WIN/”WWWWWI”WWW“"WW

134
127
THS

    

”'8“; JIllllllllfllfllfllfilllflllfilflllllllL

A I, K‘
‘1 ‘> r J j (4/

 

LIBRARY

Michigan State
University

 

 

 

This is to certify that the

thesis entitled

Learning Disability Dissimulation:
An Analogue Study

presented by

Adam Daniel Alban

has been accepted towards fulfillment
of the requirements for

MA degree in W094

llL AM

Major professor

 

 

Date 6/12! (l, 9

0-7539 MS U is an Affirmative Action/Equal Opportunity Institution

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1B8 chlRC/DataDmpfiS—p.“

LEARNING DISABILITY DISSIMULATION: AN ANALOGUE STUDY
By

Adam Daniel Alban

A THESIS

Submitted to
Michigan State University
In partial fulfillment of the requirements

for the degree of

MASTER OF ARTS

Department of Psychology

1999

.n‘l. I lirlll I

ABSTRACT
LEARNING DISABILITY DISSIMULATION: AN ANALOGUE STUDY
By

Adam Daniel Alban

This research was an effort to understand the possibility of dissimulating on
tests of learning characteristics. This project sought to answer the following
questions: Is it possible to dissimulate on tests of learning characteristics, and are
these attempts detectable? A sample of college students (N=40) was unable to
replicate an aptitude/achievement discrepancy where achievement was lower than
aptitude. Instead, they displayed an aptitude/achievement discrepancy where aptitude
was lower than achievement. Tests designed to detect dissimulation in traditional
neuropsychological assessments were found to be helpfiil in detecting an attempt to
dissimulate. In addition, participants were relatively accurately grouped into the

dissimulation and control conditions solely on the basis of these specialized tests.

Capyright by
Adam Daniel Alban
1999

For my grandparents, parents, and brother. Thank you.

iv

ACKNOWLEDGEMENTS

I would like to thank my thesis chair, Dr. Norman Abeles, for his invaluable
guidance. His support on all of my projects has been instrumental. I would also like
to thank the other members of my thesis committee, Dr. Robert Caldwell and Dr.
John Hurley, for their instructive support and feedback. In addition, I could not have
completed this project without the support of my family. Last, but certainly not least,
I would like to thank my fellow students, who continue to be a source of energy and

inspiration.

TABLE OF CONTENTS

List of Tables ................................................................................. iv
Introduction .................................................................................. 1
Learning Disabilities and the Law ............................................... 3
Court Cases ......................................................................... 4
Learning Disabilities ............................................................... 8

Learning Disabilities and Neuropsychology ................................... 10

The Woodcock Johnson Psycho-Educational Test Battery .................. 13

Neuropsychological Efforts to Detect Malingering ........................... 17

Dissimulation Measures .......................................................... 18
Hypotheses .................................................................................. 21
Methods ..................................................................................... 23

Cognitive Measure ................................................................. 23

Procedure ........................................................................... 24
Results ....................................................................................... 27
Aptitude and Achievement and Learning Disability Faking ................ 27

Tests of Malingering .............................................................. 30

Rey Fifteen-Item Test .................................................. 30

Forced Choice Test ...................................................... 32

Rey Dot Counting Test .................................................. 32

Sex Effects ......................................................................... 35
Post-Hoe Analysis ................................................................ 36
Discussion .................................................................................. 36
Future Research .................................................................. 42
References .................................................................................. 44

vi

LIST OF TABLES

Table 1 Mean z-scores for the tests of cognitive ability and the

tests of achievement .............................................................

Table 2 Descriptive Statistics for the tests from the cognitive portion of the

WJ-R, irrespective of order .....................................................

Table 3 Descriptive statistics for the tests fiom the achievement portion

of the WJ-R, irrespective of order ..............................................
Table 4 Descriptive statistics for the AMNART, irrespective of order .............
Table 5 Descriptive statistics for the Rey 15-item test, irrespective of order. . . . .
Table 6 Descriptive statistics for the Forced Choice Test, irrespective of order. . ..
Table 7 Descriptive statistics for the Dot Counting Test, irrespective of order. .

Table 8 Order effect for Dot Counting Test, errors in the ungrouped condition. .

vii

..35

...36

..36

..37

.38

.39

.42

42

INTRODUCTION

Faking or dissimulation on psychological tests has long been a concern of
professionals involved with assessment. The benefits to be gained from certain test
results may be impetus for some to attempt to alter their performance. For example,
individuals who convince a mental health professional of their lack of competence may
believe that they can secure special benefits such as social security disability payments.
If the test results for individuals accused of breaking the law indicate mental illness, the
adjudication process may be changed and their culpability diminished.

This sort of dissimulation was first recorded in biblical scriptures where David
faked insanity because he feared punishment by King Achish of Gath (Samuel I, 21 : 13-
16). The investigation of this dissimulation has an almost equally long history. Many of
the original writings on dissimulation concern avoiding military service. Militaries were
more concerned with losing able-bodied soldiers than any concern for the soldiers’
welfare (Mendelson, 1995) and conversely, many able bodied individuals valued health
over patriotism.

More recently, the Diagnostic and Statistical Manual of Mental Disorders (DSM
IV) has begun to recognize different types of illness and/or injury faking. Conditions
such as F actitious Disorders, where an individual assumes the role of a sick person and
intentionally feigns symptoms are recognized as forms of dissimulation lacking apparent
external incentives. This sort of dissimulation is distinguished fi'om other forms of
dissimulation, known as “malingering,” where an individual has a recognizable purpose
and manufactures symptoms to achieve his/her goal.

The term “malinger” is more specific than dissimulate. Whereas “dissimulate”

refers to a general act of hiding one’s motives (Webster, 1988), the term “malinger”
refers more specifically to the act of feigning incapacitation in order to avoid work
(Webster, 1988). The word “malinger” comes from the French word, “malingre”, which
is often translated to mean “sickly, puny, frail, or ailing” (Webster, 1988).

Just as some individuals may be tempted to fake a mental illness, other
individuals may be tempted to fake cognitive impairment. Just as those who receive a
diagnosis of “mentally ill” may receive certain benefits and privileges, those identified as
cognitively impaired may also receive benefits, such as social security disability
insurance, workman’s compensation insurance, and/or monetary compensation as
mandated by a court order.

The problem of malingering on tests of psychopathology has received a large
amount of attention. Much of this research deals with the MMPI and MMPI-2
(Hathaway & McKinley, 1951; Butcher, Dahlstrom, et a1., 1989) and its “fake” scale.
Research has investigated the ability of psychologists to use the MMPI and MMPI-2 to
detect and differentiate among faking both psychopathology and perfect mental health
(Lim & Butcher, 1996). Numerous studies have demonstrated this scale’s usefulness for
detecting dissimulation (Cassisi, 1992; Austin, 1992; Viglione, 1995).

Attempts to detect malingering on psychological tests in general have had mixed
results. This is partly the result of an unclear distinction between dissimulation and
psychopathology. Some researchers have argued that malingering itself is evidence of
psychopathology (Lezak, 1995).

The problem of malingering in neuropsychological assessments has involved the

interpretation of already existing tests as well as the creation of several new scales. Until

recently, the detection of malingering has relied on the clinician’s ability to detect
discrepant test results and complaints (Lezak, 1995).

Just as some may be motivated to malinger on tests of psychopathology or
cerebral dysfunction, others may be motivated to simulate a learning disability (LD).
Federal law (Americans with Disabilities Act) mandates that persons receiving the
diagnosis of a physical or mental disability are entitled to additional services and benefits.
The diagnosis of Learning Disability qualifies as just such a disability (Parry, 1992).

This research project examined the applicability of already existing tests for
malingering to the assessment of learning characteristics. More specifically,
neuropsychological tests of malingering may be useful due to the many similarities
between neuropsychological and psycho-educational tests. Are those tests commonly
used to detect malingering in neuropsychological assessments applicable to learning

disability assessments?

LgamingDisabilities and The Law

A resurgence of interest since the 19708 (Hughes & Smith, 1990) has produced a
great deal of literature on the subject of learning disabilities. New federal laws
mandating improved services for disabled students have produced demands for more
information, and this new information has, in turn, influenced policies. Frank and
Wade (1993) point out that under various laws and federal mandates such as The
Americans with Disabilities Act (ADA) (1992) and the Rehabilitation Act of 1973, all
postsecondary institutions receiving federal aid are required to “make such modifications

to [their] academic requirements as are necessary to ensure that such requirements do not

discriminate or have the effect of discriminating, on the basis of a handicap, against a
qualified handicapped applicant or student.” The protection of these two Acts is
extended to persons with learning disabilities. Public Law 94-142 defines learning

disabilities in the following way:

“A disorder in one or more of the basic psychological processes
involved in understanding or in using language, spoken or written, which
disorder may manifest itself in imperfect ability to listen, think, speak,
read, write, spell, or do mathematical calculations. Such disorders include
such conditions such as perceptual handicaps, brain injury, minimal brain
dysfunction, dyslexia, and developmental aphasia. Such term does not
include children who have learning problems which are primarily the
result of visual, hearing, or motor handicaps, of mental retardation, of
emotional disturbance, or environmental, cultural, or economic

disadvantage. (PL 94-142)”

The American Bar Association (ABA) has adopted a formal definition of learning
disabilities that is virtually identical to the original federal definition (PL 94-142), with
one important addition. The ABA adds that learning disabilities are “ . . .manifested by a
significant discrepancy between intellectual ability and actual achievement . . .(Parry,
1992)”

Postsecondary institutions are expected to make appropriate adjustments to assist

these learning disabled students, but these adjustments need not be made if the

adjustments effectively eliminate essential course content. For example, a student with a
mathematics learning disability who declares a major in mathematics does not need to be
exempted from course requirements, but the same school may relieve a clarinet major of
his/her mathematics requirement. It is up to the institution to determine the distinction
between compromising the effectiveness of the course material and assisting a student

with a disability.

§o__ur_t_(_2as_e_8.

Disabled persons have fought a long and hard battle to gain access to educational
services formerly reserved for students without disabilities. Their struggles have
sometimes led to a court decision that more clearly defines the often-vague legislation
surrounding this access.

In the United States v. Board of Trustees of the University of Alabama (1990), a
federal court found that the University of Alabama could not limit the benefits of federal
legislation to only students registered in degree granting programs. This court decision
extended the availability of disabled student services to students in nondegree-granting
programs, postgraduate students, and temporary students. The court found that the
University must provide auxiliary services to disabled students, regardless of their
enrollment status.

Other court decisions have influenced the ways in which post-secondary
institutions change their testing policies. For example, in Wynne v. Tufts University
School of Medicine (1990), a medical student who had failed several classes claimed that

the cause for these academic failures was dyslexia. The Plaintiff attempted to take a

reduced course load, but was unsuccessfiil in his efforts to do so. In addition, the medical
school refused him the opportunity to be tested with an alternate testing method (e.g., oral
examinations instead of essay examinations). The Plaintiff was, however, given a tutor
and an opportunity to take the classes a second time. Afier failing to pass all of his
classes for the second time, he was dismissed from the medical school. This suit claimed
that the methods by which he was tested did not allow him the opportunity to
demonstrate his knowledge of the material. A district court found for the medical school,
agreeing with the medical school’s argument that a major Shift in the testing format
constituted a major alteration to the medical school’s program. Upon appeal, however,
the appellate court found that the district court was incorrect to grant the medical school
the power to insist that a dyslexic student be evaluated in the one way which is most
difficult for students with dyslexia, a written examination. Upon return, the district court
found for the student.

In Campbell A. Dinsmore v. Charles C. Pugh and the Regents of the University of
California (1989; cited in Brinckerhoff, Shaw, & McGuire, 1992), the two sides reached
a settlement out of court. A student originally filed the suit after a professor refused to
give him additional time on mathematics exams. The student had adequate
documentation of dyslexia, and the full support of the University’s Disabled Student’s
Program, but the professor still refused the student additional time. The professor
claimed that to allow a student additional time for an exam was unfair to the rest of the
students. The settlement reached out of court required the University to develop a
comprehensive policy for academically accommodating disabled students, and the

professor was required to pay monetary damages to the student. Brinkerhoff, Shaw, and

McGuire (1992) noted that this settlement is important in that it disallows faculty
members the opportunity to dictate to which special services a disabled student is
entitled. Instructors were advised to follow the recommendations of an Academic
Accommodations Policy Board. While there was no court decision, the settlement agrees
with the Wynne v. Tufts University School of Medicine (1990) decision in that changing a
testing format does not, in most cases dramatically compromise the academic integrity or

constitute a major shift in the course content or curriculum. Yet, many questions remain.

The growing numbers of learning disabled students requesting special services
has forced some universities to rethink these policies. School administrators are
increasingly concerned with the large and steadily growing numbers of students
requesting LD services. Boston University, a school with an outstanding reputation for
learning disability services recently began to require that all students requesting LD
services present proof of a recent assessment and diagnosis (Lewin, 1996). Officials
concerned about preserving the academic integrity of educational programs are forcing a
major reevaluation of learning disabilities in the college-aged population. Their
uneasiness and skepticism is due in part to the relatively small amount of research
currently conducted on college students with learning disabilities.

School administrators, such as those at Boston University, recognize that there
are numerous learning disabled college students, and have recently begun to examine the
interventions used for those students. This newfound interest in the topic of learning
disabilities has answered many questions, but raised many more. Some of these

questions have to do with both the ability of individuals to simulate a learning disability,

and the ability of testing professionals to detect this malingering.

The revision of many school learning disability policies may be due in part to the
concern that some students may be simulating or exaggerating a learning deficit in an
effort to access these services. To date, I could find no publications detailing the ways in
which this sort of simulation may be detected.

The majority of college students who are self-referred to educational testing
centers for learning disability assessments are considered eligible for special educational
services (Mellard & Byme, 1993). The prospect of these benefits can motivate some to
fake a cognitive deficit or disorder to receive these services. For example, they may get
to take untimed tests, or to avoid essay exams. In other cases, they may be declared
exempt from taking certain courses (e. g., they may be exempt from fulfilling a foreign
language or math requirement). Most universities have learning disability centers, which
provide tutoring, and other assistance to their learning disabled students. These benefits

may motivate some to simulate the condition.

Le_aming Disabilities

Learning disabilities are found in equal proportions across socioeconomic groups,
different cultures, and are believed to remain relatively persistent throughout the lifespan.
Some sources estimate the prevalence of learning disabilities to be anywhere from two to
twenty percent. The large differences between studies are due to several factors,
including methods and the definition of learning disability used in the study. The
National Committee on Learning Disabilities (1987) estimates that the prevalence of

learning disabilities in this country is between five to ten percent. Obrzut & Boliek

(1991) reported that the prevalence is about ten to fifteen percent. Of learning disabled
students, males outnumber females three to one (McLesky, 1992). In addition,
approximately 15% of LD students exhibit behavior problems (McLesky, 1992).

The proper identification of learning disabled students is a task of the utmost
importance. Students who are not properly identified may experience a host of academic
and professional difficulties that may lead to emotional problems. Kahn (1988) found
that learning disabled college and junior college students exhibited many socio-emotional
difficulties, which have a strong impact on academic performance. Their difficulties
included poor time management, difficulty completing tasks, poor psycho-perceptual
skills, and poor study skills. These problems result in difficulty understanding graphics,
difficulty keeping one’s place on an examination, difficulty with standardized answer
sheets, and word reversals. Some of the emotional problems experienced by learning
disabled students are presumed to be related to their cognitive difficulties. Because of
this, their emotional difficulties do not prevent them from receiving a LD diagnosis.
Conversely, more severe emotional problems are often seen as the underlying factor in
learning disabilities. These factors often require testing prior to the diagnosis of a
learning disability.

The majority of LD students are identified as such during their early elementary
school years. In one study, a sample of 790 students revealed that the identification of
learning disabilities peaked at grade one, abated slightly during grades two and three, and
then gradually declined. Approximately 75% of learning disabled students were
identified as such by the conclusion of grade five. Severe learning disabilities may be

diagnosed earlier, and more subtle learning disabilities are identifiable as the educational

process continues (McLesky, 1992).

The diagnosis of a learning disability is a comprehensive process. Johnson and
Blalock (1987) describe their learning disability assessment for adults as composed of
four stages. In the first stage, the client discusses his/her difficulties, concerns, and
reasons for wanting an assessment. In the second stage, the diagnostician collects a
history of the client and their difficulties. This is an opportunity for the client to describe
their perceptions of his/her problems. Stage three is composed of objective testing, and
in stage four the diagnostician writes down his/her clinical observations of the client
throughout the assessment process.

The Association on Higher Education and Disability (AHEAD- Ad Hoc
Committee on Learning Disabilities, 1996) recommends that any learning disability
diagnosis be established and verified using the following guidelines: Testing performed
by a qualified professional must be comprehensive. Information should be gathered on
academic, social, developmental, and medical history. The psycho-educational or
neuropsychological testing should include aptitude, achievement, and information
processing domains. The testing must also have taken place within the past three years,
should include a specific diagnosis, and provide actual scores from technically sound
tests. In addition, the final report should include specific recommendations and provide

the rationale behind these recommendations.

Learning Disabilities and Neuropsychology

The applicability of neuropsychological assessments for use in the detection of

learning disabilities rests on the assumption that learning disabilities stem fi'om cerebral

10

dysfunction. Indeed, the National Joint Committee for Learning Disabilities defines
learning disabilities as a “generic term that refers to a heterogeneous group of disorders
manifested by significant difficulties in the acquisition and use of listening, speaking,
reading, writing, reasoning or mathematical abilities. These abilities are intrinsic to the
individual and are presumed to be due to central nervous system dysfiinction (Hammill,
Leigh, McNutt, & Larsem, 1981).” Speculation as to the etiology of learning disabilities
is disputed. Explanations range from chemical imbalances, to developmental
abnormalities, to structural damage. With regard to the latter two, the exact type and
location of this cerebral dysfunction is disputed. A review by Bigler (1992) on the
neurobiology and neuropsychology of leaming disorders discusses much of the literature
on these topics. Post-mortem studies on the brains of learning disabled persons have
revealed mixed, but interesting observations. For language based learning disorders, left
hemisphere irregularities are often present, but the location of these irregularities is
widespread and extremely heterogeneous. The same is true of magnetic resonance
imaging (MRI) and electroencephalograph (EEG) studies on the active brains of learning
disabled persons. The MRI and EEG studies on language-based LD subjects reveal
abnormal patterns of activity in language associated regions of the brain (e. g., areas in
and surrounding the left temporal and parietal lobes). However, these same studies also
reveal irregularities in other areas previously thought to be minimally related to language-
based learning disorders (e. g., anterior frontal and occipital regions). While there appears
to be a structural basis to many learning disorders, these connections are far from clear.
One common theme in the neuropsychological study of learning disabilities is that

the disabilities are not thought due to a lack of intelligence. Rather, a learning disability

11

is attributed to difficulties in processing selected kinds of stimuli. For example, dyslexic
individuals can fully comprehend auditory instructions, but have great difficulty with
written text. The information processing difficulty is related to an impairment in
processing written language, not an inability to understand language. For other learning
disabilities, such as mathematical learning, the processing difficulty may stem from
impaired mental manipulation and concentration abilities.

It is this difficulty in information processing that neuropsychological testing
attempts to address. Neuropsychological assessment batteries such'as the Halstead-
Reitan (Reitan & Wolfson, 1993) and Luna-Nebraska (Golden, Purisch, & Hammeke,
1985) have sometimes been used in conjunction with psycho-educational tests to detect
and diagnose specific learning disabilities. Rourke (1982) utilized several
neuropsychological assessments to separate learning disabled children into four
categories according to the results of neuropsychological testing. Obrzut and Boliek
(1991), emphasized that the use of neuropsychological tests for the investigation of
learning disabilities can provide information on sensory acuity, sensory recognition and
perceptual functions, motor functioning, attention, memory, psycholinguistic functioning,
as well as commonly assessed dimensions such as cognitive aptitude and achievement.

Using neuropsychological tests, Felton, Naylor, and Wood (1990) found that,
when IQ and socioeconomic status were controlled, only rapid naming, phonologic
awareness, and non-word reading discriminated significantly between learning disabled
and non-disabled participants. A similar conclusion was reached by Horn, O’Donnell,
and Leicht (1988), who found that when IQ and several demographic variables are

controlled, the assessment differences as measured by neuropsychological tests are

12

restricted to language processes in adults with language-based learning disabilities.
Bigler (1992) argued that because neuropsychological testing provides the most

sensitive and complete methods for assessing brain dysfunction in adults, these same

neuropsychological assessment techniques should be the primary method for identifying

adults with learning disabilities.

The Woodcock Johnson Psycho-Educational Test Battery

In a national survey of practicing school psychologists, Stinnett (1994) found that
the Woodcock-Johnson Psycho-Educational Battery (W J , WJ-R)(Woodcock, 1977,
1989/1990) is among the tests most frequently used for assessing aptitude and
achievement. Wendler and Roid (1984) determined that, in addition to being a popular
test among educational system assessors, the W] is also a cost-effective approach to the
diagnosis and investigation of learning disabilities.

One asset of the W], is its ability to compare a student’s aptitude to the student’s
achievements in four academic areas. If there are significant discrepancies between
aptitude and achievement in reading, mathematics, knowledge, and written language
areas, with achievement being much lower than aptitude, there may be a basis for
diagnosing a learning disability. This method of identifying aptitude-achievement
discrepancies is a key factor in the identification of LD students.

There are many tests available for identifying the aptitude-achievement
discrepancy, but the W] has two features in particular that have made it one of the
nation’s most popular tests: its standardization sample and its age range.

The Woodcock-Johnson has a clear advantage over many other psycho-

13

educational tests in that both subtests of cognitive ability and the tests of achievement
were normed using the same sample. The results from these tests are of the utmost
importance when testing for a learning disability, and the fact that they are directly
comparable gives the WJ a clear advantage for both clinical and research applications. In
addition, the normative sample of the Woodcock-Johnson ranges from the beginning of
the educational process in kindergarten, to the 65+ age group. This test’s usefirlness with
a college-age population has been demonstrated in a number of different studies (Hall,
1984; Morris, 1990;Da1ke, 1988).

McGrew and Pehl (1988) found the W] to be a good predictor of school
achievement. When compared to the WISC-R, the WJ was an equally good or better
predictor of both achievement and ability after a six-year period. The WJ interest
subtests were not predictive of future achievement.

A study by Walsh, Lowenthal, and Thompson (1989) found that the WJ highlights
this aptitude-achievement difference. It should also be noted that while the authors
support the use of the WJ to find this discrepancy, they did have criticisms of the test.
Most notably, over three different comparisons only 68% of the previously identified
learning disabled students were diagnosed as having a learning disability. The authors
used this finding to suggest that the WJ be used in conjunction with other testing methods
to effectively and accurately identify learning disabled students.

One of the most common criticisms of the Woodcock-Johnson tests of cognitive
ability is that there is more of a focus on verbal than on performance intelligence abilities.
In fact, Woodcock (1978) reported that the Cognitive Abilities scale of the WJ correlated

higher with the verbal section of the WISC-R than with the performance section of the

14

WISC-R. This discrepancy is one of the reasons why the Wechsler intelligence tests are
so often paired with the Woodcock-Johnson Psycho-Educational Battery.

A study by Buchanan and Wolf (1986) utilized the WJ in conjunction with the
WAIS-R to examine the learning and cognitive characteristics of 33 adults referred for
leaming disability assessments. They found that many of the typical learning problems
found in childhood persist into adulthood. One of the main differences is that children
identified in schools as having learning problems may be spared some of the
psychological distress that comes with not receiving professional assistance. Many of the
adults in the study were never diagnosed with a learning disability, and instead saw
themselves as poorly motivated, easily distracted, emotionally labile, lacking self-
concept, and generally unorganized. The study also suggested that using only the verbal
and performance discrepancies in the WAIS-R was insufficient for detecting learning
disabilities.

Merrell (1990) found that the WJ effectively discriminated between learning
disabled and low achieving students. Learning disabled and low achieving students’
subtest scores differ significantly in all subtests but reasoning. In this study, the
participants were given the cognitive abilities and achievement tests. On the cognitive
aptitude tests, the two groups differed most on the cognitive and memory scales. Among
the achievement tests, the two groups showed the most dramatic differences on the skills
and reading tests.

Dalke (1988) examined 72 first-year college students, half with and half without a
diagnosed learning disability, to determine if the WJ was useful in identifying learning

disabled college students. Dalke emphasized the consistency of learning difficulties for

15

the learning disabled group throughout the lifespan at a level that remains consistently
lower than those of the control peer group. This study found significant differences
between LD and non-learning disabled students on all clusters relating to cognitive
abilities and academic achievement. There were no significant differences found
between the LD and control groups for the interest measurements in the WJ.

To investigate the link between head injuries and learning difficulties, Tupper
(1990) conducted one of the first studies utilizing the WJ with a head-injured population.
In this study, 39 adults with a mean coma duration of 9.8 days were assessed upon entry
into an outpatient treatment program. The Trail Making Tests, the Halstead Category
Test, and the W] were administered to all of the subjects. The Trail Making Test A
correlated negatively and significantly with reasoning, perceptual speed, and overall
cognitive clusters on the cognitive ability cluster of the WJ. Trails B correlated
negatively and significantly with the verbal, reasoning, perceptual speed, and memory
subscales of the cognitive ability cluster, as well as overall cognitive ability. The WJ
tests in this study revealed several similarities between head injured and learning disabled
individuals. General attentional, information processing, and perceptual speed deficits
are among the most common similarities between these two groups. Tupper describes
these difficulties as problems with visual perceptual speed and new learning deficits.
Tupper’s description of these problems is interesting in that he never directly addresses
learning disabilities, although his description of head injured patients is strikingly similar
to the difficulties experienced by persons with many kinds of learning disabilities.

In another effort to document the psycho-educational effects of brain lesions,

Aram and Ekehnan (1988) tested brain-injured children using the WJ. These children

16

reported several of the same educational difficulties as LD students, such as spelling,
attentional, spatial perceptual, and memory difficulties. The WJ test results confirmed
the children’s complaints.

While the Woodcock-Johnson is currently among the most popular assessment
tools used for the evaluation of learning characteristics, it is not without criticism. One
important criticism of this test battery is that it may not adequately differentiate between
low achieving and learning disabled students (Gajar, Salvia, Gajria, & Salvia 1989;
Olivarez, Palmer, & Guillemard, 1992; Warner, Schumaker, Alley, & Deshler, 1980).
This is troublesome, given that this battery is commonly used to differentiate between
these two groups.

In addition, it is important to note that the constructs of aptitude and achievement
are not entirely separate. Aptitude influences achievement. Because of this, achievement
may be influenced by the same factors that influence aptitude test scores, such as
socioeconomic status, ethnicity, and access to resources (Kaufman, 1990). Individuals
may, because of environmental factors, evidence aptitude and achievement scores which
are not reflections of their true abilities. Therefore, conclusions from testing should be
tempered in light of potential environmental influences. Still, despite these criticisms, the
Woodcock-Johnson remains a very popular tool in the assessment of learning

characteristics.

Neuropsychologic_al Effortfis to Detect Malingering

While batteries of neuropsychological tests may assist in identifying differences

between intact and damaged cognitive functioning, their utility for detecting faking has

17

not been established (Lezak, 1995). Most individuals with some degree of impairment
are unique in the exact kind of dysfunction that they display, and this makes it difficult
for practitioners to discern between idiosyncratic and fraudulent test results.
Inconsistencies in test performance are often used to detect malingering. For example, a
patient may perform well on the WAIS-R subtests of comprehension, similarities, and
arithmetic (Wechsler, 1981), but score very low in its vocabulary and digit span subtests.
Studies utilizing the Halstead-Reitan Neuropsychological Test Battery’s (HRNB)
numerous subtests have come to differing conclusions regarding the utility of this method
(Rogers, 1997). Many authors agree that the use of discriminant function analyses to
identify inconsistent performance is far better than a qualitative evaluation of the testing
data. One broad conclusion fi'om several of these studies is that normal individuals
instructed to fake neuropsychological deficits produce more measurable deficits in the
areas of sensory-perceptual and motor functioning, as well as on measures of auditory
attention and discrimination. Compare this to individuals with verifiable brain injury,
who generally display greater impairment on indices of abstract reasoning and set-
shifting (Doerr & Carlin, 1991; Heaton, Smith, Lehman, & Vogt, 1978; Goebel, 1993;
Rogers, 1997). A review of the literature (Nies & Sweet, 1994) concluded that detecting
malingering in this way is very difficult, although sometimes possible.

A common criticism of research on neuropsychological assessment malingering is
that there are no published studies on verifiable malingerers. Individuals rarely admit to
malingering, and as such the “malingerers” used in controlled studies are presumed to be

equivalent to actual malingerers. The validity of this assumption is not known.

18

Dissimulation Measures:

Several attempts to create scales specifically to detect malingering have been
based on a number of findings in studies comparing malingerers and non-malingerers.
However, these scales have not been applied to other areas of testing. While the use of
certain tests such as the WAIS-R (W echsler, 1981) and/or the Woodcock-Johnson
(Woodcock & Johnson, 197 7) are commonly used in both neuropsychological and
leaming disability test batteries, I have found no published study that tested the
applicability of specific neuropsychological dissimulation measures to the learning
disability testing situation.

The Rey Memorization of Fifteen-Item Test (Rey, 1964) requires participants to
memorize fifteen different items, and then recreate the items following a short delay.
This test presents itself as being very difficult, when in fact it is quite easy. Because most
brain damaged patients can remember the majority of the items (Bernard & Fowler, 1990;
Paul et a1., 1992; Goldberg & Miller, 1986), anyone failing to recall these items can be
suspected of malingering, especially if there is contradictory information (e. g.,
remembering to get to the assessment appointment on time, remembering the date, etc.).
This task can be presented using either a 15-item memorization task, or a l6-item
memorization task (Paul, Franzen, Cohen, & F remouw, 1992). When attempting to
dissimulate on a test, naive malingerers tend to overestimate the deficits of memory
impaired individuals. Bernard (1990) found that naive dissirnulators performed worse
than persons with verified brain injury when testing the ability of the Rey Memorization
of Fifteen-Item Test to discriminate between malingerers and non-malingerers,.

Another neuropsychological test used to detect malingering is Rey’s (1941) Dot

l9

Counting tasks, which asks the patient to count various numbers of dots. In the first trial,
the practitioner asks the patient to count dots randomly placed on a blank index card.
The more dots present on the stimulus cards, the longer the patient should need to count
them. An only slightly more complicated version of this test adds an additional trial of
dot counting. In the second trial, the numbers of dots and the order of these numbers
remain essentially the same, but on this second trial the dots are grouped (e.g., squares,
diamonds, etc.). Times for ungrouped dots should be slower than the times for counting
the grouped dots. If either the error rate or time for counting the grouped dots exceeds
that for ungrouped stimulus dots, the patient’s cooperation is suspect.

Hiscock and Hiscock (1989) developed another malingering test, which requires
the patient to memorize a five digit number displayed on a card, and the patient is later
asked to identify this number on a card containing two five-di git numbers. The test is
given in three different trials, each with longer delays. Between trials, the clinician
remarks to the person taking the test that he/she did a good job, and that the test will be
made somewhat harder by increasing the delay period (Prigatano & Amin, 1993).
Because there is no evidence that these delays actually increase the error rate (Lezak,
1995), such increases raise the possibility of dissimulation. Indeed, a large increase
suggests that the patient is not cooperating.

On such forced choice tests, those who have sustained severe brain damage may
not correctly answer all of the questions. A patient taking the test with impaired memory
ability should, at worst, get half of the questions wrong. A subject under examination
that gets less than half of the questions wrong may be due to chance, but if this pattern

persists over multiple administrations, it might be inferred that the patient is intentionally

20

picking the wrong answers.

The Rey Memorization of Fifteen-Item Test and the F orced-Choice test both
address short-term memory, which has also been implicated in mathematics learning
disabilities. Cohen (1983) described mathematics learning disabilities as “short-term
memory difficulties,” where memory of arbitrary symbols is needed but not available.
Individuals with short-term memory difficulties may indeed perform worse on the
Fifteen-Item Test than normal subjects. However, the Fifteen Item Test’s utility lies in
that the suggested cutoff score for determining malingering norms are sufficiently low as

to rule out a majority of false positives (Millis & Kler, 1995).

Hmotheses:

The current analogue study explores the relationship between the performance on
measures designed to assess learning characteristics and dissimulation measures. The
impact of instructional sets is an additional feature. Participants were asked to simulate a
learning disability as well as asked to take a variety of cognitive and achievement

measures. The following hypotheses were investigated:

1. It is extremely difficult to simulate a learning disability, and individuals who are
instructed to dissimulate on an assessment of learning characteristics will not be able
to demonstrate the required aptitude-achievement discrepancy. When presented with
this task, participants asked to simulate a learning disability will score lower (i.e., at a
statistically significant .05 level) on:

A. Subtests of the Woodcock-Johnson Revised:

21

a. Cognitive measures: Numbers Reversed, Picture Vocabulary, and
Visual Matching,
b. Achievement measures: Word Attack, Applied Problems and
Quantitative Concepts;
B. The AMNART;
C. Specific tests of dissimulation:
a. Rey Fifteen-Item Test,
b. Rey Dot Counting Tests, grouped and ungrouped, and
c. Forced-Choice digit recognition test (Hiscock & Hiscock, 1989)

in comparison to participants asked to do their best on these tests.

2. The task of faking a learning disability is a difficult task, because it requires the
recognition of different tests, as well as a specific pattern in the results. Individuals will
not be able to show a pattern of test results similar to that of a “learning disability.” In
this study, participants will not show a discrepancy between academic achievement and
intellectual ability. That is, participants will not differ by more than one standard
deviation in their estimated cognitive ability standard scores and achievement standard
scores.

3. Tests designed to detect malingering are more effective when presented at the
beginning of a test battery. On an exploratory basis, it was hypothesized that an order
effect exists. Participants who are asked to simulate a learning disability first and are
then asked to do their best will perform differently than will participants who are asked to

do their best and then asked to simulate a learning disability.

22

It is important to note that the order of test administration may impact the
dissimulation tests in this study. Amett, Hammeke, and Schwartz (1995) suggested that
simple tests such as the ones used in this study which were designed to detect
malingering may be most useful when positioned at the begimring of a series of tests.
These dissimulation sensitive tasks perform their function in part due to a ceiling effect.
Individuals undergoing psychological testing may be more sensitive to the easy nature of
these dissimulation tasks once they have completed more difficult tasks. Dissimulation
measures are useful in that they do not measure a full range of abilities, but unfortunately
an individual with experience in any number of psychological tests may be sensitive to
these dissimulation tests with increasing exposure to other tasks.

For this reason, it is important to note that the test order may make a difference in
the results of this study. Those initially instructed to “perform their best” may be
sensitized to the difficulty of the several tests and alter their testing performance to
appear more believable. More specifically, they will perform more like those in the
control condition; they will perform “better” on these tests and therefore be more difficult

to detect as malingerers.

Methods:

Cognitive Measure

To provide an estimate of intellectual ability, all participants will be given the
American New Adult Reading Test (AMNART) (Schwartz & Saffran, 1987; cited in
Grober & Sliwinski, 1991). This is a modification of Nelson’s (1982) Adult Reading

Test (NART), based on the hypothesis that intelligence can be estimated by examining

23

highly overlearned skills which are relatively resistant to brain damage and which
correlate highly with IQ. This test scores a patient on his/her correct pronunciation of a
list of progressively more difficult words. Because the NART was normed on a British
sample, some words that appear on the test may appear more or less regularly for
American users. To remedy, the AMNART was developed using the same principles as
the NART, but substituted words more commonly used by North Americans for some of
the original test items. The AMNART correlates substantially with verbal and full-scale
WAIS IQ measures (.72), and moderately with performance IQ (.51) (Schwartz and
Saffran, 1987; cited in Grober & Sliwinski, 1991). Grober and Sliwinski (1991) report

that the test-retest reliability for this test is a high .93.

Procedure

For this study, 40 undergraduate participants were assigned to both the
malingering and non-malingering instructions. The participants were individually
assessed on two occasions by different clinicians, with at least one week between testing
sessions. Participants were initially randomly assigned (via a coin flip) to either the
malingering or non-malingering conditions.

Because the presence of those with bonafide leaming disabilities could confound
the results, all will be asked the following questions:
1. Have you ever needed a tutor to help you with your schoolwork?
2. Have you ever been placed in a remedial class?
3. Have you ever been diagnosed with a learning disability?

The answer of “yes” to any of these questions will disqualify a subject from participation

24

in the study.

The participants received written instructions prior to the assessment describing
their task. Non-malingerers were asked to give their “best performance” on the tests.
Those instructed to “malinger” were told to respond as if they wanted to receive the
diagnosis of a “learning disability” in order to obtain special academic privileges.

The following instructions (adapted from Amett, Hammeke, & Schwartz, 1995)
were given to individuals before testing in the dissimulation condition:

“You have been having trouble in school and have decided to
undergo testing for a learning disability. These tests are very important
because they will in part determine if and how much assistance you will
receive in your classes. It is possible that if your leanring disability is
severe enough, you will be exempted fiom some course requirements.

You want to cooperate fully with the tester, but at the same time you want

to make sure the tests indicate a learning disability so you can receive as

much acaderrric assistance as possible. Be careful not to appear too

obvious in your attempts, for if you are suspected of faking, your requests

for academic assistance may be dismissed.

“Your instructions are to fake a learning disability. This is a difficult task,
so in order for you to be the best possible faker here is some information about
learning disabilities and learning disability testing you should find helpfirl: A
learning disability is a disorder in one or more of the basic psychological
processes involved in understanding or in using spoken or written language.

People with learning disabilities can have difficulties performing tasks such

25

listening, thinking, speaking, reading, writing, spelling, or doing mathematical
calculations. Learning disability diagnoses are not given to people who have
learning problems which are mostly due to visual problems, hearing problems, or
motor handicaps, mental retardation, emotional disturbance, or environmental,
cultural, or economic disadvantage. A learning disability is said to be present
when there is a large enough difference between someone’s intellectual ability
and his/her academic achievement.

“So in other words, to be diagnosed with a learning disability, you should
try to make the tests show a large difference between your intellectual ability, and
your academic achievement. Your achievement should be much lower than your
intellectual ability.

Those are all the instructions you get. The rest is up to you. Good luck!”

Participants were administered the Quantitative Concepts, Word Attack, Applied

Problems, Picture Vocabulary, Numbers Reversed, and Visual Matching subscales from

the Woodcock-Johnson psycho-educational test; the AMNART, the Rey F ifteen-Item

Memorization test, the Rey Dot Counting (grouped and ungrouped) task, and the Forced-

Choice test.

After the dissimulation condition, the participants were asked to rate their efforts

on a 5-point Likert scale. The participants were asked to rate the effort they put into their

learning disability “performance” (5=tried very hard, 1=did not try at all). The

participants were also asked if they understood the dissimulation instructions (‘yes” or

26

The results were coded and submitted for statistical analysis by a repeated
measures AN OVA. Potential order effects were examined in each of the hypotheses

using this analysis.

BESBIJLS;

Forty participants took part in this study. Twenty females and twenty males
successfirlly completed two different testing sessions. Data analyses were divided into
three sections. In the first, the hypotheses conceming aptitude and achievement
differences were evaluated. In the second part of the analyses, hypotheses regarding the

tests of malingering and their effectiveness were examined.

Aptitude and Achievement and LeamirLg Disability Fak_hrg;

This study hypothesized that participants' scores on the tests of achievement and
aptitude would be lower in the malingering condition than they would be in the control
condition. However, this study also hypothesized that participants would not be able to
fake a learning disability, as defined by a one standard-deviation difference between
aptitude and achievement tests, with tests of achievement being one standard-deviation
lower than tests of aptitude.

For these analyses, all scores were converted to z-scores, based on the age
adjusted norms in the WJ-R manual. As predicted, participants’ scores in the
malingering condition (M=—.83)were judged to be more deficient than those in the control
condition (M=.42). The results of a repeated measures AN OVA indicated that this is a

statistically significant difference, E(l, 38)=60.08, p<.001.

27

Interestingly, participants were able to show a discrepancy between aptitude and

achievement (see table 1).

(Table 1) Mean z-scores for the tests of cognitive ability and the tests of achievement

 

 

 

 

AVERAGE OF Z-SCORES Aptitude tests Achievement tests
Malingering condition -1 . 12 -.54
Do your best condition .44 .41

 

 

 

The results of a repeated measures AN OVA indicated that this is a significant interaction,
E(1, 38)=17.84, p<001.

It is important to note that while participants were able to show a significant
difference between tests of aptitude and tests of cognitive ability, the difference was on
the order of approximately one-half standard-deviation in the opposite direction (see
figure 1). This difference, while significant, was less than one standard-deviation. Thus,
while the null-hypothesis that participants can not show a difference between tests of
achievement and cognitive ability is rejected, this sample did not show a large enough
difference between tests of achievement and tests of cognitive ability to indicate that
participants could fake a learning disability.

Also of note is that no significant order effects were observed in this analysis, E( 1,

38)=.487, p=.489.

28

 

(Table 2) Descriptive Statistics for the tests from the cognitive portion of the WJ—R,

 

 

 

 

 

irrespective of order

TEST (COGNITIVE): Control Malingering
Mean, SD

Numbers Reversed: .39, .96 -.83, .99
Visual Matching .62, .80 -1.93, 1.77
Picture Vocabulary .31, .66 -.61, 1.33

 

 

 

(Table 3) Descriptive statistics for the tests from the achievement portion of the WJ-R,

irrespective of order

 

 

 

 

 

TEST (ACHIEVMENT) Control Malingering
Mean, SD

Applied Problems .19, .66 -.67, 1.45
Quantitative Concepts .44, .67 -.63, 1.41
Word Attack .59, .71 -.31, .77

 

 

 

As a group, this sample of participants was unable to Show a cognitive test —

achievement test split with tests of achievement being lower than cognitive tests. Even

individually, there were no individuals out of the forty who were able to show a

discrepancy where an average of the tests of achievement was one or more standard

deviations lower than an average of the tests of cognitive ability.

29

 

 

A repeated measures ANOVA revealed that on the AMNART, participants’
scores in the malingering and control condition were significantly discrepant, resulting in
significantly different estimated verbal IQ scores E(l, 38)=17 .37, p<.001 (table 4). There

was no order effect observed for this measure, E(l , 38)=2.32, p=.633.

(Table 4) Descriptive statistics for the All/INART, irrespective of order

 

AMNART: mean, SD Control Malingering

 

 

ANINART 113.41, 5.26 108.98, 6.59

 

 

 

Tests of Malingering:

This study hypothesized that tests of malingering are sensitive to faking on
learning disability assessments. The results of the following statistical analyses would
infer that the null hypothesis can be rejected. Participants scored lower on tests of
malingering when asked to fake a learning disability than they did when asked to perform
to the best of their ability. In addition, this study hypothesized that tests of malingering
are more sensitive when placed at the beginning of a battery of tests, and thus better able
to detect malingering. The results of statistical analyses on the three tests of malingering

are as follows:

Rey Fifteen Item Test (FIT):
After the Rey F ifteen-item test stimulus card was presented, the subject was asked

to reproduce the stimuli on the card in the same fashion they were presented. The test

30

 

was scored by counting the errors the participant made when reproducing the card.

Errors on the Rey FIT were measured in two ways, the first counted the number of errors
the subjects made. This approach counts the number of symbols that were not in their
proper place. For example, many subjects had all of the symbols present on the paper,
but they reversed the order of the rows; if the original stimulus was 123, many
participants reproduced the stimulus as 321. The results of a repeated measures AN OVA
showed that this test does indeed discriminate between malingerers and controls when the
score on the FIT is measured as a function of how many symbols were not correctly
placed in their proper row and column, E(1, 38)=68.64, p<.001. There was no order
effect observed on this approach to scoring this test, E (1, 38)=2.l4, p_=.152.

The second approach to scoring the FIT counted the number of symbols the
subject remembered and placed on the paper, regardless of the placement of these
symbols. For example, if the original stimulus row was 123, and the subject responded
321 , the subject would receive credit for three correct responses. The results of a
repeated measures ANOVA showed that while this method of scoring the FIT was not
was effective as counting the number of errors, it still produces a significant effect, 2(1,

38)=13.35, p=.001. There was no order effect observed on this test _P_‘(l, 38):.10, p=.750.

(Table 5) Descriptive statistics for the Rey 15-item test, irrespective of order

 

 

 

 

Rey-FIT: Mean, SD Control Malingering
Number of items recalled 14.53, 1.20 13.10, 2.21
Number of items absent .73, 1.28 6.80, 4.71

and/or in the wrong position

 

 

 

31

 

Forced Choice Test (FCTX

The Forced Choice Test was scored as the total number of errors made across all

three trials individually, as well as the sum total of errors across all three trials. A

repeated measures AN OVA revealed that this test discriminated between subjects in the

malingering and control section in Trial A, E(1, 38)=18.46, p<.001; Trial B, _F_ ( 1,

38)=38.38, p<.001; Trial C, E(1, 38)=41.08, p<.001. The PCT was particularly powerful

across all three trials, E(l, 38)=35.09, p<.001. There was no order effect observed on

Trial A, an, 38)=.54, 9:467; Trial B, E0, 38)=.47, 9:.495; Trial 0, E0, 38)=1.19,

p=.283; or the sum total of errors across all three trials, E(l, 38)=1.08, p=.306.

(Table 6) Descriptive statistics for the Forced Choice Test, irrespective of order

 

 

 

 

 

 

Forced Choice Test (FCT): Control Malingering
Mean, SD

Errors, Trial A 0.00, 0.00 2.78, 4.06
Errors, Trial B 0.00, 0.00 4.50, 4.56
Errors, Trial C .03, .16 5.03, 4.92
Total Errors across all 3 .03, .16 12.03, 12.80

trials

 

 

 

Rev Dot Counting Test (DCle

The Dot Counting Test was administered as a series of individual cards, with a

32

 

number of dots presented on each card. The participants were instructed to count the dots
as quickly as possible. For each card, both the participant’s response (the number of dots
they counted) and the amount of time it took them to indicate their response was
recorded. The Dot Counting Test was scored several ways in an attempt to determine
which way is the most powerful discriminator between conditions. A polynomial
contrast revealed that as Rey hypothesized, there is a strong linear trend in the time it
takes to count the number of dots on each card. The amount of time it took participants
to count the number of dots increased in a linear function as the amount of dots increased.
This effect is true for both the ungrouped dots portion of the test E(l, 39)=394.11,
p<.001, as well as the grouped dots portion of the test E( 1, 39)=164.73, p_<.001.

The DCT was scored by whether or not participants increased the amount of time
it took to count the dots. On both the grouped and ungrouped conditions, subjects were
recorded as displaying an increase if the amount of time it took them to count the dots
from card one through to card six time increased or stayed the same. For example, if a
participant’s response times in seconds for counting the dots on cards one through six
were 2, 3, 5, 5, 5, and 9, respectively, the subject was recorded as displaying a
consistency in response time across the six trials. By contrast, if a subject’s response
times were 1, 3, 4, 3, 6, and 8, the subject was recorded as not displaying consistency in
response times across the six trials. Results of a repeated measures ANOVA indicated
that though there is a strong linear trend in the amount of time it takes to count the dots
on the cards, there were enough participants who displayed non-linear patterns in their
response time as to make the difference between malingering and control groups non-

significant in the ungrouped dots portion of the test E(1 , 38)=2.99, p=.092, as well as the

33

grouped dots portion of the DCT, E (1, 38):.81, p_=.374. In this case, whether or not
participants displayed a consistent increase in the amount of time it took to count the dots
as the number of dots increased was not predictive of group membership in the
malingering or control condition.

By contrast, if the DCT is scored by the number of errors in counting, the DCT
becomes a significant discriminator. There is a significant difference in the number of
errors between the malingering and control groups on the grouped dots section of the
DCT, _E(l, 38)=6.42, p=.016. In addition, there is a significant difference in the number
of errors made when one compares the malingering and control groups on the ungrouped
dots section of the DCT, 5(1, 38)=27.80, p<.001. Interestingly, there is an interaction
when looking at the order effect when counting the number of errors on the ungrouped
dots, E( 1, 38)=8.83, p_=.005. Participants made almost twice as many errors on the
ungrouped section of the DCT if they were in the “malingering first” condition (M=3.00,
S_D=l.84), than if they were in the “do your best first” condition (M=1.65, §Q=.99).
When the errors across both trials were added and used as a single variable, there was a
significant main effect, E(l , 38)=19.49, p<.001, and interaction here as well E( 1,
38)=19.49, p_=.016. Again, participants in the “malingering first” condition made more
errors CM=4.40, S_D=3.09) than did participants in the “do your best first” condition

(M.=2.50, S_D=l.82).

34

(Table 7) Descriptive statistics for the Dot Counting Test, irrespective of order

 

 

 

 

 

Dot Counting Test: Control Malingering
Mean, SD

Ungrouped dot counting .95, 1.22 2.33, 1.61
errors

Grouped dot counting errors .40, .90 1.10, 1.43
Total dot counting errors 1.35, 1.88 3.45, 2.68

 

 

 

(Table 8) Order effect for Dot Counting Test, errors in the ungrouped condition

 

 

 

 

Means and Standard Malingering First Malingering Second
Deviations

Ungrouped dot counting 3.00; 1.39 1.65; .99

errors, Malingering (Mean;

SD)

Ungrouped dot counting .85; 1.04 1.05; 1.40

errors, Control (Mean; SD)

 

 

 

Sex Effects:

There were no sex effects observed for any of the measurements.

35

 

 

Post-Hoc Analysis:

An effort was made to determine if the participants could be grouped into the
malingering and control conditions on the basis of their performance on the tests of
malingering. To evaluate this, a discriminant function analysis was used. The results of
this indicated that these participants could indeed be classified into their respective
conditions on the basis of their scores from the malingering tests, Wilks’s lambda = .49,

x2 (5, _N_=80) = 54.47, p<.001. 86.3% of the cases were correctly classified.

Discussion:

Hypothesis number one predicted that it is extremely difficult to fake a learning
disability, and that participants in this study would not be able to simulate a learning
disability. The results of the more specific aspects of this hypothesis are as follows:

Hypotheses 1A and 1B predicted that if individuals are asked to fake a learning
disability, they will score lower on tests of aptitude and achievement than they would
have scored if they were giving their “best” effort. The data supported this hypothesis.
Participants in this study performed worse on tests of aptitude and achievement when
asked to fake a leaming disability than they did when asked to do “their best.” Data
analysis showed that AMNART scores were sensitive to this manipulation as well.
Participants scored lower on the AMN ART in the malingering condition than they did in
the control condition; participants’ estimated Verbal IQs were lower in the malingering
condition than they were in the control condition.

Hypothesis 1C predicted that neuropsychological tests of malingering are

36

sensitive to faking on tests of leanring characteristics. Participants performed worse on
each of the tests of malingering when asked to fake a learning disability than they did
when asked to perform at their “best.” These instruments are sensitive enough to detect
this form of malingering. A more specific breakdown of the three malingering tests are
as follows:

The Rey-Fifteen Item Test (FIT) is a sensitive detector of malingering, whether it
is scored as a function of how many items are present, or how many items are present and
in their correct position. The FIT is a more sensitive indicator of malingering when it is
scored by how many items are present and are in their correct position.

The Forced Choice Test (F CT) is also a sensitive malingering instrument when it
comes to detecting if individuals are trying to fake a learning disability. The F CT can be
scored according to the number of errors in each individual trial, or the sum total of errors
across trials. Each of these methods were useful in discriminating between malingerers
and individuals who are doing their best, though it appeared that the most powerful
indicator on this test is the number of errors on the third trial, or trial “C.” The finding
that individuals showed the biggest difference in FCT trial “C” scores between the
malingering and control conditions supports the test’s implicit assumption that increased
errors are an indication of malingering. Successive trials “A,” “B,” and “C” are not
progressively more difficult, and should not result in increasing amounts of errors,
despite the test administrator’s instructions which state that the test becomes more
challenging.

Results from the analyses on the Rey Dot Counting Test revealed that this test,

too, was a sensitive indicator of malingering on tests of learning disabilities. However,

37

this test did not appear to function as was originally hypothesized. There is a very strong
linear function in this test - that is, the amount of time it took participants to count dots
increased linearly as the amount of dots increased linearly. This was a strong trend
regardless of whether or not people were malingering or not. However, there was no
reliable or significant difference between malingerers and controls based on whether or
not the time it took for them to count the dots increased consistently. Because it was not
possible to statistically differentiate between the malingering and control condition based
on the progressive linear nature of response time, the DCT should not be used as a
measure of whether individuals are malingering based upon if their scores are linear.
When the DCT is scored based upon the number of errors individuals make when
counting the number of dots present, it becomes a useful test for discriminating between
malingerers and individuals trying to do their best. When using this test as part of a
battery to assess learning characteristics, individuals who are trying to malinger make
significantly more errors than do individuals who are trying to do their best. The number
of errors in the grouped portion of the DTC, the ungrouped portion of the DCT, as well as
the cumulative number of errors from the grouped and ungrouped conditions were
significantly higher in individuals who were trying to fake a learning disability. Of these
three scores, the number of errors in the ungrouped section was the most useful indicator

of malingering.

Hypothesis 2 predicted that the task of producing test results with a more than one

standard deviation difference between aptitude and achievement, with achievement being

lower than aptitude, would be too difficult to complete. This hypothesis was supported

38

by the data. In most states, the difference between aptitude and achievement required for
the diagnosis of a learning disability is one to one and a half standard deviations. In this
study, while participants were able to show a statistically significant difference between
aptitude and achievement, this difference was approximately .5 standard deviations in the
opposite direction. The participants showed a tendency to score lower on tests of
aptitude than tests of achievement.

It is important to note that this study did not attempt to answer the question of
whether or not it is possible to fake a specific learning disability (i.e., Reading Disorder
or Mathematics Disorder). This was not possible given the limited amount of time in
which to conduct the assessments and thus the limitations on the kinds of tests given.
Rather than investigate whether or not it was possible to fake a specific learning
disability, this study attempted to answer the more general question of whether or not it
was possible for participants to detect the difference between tests of aptitude and tests of
achievement without having these identified beforehand. These results indicated that

9“

these participants strategy” was the opposite of what it should have been.

It is important to note that the .5 standard deviation difference between aptitude
and achievement was a mean difference, and the range of scores was variable. Some
individuals displayed a pattern of test results where achievement was less than .5 standard
deviations lower than aptitude; they were less successful at showing a "learning
disability.” Others displayed a pattern of test results where achievement was greater than
.5 standard deviations lower than aptitude; they were more successful at showing a
"learning disability. "

Participants in this study were unable to fake a learning disability. Even though

39

they were instructed on how they might do so, they were still unable to show this pattern
of scores.

Hypothesis 3 predicted that the tests of malingering would be differentially
effective depending on an individual’s prior exposure to these and the other tests of
aptitude and achievement. Interestingly, with one exception, there does not appear to be
an order effect to these three tests of malingering. Though it may make intuitive sense
that by exposing individuals undergoing testing to “real” tests, it will sensitize them to
the notion that tests of malingering seem “easy”; this does not appear to be the case.
With the exception of the number of errors made on the ungrouped section of the DCT,
there was no order effect exhibited anywhere in the testing. A possible explanation for
this finding is that the participants’ responses to the ungrouped portion of the DCT was
the final part of the portion of the test battery devoted to tests of malingering. It is
possible that participants had, by that point in the test battery, gained enough experience
with the malingering tests that the DCT became more transparent. With this single
exception, participants do not score differently on malingering tests if they have had
exposure to other tests. This result is contrary to the suggestion from Amett, Hammeke,
and Schwartz (1995) that tests of malingering may be more useful at the beginning of a
test battery.

Consequently, the results of this study support the conclusion that tests commonly
used to detect faking in neuropsychological assessment are useful in assessments of
learning characteristics. They may be helpful for detecting a poor attempt at learning
disability dissimulation. Though no participants in this study were able to show a

sufficiently large discrepancy between aptitude tests and tests of achievement, the tests of

40

malingering are helpful for detecting an attempt at deception, and may prove helpful for
more sophisticated malingerers. In addition, these tests are useful throughout a testing
battery.

There has been some discussion in the literature that many malingering studies are
flawed by virtue of their participants. For example, many published studies on the ability
of individuals to malinger neuropsychological deficits have been criticized because the
samples from these studies come from a population of college students. A sample of
college students does not accurately reflect the population of individuals requesting
neuropsychological assessment for cognitive difficulties, and these criticisms are valid in
this respect. This study’s sample, however, comes from the population it seeks to
understand. College students may be more or less sophisticated dissirnulators than other
populations when it comes to simulating cognitive deficits. However, when investigating
the likelihood that college students can fake a learning disability, it is best to draw a
sample from a population of college students. This study’s findings apply directly to the

population from which it was drawn.

On a more qualitative note, test administrators noted that most individuals in the
malingering condition had little or no perceptible strategy. The few that displayed
perceivable strategies were grouped into two general categories: the "wrong answerers"
and the "reversers."

"Wrong answerers" appeared to go out of their way to give an incorrect response.
For example, when shown a picture of a fork, they might have said "utensil," and thus

gotten the question wrong. Errors such as this were not uncommon, and displayed a

41

tendency of these participants to give a wrong response at all costs, even if an incorrect
response required more sophisticated knowledge than the correct response they were
trying not to answer. For example, when shown a picture of a fork and asked to name the
object, he/she may have responded with “utensil” and claimed no firrther understanding
of the picture

"Reversers" usually reversed phonemes and letters on the AMN ART and Word
Attack subtest of the WJ-R. In addition, they often reversed numbers on the Visual
Matching and Numbers Reversed portions of the WJ-R. These individuals appear to
subscribe to the common view that learning disabilities primarily consist of difficulties in
reading and writing that stem from reversing and switching letters on a regular basis.
This strategy appeared to be somewhat analogous to the popular, though somewhat
misguided perception of dyslexia as a disorder primarily involving letter and number
switching.

It is important to note that these two groups were derived from the assessors'
impressions of the participants’ strategies. These were clinical impressions and were not

included in the data analysis.

Future Research:

While this study has answered some very interesting questions as to the nature of
learning disability malingering as well as the application of neuropsychological tests of
malingering to leaming characteristic assessments, it is by no means a comprehensive
study. Future research on this issue may pursue the two general areas of “tests of

malingering” and “aptitude and achievement tests.”

42

Future research on tests of malingering may involve the adaptation of existing
tests of malingering to a learning characteristics assessment. For example, a forced
choice test might be created to investigate spelling or reading. A test such as this may
present the participant with two choices, and the participant would pick the correct
spelling, or the correct mathematics answer. Another potential research question would
be to investigate the popular conception of learning disabilities, such as dyslexia, as
disorders of word and number reversals. A test might be constructed that would exploit
this popular rrrisconception. An additional research idea would be to provide participants
with more explicit training on how to fake a learning disability.

Future research on tests of aptitude and achievement might pursue the
“fakeability” of different tests commonly used in learning characteristic assessments. In
this study, the Word Attack and Visual Matching subtests were differentially sensitive to
the malingering instructions. Participants’ z-scores were lower on the Visual Matching
subtest than they did on the Word Attack subtest during the malingering condition. A
very time-consuming but potentially comprehensive study would involve the
administration of an entire learning characteristics assessment to malingerers and non-

malingerers.

43

References

Ad Hoc Committee on Learning Disabilities, Association on Higher Education
and Disability (1996). Guidelines for documentation of a specific learning disability.
Latest Developments: A Newsletter of the Learning Disabilities Special Interest Group.
Summer, 10-12.

Americans with Disability Act. (1992) Public Law 94-142, U.S.C. 2101 et seq.

Aram, D. M. & Ekelman, B. L. (1988). Scholastic aptitude and achievement
among children with unilateral brain lesions. Neuropsychologia, E, 903-916.

Amett, P. A., Hammeke, T. A., & Schwartz, L. (1995). Quantitative and
qualitative performance on Rey’s 15-Item Test in neurological patients and dissirnulators.
The Clinical Neuropsychologist, (2), 17-26.

Austin, J. S. (1992). The detection of fake good and fake bad on the MMPl-Z.
Educational and Psychological Measurement, fl, 669-674.

Bernard, L. C. (1990). Prospects for faking believable memory deficits on
neuropsychological tests and the use of incentives in simulation research. Journal of
ClinicaL and Experimental Neuropsychology, Q, 715-728.

Bernard, L. C. & Fowler, W. (1990) Assessing the validity of memory
complaints; Performance of brain-damaged and normal individuals on Rey’s task to
detect malingering. Journal of Clinical Psychology, Q, 432-435.

Bigler, E. D. (1992). The neurobiology and neuropsychology of adult learning
disorders. Journal of Learning Disabilities, 2_5, 488-506.

Buchanan, M. & Wolf, J. S. (1986). A comprehansive study of learning disabled

adults. Journal of Learning Disabilities, 12, 34-38.

44

Campbell A. Dinsmore v. Charles C. Pugh and the Regents of the University of
California (1989) (as cited in Brinckerhoff, Shaw, & McGuire, 1992).

Cassisi, J. E. (1992). The detection of malingering and deception with a short
form of the MMPI-2 based on the L, F, and K scales. Journal of Clinical PsychologY. iii,
54-58.

Clawson, A. (1962). The Bender Visual Motor Gestalt Test for Children. Los

., . "“1

Angeles, Western Psychological Services.
Cohen, J. (1983). Learning disabilities and the college student: Identification

and Diagnosis. Adolescent Psychim, 11, 177—198.

F"

Dalke, C. (1988). Woodcock-Johnson Psychoeducational Test Battery profiles:

A comparative study of college freshmen with and without learning disabilities. J oumal

 

of Learning Disabilities, 21, 567-570.

Doerr, H. O. & Carlin, A. S. (Ed.) (1991). Forgraic Neuropsychology; New
York: The Guilford Press.

Felton, R. H., Naylor, C. E., & Wood, F. B. (1990). Neuropsychological profile
of adult dyslexics. Brain and Langaage, 32, 485-497.

Frank, K. & Wade, P. (1993) Disabled student services in postsecondary
education: Who is responsible for what?. Journal of College Student Development. 351,
26-30.

Gajar, A., Salvia, J ., Gajria, M., & Salvia, S. (1989). A comparison of intelligence
achievement discrepancies between learning disabled and non-learning disabled college
students. Learning Disabilities Research, 4, 119-124.

Gaj ar, A. (1992). Adults with learning disabilities: Current and firture research

45

priorities. Journal of Learning Disabilities, _2_§, 507-519.

Goebel, R. A. (1993). Detection of faking on the Halstead-Reitan
Neuropsychological Test Battery. Journal of Clinical Psychology. 39, 731-742.

Golden, C. J ., Purisch, A. D., & Hammeke, T. A. (1985). Luna-Nebraska
NeuropsychologicaiBattery: Form_s I and H. Los Angeles: Western Psychological
Services.

Gnys, J. A., Willis, W. G., & Faust, D. (1995). School psychologists’ diagnoses
of learning disabilities: A study of illusory correlation. J oumal of School Psychology,
3_3_, 59-73.

Goldberg, J .O. & Miller, H. R. (1986). Performance of psychiatric inpatients and
intellectually deficient individuals on a task assessing the validity of memory complaints.
Joumfl of Clinical Psychology, 52, 792-795.

Grober, E. & Sliwinski, M. (1991). Development and validation of a model for
estimating premorbid verbal intelligence in the elderly. Journal of Clinicalana
Experimental Neuropsychology, _1_3, 933-949.

Hall, R. J. (1984). Validity of the woodeoek-Johnson tests of achievement for
leaming-disabled students. Journal of School Psychology, 2, 193-200.

Hallahan, D. P. (1992). Some thoughts on why the prevalence of learning
disabilities has increased. Journal of Learning Disabilities, 2_5, 523-528.

Hammill, D. D., Leigh, J. B, McNutt G., & Larsen S. C. (1981) A new definition
of learning disabilities. LeamingQisability Quarterly, 4, 336-342.

Hamadeek, M. C. S. & Rourke, B. P. (1994). Principal Identifying features of

the syndrome of nonverbal learning disabilities in children. J oumaLOf Learning

46

Disabilities, fl, 144-154.

Hathaway S. R. & McKinley, J. C. (1951). The Minnesota Multiphasie
Perthy Inventory Manual (rev.). New York: Psychological Corporation.

Heaton, R. K., Smith, H. H., Lehman, R. A. W., & Vogt, A. T. (1978). Prospects
for faking believable deficits on neuropsychological testing. Journal of Conaultingfl
Clinical Psychology. 46, 892-900.

Hiscock M. & Hiscock, C. K. (1989). Refining the forced-choice method for the
detection of malingering. Journal of Clinical and Experimental Neuropsychology, l_l,
967-974.

Hem, J .L., O’Donnell, J. P., & Leicht, D. J. (1988). Phonetieally inaccurate
spelling among learning-disabled, head injured, and nondisabled young adults. Brain and
Language, 3_3_, 55-69.

Hughes, C.A. & Smith, J. O. (1990). Cognitive and academic performance of
college students with learning disabilities: A synthesis of the literature. m
Disability Quarterly. 3, 66-79.

Johnson, D. J. & Blalock, J. W. ( 1987). Adults With Learning Disabilities:
Clinical Studies. Orlando, Florida: Grune & Stratton, Inc.

Kahn, M. S. (1980). Learning problems of the secondary and junior college
learning disabled student: Suggested remedies. Journal of Learning Disabilities. _1_3, 40-
44.

Kaufrnan, A. S. (1990). Assessing Adolescent and Adult Intelligence. Boston:
Allyn & Bacon, Inc.

Lezak, M. D. (1995). Neuropsychological Assess_ment. (3rd ed.). New York:

47

Oxford University Press.

Lewin, T. (1996, February 13). College gets tougher on verifying learning
disabilities of aid applicants. New YorlaTimes, National Edition. pp. A16.

Lim, J. & Butcher, J. N. (1996). Detection of faking on the MMPI-2:
Differentiation among faking-bad, denial, and claiming extreme virtue. Journal of
Personality Assessment. (_6_7), 1-25.

McGrew, K. S. & Pehl, J. (1988). Prediction of future achievement by the
Woodcock-Johnson Psycho-Educational Battery and the WISC-R. J cm] of School
Psychology, 26, 275-281.

MeLeskey, J. (1992). Students with learning disabilities at primary, intermediate,
and secondary grade levels: Identification and characteristics. Learning Disabilig
Quarterly, 15, 13-19.

Mellard, D. F. & Byme, M. (1993). Leaming disabilities referrals, eligibility
outcomes, and services in community colleges: A four year survey. Learning Disability
Qaarterly, E, 199-218.

Mendelson, D. (1995). The expert deposes, but the court disposes: The concept
of malingering and the firnction of a medical expert witness in the forensic process.
International Journal of Law and Psychiafl, 18, 425-436.

Merrell, K. W. (1990). Differentiating low achieving students and students with
learning disabilities: An examination of performances on the Woodcock-Johnson
Psycho-Educational Battery. The Journal of Special Education. 24, 296-305.

Millis, S. R. & Kler, S. (1995). Limitations of the Rey Fifteen-Item Test in the

detection of malingering. The Clinical Neuropsychologi st, (2), 241-244.

48

Morris, M. (1990). A report of cognitive, academic, and linguistic profiles for
college students with and without learning disabilities. J ouflal of Learning Disabilities,
2_3, 355-360 & 385.

Nelson, H. E. & O’Connell, A. (1978). Dementia: The estimation of premorbid
intelligence levels using the new adult reading test. m, E, 234-244.

Nies, K. J. & Sweet, J. J. (1994). Neuropsychological assessment and
malingering: A critical review of past and present strategies. Archives of Clinical
Neuropsychology. 2, 501-552.

Olivarez, A., Palmer, D. J ., & Guillemard, L. (1992). Predictive bias with
referred and nonreferred Black, Hispanic, and White pupils. Learning Disabilig
Quarterly, 15, 175-186.

Obrzut, J. E. & Boliek, C. A., “Neuropsychological Assessment of Childhood
Learning Disabilities,” Handbook on the Assessment of Learning Disabilities: Theory
Research. and Practice. ed. H. Lee Swanson (Austin, Texas: Pro-ed, 1991) pp. 121-145.

Parry, J. (ed.) (1992). The Americans with Disabilities Act Mam}.
Washington, DC: American Bar Assiciation.

Paul, D.S., Franzen, M. D., Cohen, S. H., & Fremouw, W. (1992). An
investigation into the reliability and validity of two tests used in the detection of
dissirnilation. International Journal of Clinical Neuropsychology, M, 1-9.

Prigatano, G. P. & Amin, K. (1993). Digit Memory Test: Unequivoeal cerebral
dysfunction and suspected malingering. Journal of ClinicaL and Experimental
Neuropsychology. 1_5, 537-546.

Reitan, R. M. & Wolfson, D. (1993). The Halstead-Reitan Neuropsychological

49

Test Battery: Theoryand clinical interpretation. Tuscon, AZ: Neuropsychology Press.
Rey, A. (1941). Psychological examination of traumatic encephalopathy.
Archives de psychologia, 2_8, 286-340; sections translated by J. Corwin, & F. W. Bylsma,

The Clinical Neuropsychologial, 1993, 4-9.

Rey, A. (1964). L’examen clinique en psychologfi Paris: Presses Universitaires
de France.

Rourke, HP. (1982). Central processing deficiencies in children: Toward a
developmental neuropsychological model. Journal of Clinical Neuropsychology, 4_, l-18.

Rogers, R. (Ed.) (1997). Clinical Assessment of Malingering and Deception (2nd
ELI New York: The Guilford Press.

Schwartz & Saffran’s study (as cited in Grober & Swilinski, 1991).

Sinnett, E. R., Rogg, K. L., Benton, S. L., Downey, R. G., & Whitfill, J. M.
(1993). The Woodcock-Johnson revised: Its factor structure. Educational and
Psychological Measurement. 2, 763-769.

Stinnett, T. A. (1994). Current test useage by practicing school psychologists: A
national survey. Journal of Psychoedueational Assessment, _1_2, 331-350.

Tupper, D. (1990). Some observations on the use of the Woodcock-Johnson tests
of cognitive ability in adults with head injury. Journal of Leamig Disabilities. 2_3_, 306-
310.

United States v. Board of Trustees of the University of Alabama, 86-C-1779-S
(N .D. Ala., 1988); 408 F. 2d 740 (11th Cir. 1990).

Viglione, D. J. (1995). Maximizing internal and external validity in MMPI

malingering research: A study of a military population. Journal of Personality

50

Assessment, Q, 502-513.

Vogel, S. A. & Adehnan, P. B. (1992). The success of college students with
learning disabilities: Factors related to educational attainment. J oumal of Leamg
Disabilities, 25, 430-441.

Walsh, P. C., Lowenthal, B. & Thompson, G. (1989). Cluster score performance
of learning disabled students on the Woodcock-Johnson Psycho-Educational Battery.
Psychology in the Schools. E, 19-27.

Warner, M. M., Schumaker, J. B., Alley, G. R., & Deshler, D. D. (1980).
Learning disabled adolescents in the public schools: Are they different from other low
achievers? Exceptional Education Quarterly. 1, 27-36.

Webster’s New World Dictionm (3rd college ed.). (1988). New York: New
York, Simon and Schuster, inc.

Wechsler, D. (1974). Wech_sler Intelligence Scale for Children-Revised. New
York: Psychological Corporation.

Wechsler, D. (1981). Wechsler Adult Intelligence Scale-Revised. San-Antonio,
TX: Psychologocal Corporation.

Wendler, C. L. W. & Roid, G. H. ( 1984). A cost analysis approach to optimizing
test selection. Journal of Learning Disabilities. 12, 400-405.

Woodcock, R. & Johnson, M. (1977). Woodcoclflhnaon psychoeducatml
test battegy. Allen, TX: DLM.

Woodcock, R., & Johnson, M. (1989/1990). Woodcock-Johnson Psycho-
Educational Battery- Revised. Allen, TX: DLM Teaching Resources.

Wynne v. Tufts University School of Medicine, No. 89-1670 (lst Cir. 1990).

51

 

"Illllllllllllll'lllllll“