- 83/-- 59/-- 65/--
Trail 
Making Test
Trails A 20/-- -- 68/-- 51/-- 54/--
Trails B 18/-- -- 71/-- 51/-- 54/--
Table A3. Internal consistency and test-retest reliability values reported in D-KEFS manual
99
Task Variable Internal 
Consistency
(Range for ages 
8-14) 
Test-Retest 
Reliability 
(Reported for 
combined ages 8-
19)
Trailmaking
Combined Number + Letter Sequencing .59 to .79 n/a
Visual Scanning n/a .50
Number Sequencing n/a .77
Letter Sequencing n/a .57
Switching n/a .20
Motor Speed n/a .82
Verbal Fluency
Letter Fluency .68 to .80 .67
Category Fluency .58 to .75 .70
Category Switching Total Correct .37 to .62 .65
Category Switching Total Switching Accuracy .53 to .76 .53
Color Word Interference
Color Naming n/a .79
Word Reading n/a .77
Combined Color Naming + Word Reading 
Score
.62 to .77 n/a
Inhibition n/a .90
Inhibition/Switching n/a .80
Tower
Tower total achievement .43 to .84 .51
Table A4. Intercorrelations of D-KEFS Variables reported in the Technical Manual
nr = not reported in manual
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.
1. TM Visual Scan   -
2. TM Number Seq. .38   -
3. TM Letter Seq. .36 .54   -
4. TM L-N Switching .24 .43 .45    -
5. TM Motor Speed .28 .33 .37 .23   -
6. VF ? Letter 
Fluency total correct  nr  nr  nr .21  nr    -
7. VF ? Category 
Fluency total correct  nr  nr  nr .19  nr .55   -
8. VF ? Category 
Switch total correct  nr  nr  nr .18  nr .40 .45   -
9. VF ? Category 
Switching Accuracy  nr  nr  nr .13  nr .29 .34 .72   -
10. CW ?Color Name .16 .26 .27  nr .17  nr  nr  nr  nr   -
11. CW ? Word Read  nr .20 .25  nr .18  nr  nr  nr  nr .57   -
12. CW ? Inhibition  nr  nr  nr .32  nr .27 .26 .19 .12 .49 .45   -
13. CW ? 
Inhibition/Switching  nr  nr  nr .26  nr  nr  nr  nr  nr .41 .42 .58    -
14. Tower ? Total 
Achievement  nr .10 .08 -.10 07 -.03 -.09 -.06 -.03  nr  nr -.08 -.07
100
Appendix B: Forms
101
Participant Demographic Questionnaire
Child Age:
__________
Child Date of Birth:
(MM/DD/YY) _____ / _____ / _____
Child Sex: (circle one)
1=Male 2=Female
Child Race: (circle one)
1= Caucasian 5=Native American
2=African-American 6=Mixed (specify) _______________
3=Hispanic 7=Other (specify) _______________
4=Asian
Has your child ever received an ADHD diagnosis from a physician or psychologist? (circle one)
1=Yes 2=No
Is your child currently diagnosed with ADHD by a physician or psychologist? (circle one)
1=Yes 2=No
102
Has your child ever taken medication for ADHD? (circle one)
1=Yes 2=No
If yes, which type of ADHD medication has your child taken in the past:
(circle all that apply)
1. Adderall
2. Adderall XR
3. Attenta
4. Concerta
5. Daytrana
6. Dexedrine
7. Dexedrine SR
8. Dexedrine Spansules
9. Dextroamphetamine Sulfate
10. Dextrostat 
11. Equasym
12. Focalin
13. Metadate
14. Metadate CD
15. Metadate ER
16. Methylin
17. Methylin ER
18. Methylphenidate (generic)
19. Mixed Amphetamine Salts
20. Penid
21. Ritalin
22. Ritalin LA
23. Ritalin SR
24. Rubifen
25. Strattera
26. Vyvanse
27. Other (specify) ______________
Does your child currently take medication for ADHD? (circle one)
1=Yes 2=No
If yes, which type of ADHD medication does your child currently takes:
(circle all that apply)
1. Adderall
2. Adderall XR
3. Attenta
4. Concerta
5. Daytrana
6. Dexedrine
7. Dexedrine SR
8. Dexedrine Spansules
9. Dextroamphetamine Sulfate
10. Dextrostat 
11. Equasym
12. Focalin
13. Metadate
14. Metadate CD
15. Metadate ER
16. Methylin
17. Methylin ER
18. Methylphenidate (generic)
19. Mixed Amphetamine Salts
20. Penid
21. Ritalin
22. Ritalin LA
23. Ritalin SR
24. Rubifen
25. Strattera
26. Vyvanse
27. Other (specify) _____________
103
Has your child ever taken medication for a psychological problem (antidepressant, antianxiety, antipsychotic)?
If unsure, just list any medication(s) that your child has taken in the past for a psychological problem. (circle one)
1=Yes 2=No
If yes, write the name of the medication(s) your child has taken in the past:
____________________ ____________________ ____________________
____________________ ____________________ ____________________
Does your child currently take medication for a psychological problem (antidepressant, antianxiety, antipsychotic)?
If unsure, just list any medication(s) that your child currently takes for a psychological problem. (circle one)
1=Yes 2=No
If yes, write the name of the medication(s) your child currently takes:
____________________ ____________________ ____________________
____________________ ____________________ ____________________
Does your child have normal vision? (circle one)
1=Yes 2=No
If your child does not have normal vision, does s/he wear glasses or contacts? (circle one)
1=Yes 2=No
Does your child have colorblind vision? (circle one)
1=Yes 2=No
Does your child have any of the following: seizures, schizophrenia, brain damage? (circle one)
1=Yes 2=No
If yes, please specify:
____________________ ____________________ ________________
104