Journal of School Psychology, Vol. 40, No. 3, pp.

259 – 283, 2002

Copyright D 2002 Society for the Study of School Psychology
Printed in the USA
0022-4405/02 $ – see front matter

PII S0022-4405(02)00100-0

Academic Underachievement and Attention-Deficit/

Hyperactivity Disorder:
The Negative Impact of Symptom Severity on
School Performance
Tammy DeShazo Barry, Robert D. Lyman,
and Laura Grofer Klinger
Department of Psychology, University of Alabama, Tuscaloosa,
AL 35487-0348, USA

Many children with Attention-Deficit/Hyperactivity Disorder (ADHD) achieve

academically at a lower level than would be predicted given their intellectual
abilities. However, the extent to which this is due to behavioral problems versus
cognitive deficits associated with the disorder is unclear. In the present study, a
group of children with ADHD (with average intellectual abilities) performed
significantly below prediction in reading, writing, and mathematics skills and
demonstrated a greater discrepancy between actual and predicted achievement
than did a group of non-ADHD children. Even when controlling for performance
on a measure of executive functioning, severity of ADHD symptoms, based on
parent report, significantly predicted academic underachievement in reading,
writing, and mathematics. These results indicate that the more severe the behavioral
symptomatology of children with ADHD is, the more negatively impacted their
school performance may be. Results are discussed in terms of diagnostic and
intervention implications. D 2002 Society for the Study of School Psychology.
Published by Elsevier Science Ltd

Keywords: Academic underachievement, ADHD, Symptom severity.

INTRODUCTION
Attention-Deficit/Hyperactivity Disorder (ADHD) is the most prevalent
developmental disorder, affecting 3 –5% of school-aged children (Ameri-
can Psychiatric Association, 1994). It is estimated that at least one child in
every education classroom has ADHD (DuPaul & Stoner, 1994). The
disorder tends to negatively affect adjustment, including behavioral, social,
and academic functioning (DuPaul & Stoner, 1994). However, the current

Received 17 October 2001; received in revised form 8 March 2002; accepted 8 March 2002.
Address correspondence and reprint requests to Tammy DeShazo Barry, Department
of Psychology, University of Alabama, Box 870348, Tuscaloosa, AL 35487-0348, USA;
E-mail: tbarry@bama.ua.edu

259
260 Journal of School Psychology

diagnostic classification of ADHD (American Psychiatric Association, 1994)

focuses on the behavioral deficits (inattention, hyperactivity, and impulsiv-
ity) of the disorder and has been criticized for failing to explain the
cognitive (i.e., executive function) deficits that are associated with ADHD
(Barkley, 1997a). The focus of the present study is to examine whether the
academic underachievement that often accompanies ADHD is related to
the behavioral or cognitive impairments associated with the disorder. A
brief review of the literature on EF impairments and academic underach-
ievement in ADHD is provided below.

Executive Functions and ADHD

Executive functions (EF) are the ‘‘cognitive abilities necessary for complex
goal-directed behavior and adaptation to a range of environmental changes
and demands’’ (Loring, 1999, p. 64). The heterogeneous list of complex
mental abilities considered to be part of the EF constellation generally
include: planning and organizing; maintaining an appropriate problem-
solving set to achieve a future goal; inhibiting an inappropriate response or
deferring a response to a more appropriate time; representing a task
mentally (i.e., in working memory); cognitive flexibility; and deduction
based on limited information (e.g., Banich, 1997; Denckla, 1994; Loring,
1999; Pennington & Ozonoff, 1996; Weintraub, 1993). Considering the role
of attentional processes in EF, it appears that children with ADHD would
experience deficits in some of the abilities constituting this construct.
The theory of EF deficits associated with ADHD has been supported by
some neurological and neuropsychological research. For example,
research comparing the neuroanatomy of children with and without
ADHD shows that children with the disorder have decreased blood flow
to the frontal lobes (see Grodzinsky & Diamond, 1992, for a review).
Magnetic resonance imaging (MRI) scans have also revealed that non-
ADHD children have slightly larger right frontal lobes than left; however,
children with ADHD tend to lack this asymmetry, which may explain their
deficits in sustained attention, a process associated with the right frontal
lobe (Hynd, Semrud-Clikeman, Lorys, Novey & Eliopulas, 1991). Further
evidence for frontal lobe dysfunction in ADHD comes from case studies of
patients with frontal lesions. Such patients often display the hallmark
behavioral symptoms of ADHD: hyperactivity, distractibility, and impulsivity.
Thus, several researchers have hypothesized that the behavioral symptoms
of ADHD are linked to EF deficits (e.g., Barkley, 1997b).
Cognitive and neuropsychological research has also suggested that
certain regions of the brain may be implicated in the disorder. A review
of 18 published studies investigating the ‘‘frontal hypothesis’’ of ADHD
through neuropsychological measures revealed that children with ADHD
demonstrated significantly worse performance than non-ADHD children
 Barry, Lyman, and Klinger 261

on 67% of a total of 60 EF measures, whereas they did not perform

significantly better than controls on any of the EF measures (Pennington
& Ozonoff, 1996). In contrast, ADHD children were impaired relative to
non-ADHD children on only 35% of the non-EF measures across the
reviewed studies, indicating some specificity to the finding of EF deficits
(as opposed to deficits associated with other regions of the brain) in
ADHD. For example, one study showed that the performance of children
with ADHD was significantly diminished compared to a control group on a
battery of frontal lobe tasks (e.g., motor control and problem-solving skills)
but that they performed as well as the control group on a battery of tasks
sensitive to temporal dysfunction (e.g., memory tasks; Shue & Douglas,
1992). It is notable, however, that research examining the performance of
children with ADHD on EF tasks has yielded inconsistent results (Grodzin-
sky & Diamond, 1992). These variable findings may be indicative of more
specific EF deficits, rather than global EF deficits.

ADHD Performance on Specific Measures of EF

There are many tasks commonly used to evaluate EF in child populations,
including children with ADHD (see Pennington & Ozonoff, 1996, for a
review). However, it is not feasible to include all possible measures in either
a clinical or research setting. Therefore, performance of children with
ADHD across three common measures will be reviewed.
Four known studies have used the Tower of Hanoi (TOH) to investigate
the EF abilities of children with ADHD (Aman, Roberts & Pennington,
1998; Ozonoff & Jensen, 1999; Pennington, Groisser & Welsh, 1993;
Weyandt & Willis, 1994). Based on a quantitative review of three of these
studies by Pennington and Ozonoff (1996), the TOH appears to be
significantly sensitive to the deficits associated with ADHD, yielding an
average effect size, mean d, of 1.08. These studies have generally found
that, whereas children with ADHD displayed deficits compared to controls
on several EF tasks including the TOH, they performed equivalently to
controls on non-EF tasks (Pennington & Ozonoff, 1996).
Unlike the TOH, the Wisconsin Card Sorting Test (WCST) has been
used more often with groups of children with ADHD and, although some
studies reveal no ADHD-related deficits on this measure (e.g., Pennington
et al., 1993; Weyandt & Willis, 1994), others do (e.g., Heaton, Chelune,
Talley, Kay & Curtiss, 1993; Shue & Douglas, 1992). The WCST has
demonstrated satisfactory developmental sensitivity for the 7– 12-year-old
age range (i.e., no ceiling effects; Denckla, 1994) within a non-clinical
population. Specifically, the perseverative error measure of the WCST may
be particularly sensitive in detecting a developmental deficit in EF. Indeed,
a meta-analysis of studies using the WCST perseveration score, to compare
performance of ADHD and non-ADHD groups, revealed an average effect
262 Journal of School Psychology

size, mean d, of 0.45 (Pennington & Ozonoff, 1996). This same meta-
analysis found that total time on Trail Making Test, Part B (TMT-B)
differentiated between ADHD and nonclinical groups in four out of the
six studies in which it had been used, with a mean d of 0.75, which is
considered to be large (Pennington & Ozonoff, 1996). These findings
suggest that both the WCST and the TMT-B are valid measures for
investigating EF in children with ADHD.
In a study using a battery of measures that included the WCST and the
TMT-B, Shue and Douglas (1992) found that the ADHD group, when
compared to normal controls, achieved fewer categories on the WCST;
made more errors (both perseverative and nonperseverative) on the
WCST; and made more errors and required more time to complete the
TMT-B. Their conclusion was that children with ADHD experience prob-
lems with organized, goal-directed behavior and have ‘‘difficulty formulat-
ing and testing hypotheses and using feedback to guide responding’’ (Shue
& Douglas, 1992, p. 114).
Although the findings of Shue and Douglas (1992) are convincing, other
studies have found opposing results. For example, Grodzinsky and Dia-
mond (1992) found no impairments associated with ADHD on the TMT-B
and no significant differences on most measures of the WCST. The mixed
pattern of empirical findings suggest that some cognitive deficits are
associated with ADHD, including possible deficits in EF, but no explicit
and definitive pattern of EF deficits has emerged from the research. That is,
some studies have presented with impressive evidence for an EF deficit
theory of ADHD, whereas other studies, using comparable measures, have
not clearly supported such a theory.
In summary, the literature investigating potential EF deficits in children
with ADHD has been somewhat unclear, which may indicate that some
components of EF are impaired in children with ADHD, whereas others are
intact. Furthermore, although children with ADHD may show impairments
on cognitive measures relative to non-ADHD children, the DSM-IV diag-
nostic criteria specifically include only behavioral criteria (American Psy-
chiatric Association, 1994). To better understand potential underlying core
deficits associated with ADHD, it is important to corroborate the behavioral
evidence with indications of cognitive deficits.

ADHD and Academic Underachievement

The conceptualization of ADHD as a deficit in higher order cognitive
processes has important implications for both clinical and academic
interventions. For example, research has shown that children with ADHD
are traditionally academic underachievers, a problem that appears to be
specific to ADHD when compared to other disruptive behavior disorders,
such as conduct disorder (Frick et al., 1991). As many as 30% of children
 Barry, Lyman, and Klinger 263

with ADHD do not achieve academically at the level predicted by their age
or IQ (Frick & Lahey, 1991; Kamphaus & Frick, 1996). Although this
finding may be interpreted as a problem in behavior (e.g., off-task, not
completing work, etc.), which then interferes with the child’s ability to
achieve to his or her potential, it may also be that children with ADHD have
specific core cognitive impairments that hinder learning.
Some studies have suggested that deficits in problem-solving and other
cognitive processes are likely linked with the poor academic performance
associated with ADHD. For example, in addition to slower computational
performance in mathematics, which may be behavioral (i.e., slow output of
work), children with ADHD also have been shown to score lower on
measures of their problem-solving ability in conceptual math (Zentall,
Smith, Lee & Wieczorek, 1994). Clarification of these cognitive impair-
ments and their relation to academic achievement is needed to begin the
subsequent process of identifying the possible causes of underachievement.
Knowledge of the differential contribution of behavioral difficulties and
complications with cognitive processes could also help guide academic
intervention efforts.

PRESENT STUDY
The present study examined the occurrence of academic underachieve-
ment in a group of children diagnosed with ADHD, compared to a group
of non-ADHD children. Furthermore, in an effort to explain why such
academic underachievement co-occurs with ADHD, this study explored the
relative extent to which EF and behavioral criteria predict discrepancies
between intellectual ability and academic achievement. The neuropsycho-
logical measures chosen for the EF battery included those that assess higher
order EF processes, such as cognitive flexibility and planning.

METHOD
Participants
Participants included 66 children, ranging in age from 8 years 9 months to
14 years 5 months (M =11 years 2 months, SD=1 year 3 months). The ADHD
group consisted of 33 children (21 males and 12 females) who met the
Diagnostic and Statistical Manual of Mental Disorders (DSM-IV; American
Psychiatric Association, 1994) criteria for ADHD. These children were
recruited from pediatric practices, local mental health centers, a university
research database of children with ADHD, and local schools (informational
flyers about the project were sent home with children in the fourth, fifth,
and sixth grades known by school personnel to be diagnosed with ADHD,
264 Journal of School Psychology

and interested parents returned an attached form). To be included in the

ADHD group, each child had an independent diagnosis of ADHD from a
physician or mental health care professional and was currently under the
care of a pediatrician or child psychiatrist to treat their symptoms. The
presence of ADHD was further verified by several measures completed by
the child’s parent at the time of participation, and each child was required
to meet the ADHD threshold on all parent measures (discussed below).
Finally, each child was required to obtain an IQ score of 70 or greater,
based on the Composite IQ score of the Kaufman Brief Intelligence Test
(K-BIT; Kaufman & Kaufman, 1990). An additional six participants with
ADHD participated in the study but were excluded due to failure to meet
one or more of these criteria. Based on parent ratings, 30 of the final
sample of ADHD participants were classified as meeting criteria for ADHD-
Combined Type, and three were classified as meeting criteria for ADHD-
Predominantly Inattentive Type.
Thirty-three non-ADHD children (15 males and 18 females) served as
control participants. Non-ADHD participants were recruited from a
university research database of non-ADHD children who had partici-
pated in previous research experiments (parents were contacted to de-
termine if they would be interested in participating in the current
project) and local schools (informational flyers about the project were
sent home with children in the fourth, fifth, and sixth grades, and inte-
rested parents returned an attached form). To be included in the non-
ADHD group, a child had never received a diagnosis of ADHD, had no
current diagnosis of another emotional or behavioral problem, and was
within the specified non-ADHD range on all parent measures (discussed
below). Finally, each child was required to obtain a Composite IQ score
of 70 or greater. An additional 19 non-ADHD children participated in
the study but were excluded for failure to meet one or more of the in-
clusion criteria.
Based on two-way contingency table analyses and independent samples
t-tests, the ADHD and non-ADHD groups were equivalent across most
demographic characteristics, including average education of caregivers
and household income (general estimates of socioeconomic status), race,
and gender. While both groups were within an average range of intellec-
tual functioning, the non-ADHD group did obtain significantly higher IQ
scores (see Table 1).
Additional analyses found diagnostic group to be significantly related to
having a learning disability (LD) diagnosis (see Table 1). Consistent with
literature regarding the comorbidity of ADHD and LD (Frick et al., 1991),
eight of the participants with ADHD had a comorbid diagnosis of an LD. In
contrast, none of the non-ADHD children had received an LD diagnosis.
Diagnostic group was also significantly related to being placed in a special
education class (see Table 1). Specifically, a total of 14 children in the
 Barry, Lyman, and Klinger 265

Table 1
Demographic Characteristics of the ADHD and Non-ADHD Groups

ADHD (n = 33) Non-ADHD (n = 33)

Gender (% males) (63.6) (45.5)

Pearson v2 (1, N = 66) = 2.20, p = .14
Race (% Caucasian) (87.9) (84.8)
Pearson v2 (2, N = 66) = 1.02, p = .60
Comorbid Diagnosis of (24.2) (0.0)
Learning Disability (% Yes)
Pearson v2 (1, N = 66) = 9.10, p = .003
Current or Previous Placement (42.4) (6.1)
in Special Education (% Yes)
Pearson v2 (1, N = 63) = 10.61, p = .001

Mean (SD) Mean (SD) t-Value

Age in Months 132.67 (15.59) 134.88 (13.87) 0.61

Grade in School 5.30 (1.24) 5.45 (1.18) 1.43
Average Education of 13.36 (2.07) 13.99 (2.05) 1.23
Caregivers (in years)
Household Income 51.99 (30.80) 52.75 (22.68) 0.11
(in thousand dollars)
K-BIT Vocabulary 100.12 (12.48) 106.79 (8.59) 2.53*
Standard Score
K-BIT Matrices 105.21 (16.73) 114.58 (11.96) 2.62*
Standard Score
K-BIT IQ Composite 103.03 (14.39) 112.03 (9.35) 3.01**
Note. K-BIT=Kaufman Brief Intelligence Test.
*p < 0.05.
**p < 0.01.

ADHD group had been placed in special education classes at some point in
their academic career, whereas only two non-ADHD children had been in
special education. This excludes those participants whose parents reported
that they had received gifted education as special education classes (n = 3,
from the non-ADHD group), given that gifted education is not associated
with learning impairments or a special need. The nature of the special
education ranged from speech classes in early elementary (kindergarten
and first grade) to self-contained academic classes (e.g., for reading). As
would be expected, independent samples t-tests revealed that the groups
differed significantly across the behavioral measures that were used to
determine inclusion (see Table 2), confirming that the groups did signifi-
cantly differ in ADHD symptomatology.
Finally, parents of only four participants (one in the non-ADHD group
and three in the ADHD group) reported that their child had ever received
another diagnosis (other than an LD). The non-ADHD participant had a
previous diagnosis of school phobia, but this was reported to no longer be a
problem. The comorbid diagnoses of the three participants with ADHD
266 Journal of School Psychology

Table 2
Differences Between ADHD and non-ADHD Groups on ADHD Behavioral Measures

ADHD (n = 33) Non-ADHD (n = 33)

Behavioral Measures Mean (SD) Mean (SD) t-Value

DSM-IV ADHD Checklist

Number Inattention Symptoms 7.45 (1.68) 0.30 (0.81) 22.04***
Number Hyperactivity Symptoms 5.06 (3.12) 0.18 (0.53) 8.85***
Total Rating on Inattention 19.73 (4.14) 3.67 (2.79) 18.48***
Total Rating on Hyperactivity 14.73 (6.49) 1.91 (2.16) 10.77***
Inattention Percentile 95.42 (4.30) 41.45 (25.02) 12.21***
Hyperactivity Percentile 91.03 (8.54) 31.21 (29.53) 11.18***

BASC-PRS
Attention Problems T-Score 70.36 (6.59) 45.97 (7.83) 13.69***
Hyperactivity T-Score 71.42 (14.45) 40.55 (6.09) 11.31***

HSQ
HSQ Total Score 46.18 (22.92) 4.79 (6.47) 41.39***
Note. DSM-IV=Diagnostic and Statistical Manual, 4th Ed. (American Psychiatric Association,
1994); BASC – PRS =Behavior Assessment System for Children – Parent Rating Scale;
HSQ = Home Situations Questionnaire.
***p < 0.001.

included depression, Oppositional Defiant Disorder (ODD), and a type of

‘‘thought disorder.’’

Measures
Tower of Hanoi (Borys, Spitz & Dorans, 1982; Leon-Carrion, in press). A
computerized TOH, a subtest of the Sevilla Computerized Neuropsycho-
logical Test Battery (Leon-Carrion, in press), was included as a measure of
EF. The objective is to move a tower of four graduated disks, varying in size
and color, from the first of three posts to the third. Each disk can be moved
one at a time as long as the following restrictions are observed: a larger disk
may not be placed on a smaller disk; no movement can be made from an
empty post; and no disk can be moved to the same position that it currently
occupies. A restriction violation is counted as an illegal movement. Partic-
ipants move the disks by pressing the number of the origin post (1, 2, or 3)
and then pressing the number of the destination post on the numerical
keypad of a standard keyboard. They are allowed to move only the top disk
on any post. The current move is displayed on the screen for their
reference. Once the origin post has been selected, it cannot be changed.
The dependent measures collected by the computer program include total
time to solve the problem (in milliseconds), pause time per response and
an overall mean pause time, number of moves required to solve the
problem, and frequency and percentage of three types of illegal move-
 Barry, Lyman, and Klinger 267

ments (i.e., those that violate a restriction). The optimum solution for the
four-disk puzzle is 15 moves (Leon-Carrion, in press). Administration time
for the TOH varies somewhat based on number of moves but is generally
less than 10 min, including instructions and practice.

Wisconsin Card Sorting Test (Grant & Berg, 1948; Loong, 1990). The
present study utilized a computerized version of the WCST, which displays
four stimulus ‘‘cards’’ at the top of the computer screen and a response
‘‘card’’ in the bottom left corner of the screen. Each card contains figures of
varying forms (triangles, stars, crosses, or circles), colors (red, green, yellow,
or blue), and numbers of figures (ranging from one to four). The object of
the WCST is to match each response card with the correct stimulus card
based on one of three sorting categories (color, form, or number).
Participants responded by pressing the number of the stimulus ‘‘card’’ to
which they wanted to match (either 1, 2, 3, or 4) on a standard keyboard.
The participant is not told the sorting category and must discover it based
on feedback about performance, which is provided by the computer. Once
the participant obtains 10 consecutive correct responses in a category, the
category is considered completed and the sorting category shifts to the next
category without any warning to the participant. The test ends when the
participant completes a total of six categories or has used all of the 128
response cards. Administration time varies depending on the number of
cards used but generally does not exceed 15 min.
The computer program provides a number of scores to be used as
measures of performance, including: duration of the test and average
response time; total number of categories completed, total number of trials
(with a maximum of 128), and total number of trials needed to complete
the first category (a measure of initial learning); total number and percent
of correct responses; and total number and percent of errors. Errors are
further defined as either perseverative errors (continuing to match on the
same sorting category even though the feedback tells them it is incorrect),
nonperseverative errors (sorting by a category that is neither the correct
category nor a perseverative response), and ‘‘other’’ responses (sorting
based on neither color, form, nor number). Additionally, the computer
program provides information regarding the percent of conceptual level of
responses, which is the ‘‘number of all correct responses in the test that
occur in runs of three or more, divided by the total number of trials’’
(Loong, 1990, p. 4). The total number of times the participant failed to
maintain set, indicated by getting five or more consecutive correct
responses but failing to complete a category, is also computed. Finally, a
learning-to-learn score, which demonstrates how efficiently the participant
learned to complete categories as the test progressed, is calculated. Studies
examining the test – retest reliability of the WCST have yielded a coefficient
268 Journal of School Psychology

of 0.60, which is considered to be moderate to good reliability (Heaton

et al., 1993).

Trail Making Test, Part B (Reitan, 1958; Reitan & Wolfson, 1993). This
paper-and-pencil test requires participants to connect 25 randomly placed
circles (numbered from 1 to 13 and lettered from A to L) in sequence,
while alternating between numbers and letters. The task is timed, and the
participants are told to complete the task as quickly as possible without
making mistakes. When an error is made, the examiner immediately stops
the participant and returns the participant to the last correct circle, so that
the participant can proceed correctly. Administration time, including
instructions and practice, is approximately 5 min.

Kaufman Brief Intelligence Test (Kaufman & Kaufman, 1990). The K-BIT
is a brief intelligence measure that consists of two parts: Vocabulary, which
measures verbal intelligence, and Matrices, which measures nonverbal
intelligence. Administration time ranges from 15 to 30 min. Age-based
standard scores are available for both the Vocabulary and Matrices sections
and for the K-BIT IQ Composite, an overall measure of intelligence. The
K-BIT IQ Composite has a test –retest reliability estimate of 0.92 and
correlates 0.80 with the Full Scale IQ score from the WISC-R. Any
participants with an estimated level of intelligence in a severely impaired
range (i.e., less than 70) were excluded from the study.

Woodcock –McGrew – Werder Mini-Battery of Achievement (Woodcock,

McGrew & Werder, 1994). The Mini-Battery of Achievement (MBA) is a
brief, wide-range test of basic academic skills and knowledge. Test 1
(Reading) consists of three subtests which measure reading recognition
skills, vocabulary, and comprehension. Test 2 (Writing) is composed of one
subtest that requires written response to a variety of questions and a second
subtest that involves identifying and correcting mistakes in written passages.
Test 3 (Mathematics) consists of two subtests that measure calculations
skills, as well as mathematical concepts and reasoning. The Reading,
Writing, and Mathematics scores are also combined into a Basic Skills
Cluster score. Administration time for the MBA ranges from 30 to 40 min.
The academic achievement scores from the MBA were used in determining
significant discrepancies between predicted achievement based on intellec-
tual functioning as measured by the K-BIT and actual achievement.

DSM-IV ADHD Checklist, adapted from the Diagnostic and Statistical

Manual of Mental Disorders (American Psychiatric Association,
1994). The DSM-IV ADHD Checklist includes a list of the nine inattention
symptoms and the nine hyperactivity-impulsivity symptoms of ADHD.
Parents indicate how frequently they observe their child engaging in such
 Barry, Lyman, and Klinger 269

behaviors on a Likert-type scale ranging from 0 to 3. Based on the DSM-IV

criteria, a child must display a significant number of symptoms in at least
one of the two areas (inattention or hyperactivity-impulsivity) to be
diagnosed with ADHD. Coefficient alphas were 0.97 and 0.96 for the
Inattention and Hyperactivity-Impulsivity domains, respectively. For this
sample, the DSM-IV Inattention Total Rating correlated 0.89 with the
Behavior Assessment System for Children (BASC) Attention Problems
T-score, and the DSM-IV Hyperactivity-Impulsivity Total Rating correlated
0.89 with the BASC Hyperactivity T-score, demonstrating validity of the
measure in assessing the constructs of inattention and hyperactivity.
Available normative data for Likert scale ratings of each of the 18 DSM-IV
(American Psychiatric Association, 1994) criteria for ADHD provide per-
centiles for various total scores across both the Inattention and Hyper-
activity-Impulsivity domains (DuPaul, Power, Anastopoulos, & Reid, 1998).
These percentiles were used in the present study to determine significant
cut-off scores for the ADHD and non-ADHD groups. To be included in the
ADHD group, children had to rank at the 85th percentile or higher (i.e.,
greater than one standard deviation above the norm) on both the
Inattention and the Hyperactivity-Impulsivity domains (Combined Type)
or on the Inattention domain alone (Predominantly Inattentive Type). Any
participants in the non-ADHD group who ranked greater than the 80th
percentile on either the Inattention or the Hyperactivity-Impulsivity
domains of the DSM-IV Checklist were excluded from the study.

Behavior Assessment System for Children – Parent Rating Scale (Reynolds &
Kamphaus, 1992). The BASC is an omnibus rating scale designed to
broadly sample a child’s behavior. Parents rate how often they observe
their child engaging in various behaviors on a 4-point scale, ranging from
‘‘never’’ to ‘‘always.’’ The ratings result in scores on 12 clinical and adaptive
scales, including two scales that specifically measure ADHD-like behaviors
(Attention Problems scale and Hyperactivity scale), which are presented as
T-scores. For clinical scales, any T-score of 60 or higher (i.e., at least one
standard deviation above the mean) is considered to be ‘‘at-risk.’’
Another criterion for the present study was that all participants included
in the ADHD group must obtain a general norm group T-score at least one
standard deviation above the mean (60 or higher; at-risk range) on both
the Attention Problems scale and the Hyperactive scale of the Behavior
Assessment System for Children – Parent Rating Scale (BASC-PRS) (Com-
bined Type) or on only the Attention Problems scale (Predominantly
Inattentive Type). This ensured that the ADHD group’s symptomatology
reached a specified threshold of severity. Likewise, a criterion for the non-
ADHD group was to obtain a general norm group T-score of less than 60
(Average range) on both the Attention Problems and the Hyperactive
scales of the BASC-PRS.
270 Journal of School Psychology

Home Situations Questionnaire (Barkley, 1987). The Home Situations

Questionnaire (HSQ) lists 16 common situations which often occur when
children are with their parents (e.g., at home, in the car, in public places).
Parents indicate if their child displays problematic behavior in each of the
situations and, if so, rates the severity of the problems on a scale from 1 to 9.
For this study, a HSQ Total Score (a summation of all severity ratings for
situations rated as being problematic), a HSQ Number of Problematic
Settings score (the total number of the 16 possible problems that were
identified as being problematic), and a Mean Severity Rating (Total Score
divided by Number of Problematic Settings) were calculated. These scores
provided not only a measure of the severity of problems, but also a measure
of problem pervasiveness.

Procedure
Prior to testing, the parent or legal guardian signed a statement of
informed consent and the participant provided their verbal assent.
Parents provided general background information and completed the
parent-report behavioral rating measures in a room separate from that
in which the experiment took place. All participants with ADHD had
been prescribed one of the following three medications by their
pediatrician or child psychiatrist: MPH (Ritalin), D,L-amphetamine (Ad-
derall), or D-amphetamine (Dexedrine). The three stimulant medica-
tions included in the study require from 30 min to 2 h to take effect
(Copeland, 1994). The duration of action of each drug is as follows: 3– 5 h
for MPH (Ritalin); 5 –8 h for the sustained release version of MPH (or
Ritalin-SR); 5– 7 h for D,L-amphetamine (Adderall); 4 –5 h for D-amphe-
tamine (Dexedrine); and 7 –8 h for the sustained release version of D-am-
phetamine (or Dexedrine spansules; Bezchlibnyk-Butler & Jeffries, 1998).
Participants with ADHD were tested when off their stimulant medication,
with the testing session scheduled a minimum of 10 h following a scheduled
dose of stimulant medication, to ensure that the medication no longer had
a beneficial effect on cognitive performance. The time elapsed since the
ADHD group’s last dosage of medication ranged from 10 to 336 h
(M = 51.05, SD= 78.38, Mdn =25.00).
In a separate testing room, participants practiced on a two-disk compu-
terized TOH problem to familiarize themselves with the testing apparatus
and to ensure that they understood the rules and the object of the task. All
participants were able to independently solve the two-disk problem in the
optimal number of moves and were allowed to continue in the study.
Participants were then given the four-disk TOH problem for one trial.
Following this, participants completed the computerized WCST. Although
no practice trials were given for the WCST, participants received instruc-
tions both verbally and on the computer screen regarding how to complete
 Barry, Lyman, and Klinger 271

the task and all appeared to understand. They then engaged in the WCST
for a maximum of 128 trials. Next, participants were given standardized
instructions and a sample task for the TMT-B. The sample task consisted of
eight circles, numbered from 1 to 4 and lettered from A to D. Any errors
were corrected as they were made and corrective feedback was given. All
participants were able to complete the sample TMT-B and, therefore, were
allowed to continue in the study. Once it was determined that the
participants understood the task, standardized instructions, and adminis-
tration of the full TMT-B was completed. Following the EF battery, all
participants were offered a short break (e.g., 5 min), following which they
were administered the K-BIT IQ test and the MBA academic achievement
test. The total time for participation of each individual participant was
approximately 90 min. Participants received ice cream or fast food gift
certificates as a thank you for participation.

RESULTS
Computation of Composites and Comparison of Group Performance
It should be noted that EF data were analyzed by examining performance
on seven key variables collected across the three measures included in the
executive battery, as well as by examining performance on a standardized
EF composite score that summarized the participants’ performance across
17 variables from the EF measures. The seven key variables included Total
Moves on the TOH; Total Categories, Trials to First Category, Percent Error,
and Perseverative Responses on the WCST; and Time and Errors on the
TMT-B. These variables were identified as key variables to include in
analyses based on previous literature (e.g., Grodzinsky & Diamond, 1992;
Pennington et al., 1993; Shue & Douglas, 1992).
An EF composite score that summarized the participants’ performance
across the three EF measures was calculated based on item analyses from
each of the variables collected from the TOH, the WCST, and the TMT-B,
provided that the variable was not completely redundant with another
variable (e.g., percent correct is redundant with percent errors). A total of
19 variables were included in the item analyses. Because most of these EF
variables are measures of impairment (e.g., errors), variables were reversed
scored where necessary so that a higher score represented a poorer perform-
ance on the EF measures. Data from an additional 19 participants (i.e.,
from those who were excluded from group analyses for failure to meet
either an ADHD or non-ADHD cut-off score) were used to increase the
power of the item analyses. (The six participants remaining from the 25
excluded participants either had significant missing data or scored within
the mentally retarded range on the K-BIT and, therefore, were not
included in the item analyses.) Two variables were excluded for having
272 Journal of School Psychology

low or negative correlations, and all item-total correlations for the revised
17-variable EF composite score were considered satisfactory (all greater
than 0.10). The standardized item coefficient alpha for the EF composite
score was 0.87. The raw scores obtained by each of the participants on the
17 variables were standardized by converting them to z-scores (M = 0, SD =1),
and these z-scores were averaged to obtain the EF composite score.
Analyses of covariance (ANCOVAs), with IQ as a covariate because of
group differences in IQ, were conducted to compare the performance of
the ADHD and non-ADHD groups across the seven key variables measuring
EF (see Table 3). Results indicated no significant group differences. An
ANCOVA, with IQ as a covariate, also was conducted to compare the
performance of the two groups on the EF composite score and revealed
that the two groups did not differ in their performance, with an average EF
composite score of 0.04 for the ADHD group (SD = 0.63) and 0.08 for the
non-ADHD group (SD =0.60), F(1,63)= 0.83, p > .05. Therefore, although
the scores were in the expected direction (i.e., ADHD group had a slightly
higher score, where a higher score represents a poorer performance), no
significant relative EF deficit occurred within the ADHD group.
Behavioral data were averaged to form an ADHD severity index. This
index was based on five variables: number of Inattentive symptoms and
number of Hyperactivity/Impulsivity Symptoms endorsed on the DSM-
IV Checklist; BASC Attention Problems T-score and BASC Hyperactivity
T-score; and the HSQ Total Score. Thus, the ADHD severity index provided
a composite of the number, severity, and pervasiveness of symptoms. Item
analyses yielded excellent item-total correlations (all greater than 0.70).
The standardized item coefficient alpha for the ADHD severity index was

Table 3
Comparison of the ADHD and Non-ADHD Groups on Key Measures of EF a

ADHD (n = 33) Non-ADHD (n = 33)

Mean (SD) Mean (SD) F

TOH
Total Moves 39.36 (24.93) 42.15 (26.67) 1.01

WCST
Total Categories 4.24 (1.79) 4.70 (1.74) 0.02
Trials to First Category 18.70 (15.16) 22.85 (22.62) 0.22
Percent Errors 33.31 (17.29) 29.96 (16.40) 0.54
Perseverative Responses 22.82 (18.15) 19.91 (15.01) 0.32

TMT-B
Time (s) 101.36 (57.86) 80.03 (30.88) 0.07
Errors 1.03 (2.16) 0.55 (0.94) 0.04
Note. TOH=Tower of Hanoi; WCST = Wisconsin Card Sorting Test; TMT-B = Trail Making Test,
Part B; EF = Executive Functioning.
a
Controlling for IQ.
 Barry, Lyman, and Klinger 273

0.94. Raw scores for each of the five variables were converted to z-scores
and then averaged to form the ADHD severity index. As expected, the
ADHD group demonstrated a significantly higher ADHD severity index
(M =0.90, SD =0.57) than did the non-ADHD group (M= 0.87, SD=0.23),
t(64)= 16.45, p < 0.001.

Academic Achievement
Results of independent-samples t-tests demonstrated that the ADHD group
obtained lower scores in all academic subjects measured by the MBA, when
compared to the non-ADHD group (all ps > 0.05). However, because the
ADHD group also obtained significantly lower IQ scores, it is more mean-
ingful to investigate potential group differences using a discrepancy score
(i.e., between predicted achievement, based on IQ, and actual achieve-
ment). Accordingly, for each participant, a predicted achievement score-
based on the obtained IQ score was calculated using a model that accounts
for both regression to the mean (i.e., extremely high and extremely low IQ
scores would be associated with less extreme predicted achievement scores)
and the fact that intellectual and achievement tests are correlated (i.e., the
model assumes a 0.65 level of correlation; Alabama State Department of
Education Special Education Services, 1999).
Based on this predicted achievement score, a discrepancy score repre-
senting any disparity from predicted achievement and actual achievement
was calculated for the three measured academic subject areas, as well as an
overall skills composite scale. The discrepancy scores were computed so that
a positive score represents actual achievement above that which would be
predicted given the participant’s estimated intellectual level, and a negative
score represents actual achievement below that which would be predicted.
Overall, the ADHD group scored below prediction in all academic areas,
whether compared to same-age peers or to peers at their same grade-level
(i.e., all discrepancy scores were negative). In contrast, the non-ADHD
group’s age-based and grade-based achievement scores were below predic-
tion only in Writing. The non-ADHD group scored above predicted
achievement scores in Reading and Mathematics, as well as on the Basic
Skills cluster score. Because IQ is an age-based index, the discrepancy
scores derived from age-based achievement scores are the most meaningful
and will be the focus of further analyses. Independent-samples t-tests
revealed that the ADHD group’s discrepancy between predicted and actual
achievement was significantly greater than the non-ADHD group’s aca-
demic discrepancy in all academic subject areas (see Table 4).
An LD involves achievement below expectation based on intellectual
level (American Psychiatric Association, 1994) and is usually diagnosed in
school systems based on a discrepancy formula comparing predicted
academic achievement and actual achievement (Alabama State Depart-
274 Journal of School Psychology

Table 4
Discrepancya Between Predicted Achievement and Actual Achievementb by Diagnostic Group

ADHD (n = 33) Non-ADHD (n = 33)

Discrepancy Score Mean (SD) Mean (SD) t-Value

Basic Skills 10.00 (10.28) 0.64 (7.41) 4.82***

Reading 5.94 (11.42) 3.52 (9.87) 3.60**
Writing 14.64 (10.53) 4.79 (10.41) 3.82***
Mathematics 4.06 (15.04) 10.39 (12.77) 4.21***

ADHD/non-LD (n = 25) Non-ADHD (n = 33)

Discrepancy Score Mean (SD) Mean (SD) t -Value

Basic Skills 5.60 (6.73) 0.64 (7.41) 3.30**

Reading 1.80 (9.09) 3.52 (9.87) 2.10*
Writing 10.56 (7.57) 4.79 (10.41) 2.34*
Mathematics 1.52 (12.38) 10.39 (12.77) 2.66*
a
Represents discrepancy between predicted achievement based on the Kaufman Brief
Intelligence Test Composite IQ (Kaufman & Kaufman, 1990) and actual achievement on the
MBA (Woodcock et al., 1994), using a regression to the mean model (Alabama State
Department of Education Special Education Services, 1999).
b
Age-based standard scores.
*p < 0.05.
**p < 0.01.
***p < 0.001.

ment of Education Special Education Services, 1999). Eight of the partic-

ipants with ADHD in this sample had a comorbid diagnosis of an LD, based
on parental report, whereas none of the non-ADHD participants were
reported by their parents to have an LD diagnosis. Given that an LD, by
definition, involves academic underachievement, the data were reanalyzed,
excluding the eight participants with comorbid LD and ADHD. Results
revealed that the ADHD group (n =25) continued to exhibit a greater
discrepancy between intellectual level and academic achievement (in a
negative direction) in each academic area measured, when compared to
the non-ADHD group (n =33; see Table 4).

Predictors of Academic Underachievement

To determine the relative effects of either severity of ADHD behaviors or
EF impairments on academic underachievement, hierarchical stepwise
multiple regression analyses that reversed the order of the entry of the
ADHD severity index and the EF composite score were conducted to
predict the age-based discrepancy scores in each academic area. These
analyses included all participants in the ADHD and non-ADHD groups
(n = 66). The first set of analyses evaluated the extent to which the ADHD
severity index predicted academic underachievement over and above
 Barry, Lyman, and Klinger 275

performance on measures of EF, based on the EF composite score, for each

academic area. The second set of analyses evaluated the extent to which the
EF composite score predicted academic underachievement over and above
the ADHD severity index for each academic area.
Results indicated that the ADHD severity index accounted for a signifi-
cant proportion of the variance in the academic discrepancy scores, in all
academic areas, after controlling for any effects of EF. In contrast, perform-
ance on EF measures accounted for a significant proportion of the variance
in the academic discrepancy scores for the Basic Skills cluster and Math-
ematics only (see Table 5). These results indicate that children who had
more severe symptoms of ADHD had lower academic discrepancy scores
(i.e., tended to be academic underachievers) in all areas, whereas children
who had more impaired performance on EF measures had lower academic
discrepancy scores in Mathematics only.
The series of hierarchical stepwise multiple regression analyses were
repeated excluding any ADHD participants with a comorbid diagnosis of an
LD (total n =58). The pattern of results for the ADHD severity index was
similar to that of the larger sample, with greater levels of ADHD sympto-
matology predicting academic underachievement, when controlling for EF.
However, in this non-LD sample, the EF composite score did not account

Table 5
Multiple Regression Analyses: ADHD Behaviors and EF Deficits as Predictors of
Academic Underachievement

Predictor Variables

ADHD severity index EF composite score

Criterion
2 2
Variables R change F change (1,63) b R change F change (1,63) b

n = 66a
Basic Skills 0.21 19.19*** 0.46 0.07 6.23* 0.26
Reading 0.12 8.98** 0.35 0.03 2.27 0.17
Writing 0.19 15.36*** 0.44 0.01 0.82 0.10
Mathematics 0.15 13.02** 0.39 0.12 10.33** 0.34

Criterion
Variables R 2 change F change (1,55) b R2 change F change (1,55) b

n = 58b
Basic Skills 0.18 12.43** 0.43 0.004 0.25 0.06
Reading 0.08 4.46* 0.27 0.000 0.12 0.01
Writing 0.14 8.71** 0.37 0.003 0.21 0.06
Mathematics 0.10 6.57* 0.32 0.027 1.73 0.17
Note. Discrepancy scores based on age-based academic achievement scores.
a
Includes comordid ADHD and LD participants.
b
Excludes comordid ADHD and LD participants.
*p < 0.05.
**p < 0.01.
***p < 0.001.
276 Journal of School Psychology

for a significant proportion of unique variance in the academic discrepancy

scores in any of the academic areas (see Table 5).
In a structural equation modeling (SEM) framework, the data were
analyzed using LISREL v.8.51 (Jöreskog & Sörbom, 2001). The models
tested were covariance structure models with multiple indicators for all
latent constructs, which allowed a test of the EF and ADHD severity
variables as simultaneous predictors of academic underachievement,
with each academic subject area also tested simultaneously. Data for the
85 participants with IQ >70 and for which all data had been obtained were
used in an effort to maximize sample size, which is important for this
statistical procedure. In this approach, performance across the three EF
measures (TOH, WCST, and TMT-B) were used to represent the latent
construct of EF; ratings across the three behavioral measures (DSM-IV
Checklist-Inattention and Hyperactivity, BASC-Attention Problems and
Hyperactivity, and HSQ) were used to represent the latent construct of
ADHD severity; and academic discrepancy scores for reading, writing, and
mathematics (based on performance on the MBA) were used to represent
the latent construct of academic underachievement. In the first model,
path coefficients predicting the latent construct of academic underachieve-
ment from the latent constructs of ADHD severity and EF were estimated
and demonstrated an acceptable fit, v2 (24, N= 85) =38.335, p <0.05, good-
ness of fit index (GFI) = 0.990. Both paths were significant, 0.41 and 0.49,
for ADHD severity and EF, respectively (both ps < 0.05), indicating that both
constructs significantly predicted academic underachievement. Next, a
model was tested in order to determine if the relative predictive utility was
essentially identical for ADHD severity and EF. This model, for which the
path coefficients between ADHD severity and academic underachievement
and between EF and academic underachievement were held equal, also
provided an acceptable fit, Dv2 (1, N = 85)=1.59, not significant. Therefore,
the magnitude of the effect of the two constructs (EF and ADHD severity)
on overall academic underachievement is considered to be the same, and
both are good predictors independent of one another.

Gender Differences
A series of two-way ANOVAs were conducted with diagnostic group and
gender as the between-subjects variables and the four age-based discrep-
ancy scores as the dependent variables for each of the participants (n =66).
In addition to the main effect for diagnostic group for each subject area, an
interaction between diagnostic group and gender emerged for the Basic
Skills cluster, F(1,62)=5.45, p = 0.023, partial g2 =0.08, and Reading,
F(1,62)=9.23, p = 0.003, partial g2 = 0.13. For Writing, a main effect for
gender was found, F(1,62)=7.03, p = 0.010, partial g2 = 0.10, with males tend-
ing to exhibit greater academic underachievement (M = 12.98, SE =1.69)
 Barry, Lyman, and Klinger 277

Figure 1. Gender by diagnostic group interaction for academic underachievement in Reading.

than females (M = 6.32, SE=1.86) in this academic area. Finally, for Math-
ematics, no effect for gender or interaction between gender and diagnostic
group was evidenced. The results for Reading and Writing remained
significant when applying the Bonferonni correction to the post hoc analyses
for these four dependent variables (significance level set at .0125). Each of
these analyses was repeated excluding all children with a comorbid diagnoses
of ADHD and LD (n =58 included) with a similar pattern of findings.
Because the interaction between gender and diagnostic group was
significant for Reading, potential differences among the two diagnostic
groups for males and females were explored separately (see Figure 1).
There was no significant difference between diagnostic groups for males,
F(1,34)=0.29, p = 0.595; however, there was for females, F(1,34) = 26.51,
p < 0.001. Results indicated that female participants with ADHD exhibited
significantly greater academic underachievement in Reading (M= 8.83,
SD = 2.59) than did non-ADHD females (M = 8.39, SD = 2.12). The non-
ADHD females tended to achieve much higher than would be predicted
given their level of intellectual functioning in Reading, which is tradition-
ally an area of academic strength for females (Vogel, 1990).

DISCUSSION
The children with ADHD in the present sample did not exhibit significant
impairments in EF. Nevertheless, the ADHD group was clearly impaired
behaviorally, as evidenced on the behavioral rating forms completed by the
parents. Likewise, the children with ADHD were significantly more
278 Journal of School Psychology

impaired academically. First, eight of the children in the ADHD group had a
comorbid diagnosis of an LD, whereas none of the non-ADHD children had
an LD diagnosis. Second, the children within the ADHD were more likely to
have been placed in a special education class at some time within their
academic career than were non-ADHD children. Finally, on standardized
measures of academic achievement administered in the laboratory, children
with ADHD exhibited significant academic impairment both by clinical
standards and when compared to the group of non-ADHD children.
The support of the a priori hypothesis that the ADHD group would
exhibit academic underachievement is a replication of findings of previous
researchers (e.g., Frick et al., 1991). This was a robust finding, with the
ADHD participants having a significantly greater discrepancy, in the
negative direction, than the non-ADHD group in each academic area
measured. Furthermore, these findings held even when the eight partic-
ipants with comorbid ADHD and LD were excluded from the analyses,
suggesting that the academic underachievement observed within the
ADHD group cannot be explained exclusively in terms of comorbid
learning disabilities.

Clinical Implications
In addition to diagnostic group differences in academic underachieve-
ment, regression analyses indicated that, as children exhibited a greater
number, severity, and pervasiveness of ADHD behaviors (measured by the
ADHD severity index), they were more likely to achieve academically at a
level below what would be predicted by their intellectual level. Hence, not
only a categorical diagnosis of ADHD, but also the severity and pervasive-
ness of the ADHD symptoms, is a good predictor of academic underach-
ievement. In addition, the ADHD behaviors predicted academic
underachievement over and above performance on measures of EF for
each of the academic areas. The predictive utility of the ADHD severity
index remained even when excluding the participants with comorbid
ADHD and LD. In contrast, EF performance predicted academic under-
achievement over and above ADHD behaviors in only one academic
subject, Mathematics, and, this was no longer the case when participants
with LD were excluded. Examination of a larger sample (including
children with a magnitude of ADHD behaviors that excluded them from
a non-ADHD group but which were subthreshold for a diagnosis of ADHD)
within an SEM framework also demonstrated that severity of ADHD
behaviors is a good predictor of academic underachievement, independent
of EF. In sum, results of this study indicate that more severe ADHD
behaviors were associated with more severe impairment in academic
functioning, and this impairment was not accounted for by comorbid LD
or by cognitive deficits associated with EF.
 Barry, Lyman, and Klinger 279

Children with ADHD may have ‘‘what may be called a producing

disability impeding their academic careers. They may be learning but still
underachieving because they do not produce on papers and tests’’
(Denckla, 1994, p. 117). Although Denckla notes that this ‘‘producing
disability’’ may be related to executive dysfunction, it is possible that it is
related to other behavioral manifestations of ADHD (e.g., off-task behav-
iors) or to more specific frontal lobe dysfunction (e.g., response disinhibi-
tion, poor sustained attention), rather than to global EF impairments. If so,
children with ADHD may have intact higher order problem-solving skills
but are simply engaged in other off-task behaviors and are not focused a
sufficient amount of time to apply these skills.
These findings indicate that, at least for some academic areas (e.g.,
reading, writing), children with ADHD may be more negatively impacted
by their disruptive behaviors, which likely impede their ability to learn and
master new academic material, rather than by an innate neurocognitive
deficit. This is consistent with findings of research specifically evaluating
the role of planning in school achievement outside of the context of
ADHD (Cohen, Bronson & Casey, 1995). It was found that neuropsycho-
logical measures of planning (i.e., the WCST and the TMT-B) did not
significantly predict academic performance (i.e., grades across academic
subjects) when IQ was controlled, whereas teacher ratings of behavior
(e.g., uses time wisely) did significantly predict school achievement
(Cohen et al., 1995).
Examining the relation between ADHD and academic underachieve-
ment through the use of a quantitative measure of ADHD (i.e., the ADHD
severity index), rather than simply comparing ADHD and non-ADHD
groups, demonstrated that children with more severe ADHD behaviors
generally had worse achievement than those children with less severe
behaviors. However, our current diagnostic classification system (DSM-IV;
American Psychiatric Association, 1994) is a categorical system which
describes the absence or presence of an ADHD diagnosis but does not
capture the severity of the problems. These results, which demonstrate that
severity of problems is a significant predictor of academic impairment,
suggest that a consideration of the severity of problems (not just the
presence or absence of a diagnosis) is also important, particularly in terms
of academic outcomes.
Children exhibiting academic underachievement will likely benefit from
educational interventions that directly address school achievement, such as
remedial instruction, tutoring, and training in study skills. Furthermore,
these results suggest that children with ADHD (or even those at a subthres-
hold diagnostic level) will benefit from behavioral strategies that address
their symptoms of ADHD and help them to achieve to their potential in the
classroom. Behavioral interventions, such as token reinforcement programs,
contingency contracting, time-out from positive reinforcement, home-based
280 Journal of School Psychology

contingencies, home-school note programs, and self-management interven-

tions have proven effective in decreasing the negative behaviors associated
with ADHD (see DuPaul & Stoner, 1994, for a review) and may, subse-
quently, lead to improvement in school performance.
Furthermore, this study suggests that the gender of the child may
indicate certain academic areas that require more attention through
interventions than others. For example, girls may require more behavioral
structure in Reading classes or classes for which reading is essential for
success (e.g., science, social studies).
Finally, the undiminished performance of children with ADHD on
problem-solving tasks and tests of cognitive flexibility suggests that these
children have cognitive strengths that can be capitalized upon. Future
treatment efforts for children with ADHD, particularly academic-related
programs, should not only address the children’s cognitive deficits but also
build upon their noted cognitive strengths.

Limitations of the Present Study and Suggestions for Future Research

The limitations of the present study should be considered when planning
further exploration of these research questions. First, the sample partic-
ipating in this study was predominantly Caucasian and predominantly
middle class. Consequently, it is unclear whether or not these findings
would extend to other populations (e.g., African-American; lower SES).
Likewise, inclusion of a small number of children with ADHD-Predom-
inantly Inattentive Type precluded any analyses by subtype of the disorder.
In addition, given that the test administration was constant, rather than
counterbalanced, the differential effects of fatigue during the relatively
long test administration time (i.e., 90 min) between the ADHD and non-
ADHD groups cannot be determined. Specifically, it is not possible to
completely rule-out fatigue as a factor contributing to the impaired
performance of the ADHD group on the academic achievement tests,
which were consistently administered last.
In addition, although the relation between ADHD behaviors and aca-
demic underachievement was explored excluding data from participants
with comorbid LD, it should be noted that the remaining ‘‘ADHD-only’’
group may have included some children with an undiagnosed LD. Unfortu-
nately, this cannot be ruled-out by the data available. A substantial number
of the ‘‘ADHD-only’’ group obtained an academic underachievement score
for at least one academic subject that was more than one standard deviation
below their predicted achievement score. This would be in line with the
standard school system criteria for diagnosing an LD (Alabama State
Department of Education Special Education Services, 1999). Although
the presence of undiagnosed learning disabilities was also a possibility
within the non-ADHD group (i.e., there were some non-ADHD children
 Barry, Lyman, and Klinger 281

underachieving in a clinically relevant range), it would be more likely to

have occurred within the ADHD group, given that ADHD and LD are often
comorbid. Likewise, the current study did not exclude individuals who had
comorbid diagnoses. Although the base rate of comorbidity (other than LD
within the ADHD group; n =8) was quite low for both the ADHD group
(n =3) and the non-ADHD group (n= 1), it is still not possible to determine
the extent to which group differences may have been influenced by these
other disorders.
Despite these limitations, the present study provides further clarification
to some issues surrounding ADHD and also raises engaging questions that
should be addressed in future research endeavors, such as a focus on the
efficacy of treatments addressing the academic underachievement
observed in children with ADHD. Specifically, studies evaluating behav-
ioral strategies that minimize the negative impact of ADHD behaviors in
the classroom, as well as those evaluating the effectiveness of academic
programs that build upon the cognitive strengths of children with ADHD,
are indicated. The ultimate goal should be to identify treatment strategies
that assist children with ADHD in achieving academically to their intel-
lectual potential.

ACKNOWLEDGMENTS

The completion of this manuscript has been partially supported by a

grant from the National Institute on Drug Abuse (DA08453) to Dr. John E.
Lochman (Principal Investigator). The authors are appreciative of the
parents and children who gave of their time to participate in this study.
Appreciation is also expressed to Dr. Michael Taylor, Associate Professor
and Chairman of Pediatrics, The University of Alabama College of
Community Health Sciences/Capstone Medical Center, as well as to the
faculty and staff at the Tuscaloosa City School System and the Tuscaloosa
County School System, for their invaluable contribution to the recruitment
of participants.

REFERENCES

Alabama State Department of Education Special Education Services (1999, July). Table SLD-1.
Aman, C. J., Roberts, R. J., & Pennington, B. F. (1998). A neuropsychological examination of
the underlying deficit in ADHD: the frontal lobe vs. right parietal lobe theories. Develop-
mental Psychology, 34, 956 – 969.
American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders.
(4th ed.). Washington, DC: Author.
Banich, M. T. (1997). Executive function. In M. T. Banich (Ed.), Neuropsychology: the neural bases
of mental function ( pp. 368 – 397). Boston, MA: Houghton Mifflin.
282 Journal of School Psychology

Barkley, R. A. (1987). Defiant children: parent – teacher assignments. New York: Guilford Press.
Barkley, R. A. (1997a). Attention-deficit/hyperactivity disorder, self-regulation, and time: to-
ward a more comprehensive theory. Developmental and Behavioral Pediatrics, 18, 271 – 279.
Barkley, R. A. (1997b). Behavioral inhibition, sustained attention, and executive functions:
constructing a unifying theory of ADHD. Psychological Bulletin, 121, 65 – 94.
Bezchlibnyk-Butler, K. Z., & Jeffries, J. J. (1998). Clinical handbook of psychotropic drugs, 8th revised
edition. Toronto, Canada: Hogrefe and Huber Publishers.
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.
Cohen, G., Bronson, M. B., & Casey, M. B. (1995). Planning as a factor in school achievement.
Journal of Applied Developmental Psychology, 18, 405 – 428.
Copeland, E. D. (1994). Medications for attention disorders (ADHD/ADD) and related medical prob-
lems: a comprehensive handbook. Atlanta, GA: Resurgens Press.
Denckla, M. B. (1994). Measurements of executive function. In G. R. Lyons (Ed.), Frames of
reference for the assessment of learning disabilities: new views on measurement issues ( pp. 117 – 142).
Baltimore, MD: Paul H. Brookes Publishing.
DuPaul, G. J., & Stoner, G. (1994). ADHD in the Schools: assessment and intervention strategies. New
York: Guilford Press.
DuPaul, G. J., Power, T. J., Anastopoulos, A. D., & Reid, R. (1998). ADHD rating scale-IV:
checklists, norms, and clinical interpretation. New York, NY: Guilford Publications.
Frick, P. J., & Lahey, B. B. (1991). The nature and characteristics of attention-deficit hyper-
activity disorder. School Psychology Review, 20, 163 – 173.
Frick, P. J., Kamphaus, R. W., Lahey, B. B., Loeber, R., Christ, M. G., Hart, E. L., & Tannen-
baum, L. E. (1991). Academic underachievement and the disruptive behavior disorders.
Journal of Consulting and Clinical Psychology, 59, 289 – 294.
Grant, D. A., & Berg, E. A. (1948). The Wisconsin card sorting test: directions for administration and
scoring. Odessa, FL: Psychological Assessment Resources.
Grodzinsky, G. M., & Diamond, R. (1992). Frontal lobe functioning in boys with attention-
deficit hyperactivity disorder. Developmental Neuropsychology, 8, 427 – 445.
Heaton, R. K., Chelune, G. J., Talley, J. L., Kay, G. G., & Curtiss, G. (1993). Wisconsin card sorting
test manual, revised and expanded. Odessa, FL: Psychological Assessment Resources.
Hynd, G. W., Semrud-Clikeman, M., Lorys, A. R., Novey, E. S., & Eliopulas, D. (1991). Brain
morphology in developmental dyslexia and attention deficit disorder/hyperactivity. Archives
of Neurology, 47, 919 – 926.
Jöreskog, K. G., & Sörbom, D. (2001). LISREL 8.51. Lincolnwood, IL: Scientific Software
International.
Kamphaus, R. W., & Frick, P. J. (1996). Clinical assessment of child and adolescent personality and
behavior. Needham Heights, MA: Allyn and Bacon.
Kaufman, A. S., & Kaufman, N. L. (1990). Kaufman brief intelligence test (K-BIT). Circle Pines,
MN: American Guidance Service.
Leon-Carrion, J. (in press). Sevilla computerized neuropsychological test battery. The Univer-
sity of Sevilla, Spain.
Loong, J. W. (1990). The Wisconsin card sorting test. Los Angeles, CA: Western Psychological
Services.
Loring, D. W. (1999). INS dictionary of neuropsychology. New York, NY: Oxford Univ. Press.
Ozonoff, S., & Jensen, J. (1999). Brief report: specific executive function profiles in three
neurodevelopmental disorders. Journal of Autism and Developmental Disorder, 29, 171 – 177.
Pennington, B. F., & Ozonoff, S. (1996). Executive functions and developmental psychopathol-
ogy. Journal of Child Psychology and Psychiatry, 37, 51 – 87.
Pennington, B. F., Groisser, D., & Welsh, M. C. (1993). Contrasting cognitive deficits in attention
deficit hyperactivity disorder versus reading disability. Developmental Psychology, 29, 511 – 523.
 Barry, Lyman, and Klinger 283

Reitan, R. M. (1958). Validity of the trail making test as an indicator of organic brain damage.
Perceptual and Motor Skills, 8, 271 – 276.
Reitan, R. M., & Wolfson, D. (1993). The Halstead – Reitan neuropsychological test battery:
theory and clinical interpretation. (2nd ed.). Tucson, AZ: Neuropsychology Press.
Reynolds, C. R., & Kamphaus, R. W. (1992). Behavior assessment system for children (BASC). Circle
Pines, MN: American Guidance Services.
Shue, K. L., & Douglas, V. I. (1992). Attention deficit hyperactivity disorder and the frontal
lobe syndrome. Special Issue: The role of frontal lobe maturation in cognitive and social
development. Brain and Cognition, 20, 104 – 124.
Vogel, S. A. (1990). Gender differences in intelligence, language, visual-motor abilities, and
academic achievement in students with learning disabilities: a review of the literature.
Journal of Learning Disabilities, 23, 44 – 52.
Weintraub, S. (1993). Neuropsychological assessment in focal lesions and learning disabilities.
In M. M. Mesulam (Ed.), Neurology of behavior ( pp. 22 – 36). Boston, MA: Parker House.
Weyandt, L. L., & Willis, W. G. (1994). Executive functions in school-aged children: potential
efficacy of tasks in discriminating clinical groups. Developmental Neuropsychology, 10, 27 – 38.
Woodcock, R. W., McGrew, K. S., & Werder, J. K. (1994). Woodcock – McGrew – Werder Mini-Battery
of Achievement (MBA). Chicago, IL: Riverside Publishing.
Zentall, S. S., Smith, Y. N., Lee, Y. B., & Wieczorek, C. (1994). Mathematical outcomes of
attention-deficit hyperactivity disorder. Journal of Learning Disabilities, 27, 510 – 519.