Wechsler Adult Intelligence Scale-III

2
Wechsler Adult Intelligence Scale-III
Joseph J. Ryan and Shane J. Lopez
INTRODUCTION
The Wechsler Adult Intelligence Scale-III (WAIS-III) comes from a tradition of mental ability
testing that began in 1939 with the publication of the Wechsler-Bellevue Intelligence Scale,
Form I (W-B I). The W-B I, which was named after David Wechsler and the Bellevue Hospital
where he was employed as chief psychologist, was considered a unique clinical instrument
because it possessed good face validity with adolescents and adults, grouped items into subtests,
and provided extensive assessment of both verbal and nonverbal abilities. The scale also uti-
lized standard scores and deviation IQs instead of the mental age values and ratio IQs pro-
vided by contemporary tests such as the Stanford-Binet Intelligence Scale, Forms Land M.
David Wechsler was known for his pragmatic view of human intelligence. Early in his
career as a test developer, he defined the construct of intelligence as "the aggregate or global
capacity of the individual to act purposefully, to think rationally, and to deal effectively with
his environment." This definition is not exclusively cognitive, but also includes affective (emo-
tional) and conative (e.g., motivational) factors. Thus, Wechsler felt that intelligence was
multifaceted, multidetermined, and a part of personality as a whole. To measure intelligence
so defined, he selected items and procedures from ability tests already in use, such as the
Army Alpha, Army Beta, Kohs Block Designs, Stanford-Binet, and Healy Picture Comple-
tion II. Some of the items were modified to increase their clinical potency, whereas others
were retained in their original form. The resulting product consisted of six Verbal Scale subtests
and five Performance Scale subtests. Wechsler considered the scales as different "languages"
through which the underlying construct of general intelligence may be expressed. This format
provided a separate Verbal IQ, Performance IQ, and overall composite score known as the
Full Scale IQ. For ease of interpretation, these scores shared the same distributional character-
istics (e.g., means and standard deviations) at all ages.
Joseph J. Ryan. Department of Psychology and Counselor Education, Central Missouri State University,
Warrensburg, Missouri 64093 Shane j. Lopez • Department of Psychology and Research in Education, University
of Kansas, Lawrence, Kansas 66045.
Understanding Psychological Assessment. edited by Dorfman and Hersen. Kluwer Academic/Plenum Publishers,
New York, 2001.
19
20 JOSEPH J. RYAN AND SHANE J. LOPEZ
The W-B I was a very popular instrument, but it had numerous structural shortcomings
and an inadequate standardization sample. In 1955, these limitations were rectified with the
publication of the Wechsler Adult Intelligence Scale (WAfS). This test was standardized on a
large national sample of persons 16 to 75 years of age and measured a wider range of cogni-
tive functions than its predecessor. Within a few years of its appearance, the WAIS was the
most widely used psychological test in the United States. By the late 1970s, it was obvious
that the WAIS needed to be modified, revised, and restandardized. This project was completed
in 1981 with publication of the Wechsler Adult Intelligence Scale-Revised (WAIS-R).
TEST CONSTRUCTION AND DEVELOPMENT
A decade after publication of the WAIS-R, the test was in need of revision and
restandardization. Psychologists and statisticians at The Psychological Corporation (TPC)
recognized this fact and responded by assembling an in-house working group directed by Dr.
David Tulsky to produce the WAIS-Ill. Tulsky and associates assembled an external panel of
experts in the field of intellectual assessment to serve in an advisory capacity throughout the
revision project. They also promulgated specific test development goals for the third edition
which included: (a) retention of the basic structure of the previous editions, (b) updating the
norms, (c) improving individual items, (d) enhancing clinical utility and strengthening the
theoretical basis, and (e) providing extensive data on scale reliability and validity.
Structure of the WAI5-111

As with previous editions of the scale, the WAIS-III retains the 11 traditional subtests and
yields Verbal, Performance, and Full Scale IQs. The revised Verbal Scale contains seven subtests.
The Vocabulary, Similarities, Arithmetic, Digit Span, Information, and Comprehension subtests
are combined to yield the Verbal IQ. The Letter-Number Sequencing subtest, which appears
in a Wechsler Scale for the first time, is designated as a supplementary measure. It contributes
to the Verbal IQ only when the Digit Span subtest has been spoiled (Le., Digit Span adminis-
tration is invalid due to an interruption in the test procedure or some other violation of stan-
dardization). Similarly, the Performance Scale consists of seven subtests where Picture Comple-
tion, Digit Symbol-Coding, Block Design, Matrix Reasoning, and Picture Arrangement are
combined to yield the Performance IQ. The Symbol Search and Object Assembly subtests are
designated as supplementary and optional tasks, respectively. The former may be substituted
only for a spoiled Digit Symbol-Coding administration, whereas the latter may replace any
spoiled Performance Scale subtest if the examinee is in the age range 16 to 74 years. The
Matrix Reasoning and Symbol Search subtests are new to the third edition.
The six Verbal Scale and five Performance Scale subtests are combined to calculate the
Full Scale IQ. After each subtest is scored, raw point totals are converted to scaled scores
according to the examinee's age range in Table A.l from the WAfS-lIl Administration and Scor-
ing Manual. These scaled scores have a mean of 10 and a standard deviation of 3. Sums of
scaled scores then are computed separately for the six Verbal Scale subtests, five Performance
Scale subtests, and all 11 subtests which constitute the Full Scale. The sums are converted to
deviation IQs using Tables A.3 through A.5 in the WAfS-III Administration and Scoring Manual.
The IQs generated have a mean of 100 and a standard deviation of 15 at all age levels.
The WAIS-III introduces four new composites which have been designated as ''index scores."
They are based on a series of factor analyses performed on the standardization sample data.
WECHSLER ADULT INTELLIGENCE SCALE-III 21
The Verbal Comprehension Index is composed of the Vocabulary, Similarities, and Informa-
tion subtests. The Perceptual Organization Index includes the Picture Completion, Block
Design, and Matrix Reasoning subtests. A Working Memory Index is comprised of Arith-
metic, Digit Span, and Letter-Number Sequencing. The fourth Index represents the construct
of Processing Speed and contains two subtests, Digit Symbol-Coding and Symbol Search.
The index scores are obtained by summing the scaled scores that constitute each of the com-
posites and then converting the sums to deviation quotients using Tables A.6 through A.9 in
the WAIS-lIf Administration and Scoring Manual. Each index score has a mean of 100 and a
standard deviation of 15. Table 2.1 presents a brief description of the WAIS-I1I subtests, op-
tional procedures, IQs, and indexes.
Table A.2 in the WAIS-lIf Administration and Scoring Manual provides optional norms
that allow examiners to convert raw subtest scores to scaled scores relative to a group of 600
young adults 20-34 years of age. These scaled scores are not age-appropriate and are never
used to compute the IQs and indexes. However, they can be consulted when a comparison of
an older examinee's achievement levels with that of an optimally functioning group of young
adults becomes relevant.
Updated Norms
The WAIS-III was normed on a stratified sample of 2450 individuals with negative histories for
psychiatric conditions, medical conditions, or both that could potentially affect cognitive func-
tioning. The stratification percentages for age, sex, race/ethnicity, educational level, and geo-
graphic region were based on the 1995 United States Bureau of the Census data. For each
stratification variable, there was a very close match between the normative sample and the
census figures. Table 2.2 summarizes the demographic characteristics of the WAIS-II1 stan-
dardization sample.
Item Development
The panel of experts assembled by TPC reviewed and critiqued the WAIS-III content and
made suggestions for updating existing items, the creation of new items, and the addition of
new subtests. Problematic items were identified and deleted from the Scale. With the assis-
tance of the advisory panel, TPC conducted a series of pilot investigations and tryout studies.
They assessed the psychometric characteristics of the modified Scale, determined item diffi-
culty levels, and analyzed test content for potential bias. A variety of patient samples were
tested in order to evaluate the clinical utility of new items, new subtests, and revised adminis-
tration and scoring procedures.
Enhancing Clinical Utility/Strengthening the Theoretical Basis

The WAIS and the WAIS-R were often criticized because of their limited floors and ceil-
ings. The IQs and subtest scores generated by these scales did not extend sufficiently down-
ward to allow for discrimination among examinees with mild to moderate mental retardation
or high enough to properly assess the intellectually gifted. To correct the former limitation,
less difficult items were added to the third edition, the range of possible subtest scores was
extended downward, and extrapolated IQs were generated at the lower end of the ability dis-
tribution. To address the second shortcoming, the ceiling of the WAIS-III was raised by five
points. However, as the range of WAIS-II1 Full Scale IQs is only 45 to ISS, the test is still
Table 2.1. WAIS-III Subtests, Optional Procedures, IQs, and Indexes
Subtest/IQ/lndex Description
Vocabulary Contains 33 words arranged in order of increasing difficulty. The examinee defines
each word aloud. It measures word knowledge, contributes to the Verbal Scale IQ
and the Verbal Comprehension Index.
Similarities Contains 19 pairs of words. The examinee is asked to identify similarities between
each pair. The subtest appears to measure verbal concept formation and abstract
thinking. It contributes to the Verbal Scale IQ and the Verbal Comprehension Index.
Arithmetic Contains 20 arithmetic problems that require counting, addition, subtraction,
multiplication, basic probability, or a combination of these. All items are timed, 3 use
blocks with oral instructions, and 17 are word problems. The required skill level does
not exceed 8th grade. The subtest always contributes to the Verbal Scale IQ and
Working Memory Index.
Digit Span Contains separate tasks of digits forward (series of numbers ranging from 2 to 9 digits
in length) and digits backward (series of numbers ranging from 2 to 8 digits) and
requires examinees to repeat digits in a prescribed order. Digits forward measures
auditory attention and auditory sequencing. Digits backward appears to measure
mental tracking, short term memory, and internal visual scanning. It contributes to the
Verbal Scale IQ and the Working Memory Index.
Information Contains 28 oral questions that assess general knowledge of objects, events, and places.
It contributes to the Verbal Scale IQ and the Verbal Comprehension Index.
Comprehension Contains 18 orally presented questions. Sixteen deal with everyday situations that assess
knowledge of social conventionalities and common sense. The examinee is also asked
to provide the meaning of two proverbs. It contributes to the Verbal Scale IQ.
letter-Number Contains seven items that assess the ability to order random series of orally presented
Sequencing numbers and letters. The subtest measures auditory tracking, mental flexibility, and
divided attention. It is a supplementary task that can substitute for Digit Span, if that
subtest is spoiled. If substituted for Digit Span, it contributes to the Verbal Scale IQ. It
always contributes to the Working Memory Index.
Picture Completion Contains 25 drawings of people, animals, common objects, and scenes. Each drawing
has a missing part that must be identified within 20 seconds. It measures the ability to
differentiate essential from nonessential details and contributes to the Performance
Scale IQ and the Perceptual Organization Index.
Digit Symbol- Contains a key with nine number-symbol pairs as well as 133 boxes with a number in
Coding the top portion and a blank space in the lower portion. The task requires drawing in
the lower portion of each box the symbol that was paired with the number in the key.
The subtest has a 120 second time limit and requires repeated visual scanning, visual
tracking, handwriting speed, and paired-associate learning. It contributes to the
Performance Scale IQ and the Processing Speed Index.
Block Design Contains 14 items that require the reproduction of design patterns using colored blocks.
All items are timed. It measures visuospatial problem solving, nonverbal concept
formation, and constructional ability. It contributes the Performance Scale IQ and the
Perceptual Organization Index.
Matrix Reasoning Contains 26 nonverbal, multiple-choice items that involve pattern completion, serial
reasoning, and/or classification. It has also been deSignated as a measure of fluid
intelligence. It contributes to the Performance Scale IQ and the Perceptual Organization
Index.
Picture Arrangement Contains 11 individual sets of pictures. Each set is presented in a mixed-up order and
needs to be put into a logical sequence within a specified time frame. It measures
visual sequencing and the ability to plan and anticipate within a social context. It
contributes to the Performance Scale IQ.
WECHSLER ADULT INTELLIGENCE SCALE·'" 23
Table 2.1. Continued

Subtest/lQ/lndex Description
Symbol Search Contains 60 items that each present two symbols on the left side of a page and five on
the right. Each requires a quick decision whether either symbol on the left is also
presented on the right. The time limit is 120 seconds and the subtest measures
perceptual discrimination, speed and accuracy, and sustained attention. It is a
supplementary task that can substitute for Digit Symbol-Coding. if that subtest is spoiled.
If substituted for Digit Symbol-Coding, it contributes to the Performance Scale IQ. It
always contributes to the Processing Speed Index.
Object Assembly Contains five jigsaw puzzles that must be solved within specified time limits. It measures
visual organization, appreciation of part-whole relationships, and constructional ability.
This is an optional task that may be administered for clinical reasons and may also be
substituted for any spoiled Performance Scale subtest if the examinee is in the age
range 16-74. If substituted for another subtest, it contributes the Performance Scale
IQ.
Digit Symbol- Administered immediately after the Digit Symbol-Coding subtest. It evaluates memory
Incidental for the associated number-symbol pairs and the individual symbols independent of
Learning the numbers. Thus, it is a measure of incidental learning of rehearsed material. This is
an optional procedure that may assist with the interpretation of the Digit Symbol-
Coding subtest.
Digit Symbol-Copy Administered as the final component of a WAIS-i11 evaluation. It consists of the nine
symbols used in the Digit Symbol-Coding subtest. Each item presents a symbol, with a
blank space directly below. The examinee copies as many symbols as possible within
90 seconds. It measures perceptual and graphopmotor speed. This is an optional
procedure that may assist with the interpretation of the Digit Symbol-Coding subtest.
VerballQ Comprised of the Vocabulary, Similarities, Arithmetic, Digit Span, Information, and
Comprehension subtests. The Letter-Number Sequencing subtest is a designated
supplementary measure. It contributes to the VerballQ only when the Digit Span
subtest has been spoiled. This composite score reflects general verbal abilities including
acquired knowledge, verbal reasoning. and attention to verbal information.
Performance IQ Comprised of the Picture Completion, Digit Symbol-Coding. Block Design, Matrix
Reasoning. and Picture Arrangement subtests. The Symbol Search and Object Assembly
subtests are designated as supplementary and optional tasks, respectively. The former
may be substituted only for a spoiled Digit Symbol-Coding administration, whereas
the latter may replace any spoiled Performance Scale subtest if the examinee is in the
age range 16 to 74 years. This composite score reflects perceptual organization, fluid
reasoning. and attention to detail.
Full Scale IQ Comprised of the 6 Verbal Scale and 5 Performance Scale subtests. This composite
score serves as a global measure of intellectual functioning.
Verbal Comprised of the Vocabulary, Similarities, and Information subtests. It measures
Comprehension verbal knowledge and understanding and. the ability to apply verbal skills across
Index situations.
Perceptual Comprised of the Picture Completion, Block Design, and Matrix Reasoning subtests.
Organization Index It measures the ability to interpret and organize visually perceived material.
Working Memory Comprised of Arithmet!c, Digit Span, and Letter-Number Sequencing subtests.
Index It measures auditory attention-concentration and verbal information processing
capacity.
Processing Speed Comprised of the Digit Symbol-Coding and Symbol Search subtests.
Index It measures the ability to process visual information quickly.
Table 2.2. Demographic Characteristics of the WAIS-II1 Standardization Sample

Variable Description
Age The 13 age groups were 16-17, 18-19, 20-24, 25-29, 30-34, 35-44, 45-54, 55-64,
65-69,70-74,75-79,80-84,85-89 years.
Sex For all age groups between 16-17 years and 55-64 years, there were 100 males and
100 females. There were 90 males and 110 females in the age range 65-69 years and
88 males and 112 females for those 70-74 years. For individuals in the age ranges 75-
79 years, 80-84 years, and 85-89 years, the male-female numbers were, respectively,
83 and 117,54 and 96, and 32 and 68.
Race/Ethnicity There were 1,925 whites, 279 African Americans, 181 Hispanics, and 65 individuals
from other racial/ethnic groups.
Educational Level Five educational categories were represented. The groups were ,,8 years (n = 284), 9-
11 years (n = 289),12 years (n = 853),13-15 years (n = 579), and ",16 years (n = 445).
Individuals 16-19 years of age were classified according to years of parental education.
Geographic Region The u.S. was divided into four regions consisting of nine northeastern states, 12 north
central states, 17 southern states, and 12 western states. Individuals were recruited
from each of these areas.
inappropriate for the assessment of examinees with moderate to severe mental retardation or
persons identified as extremely gifted.
To increase the clinical information provided by the WAIS-III, a new measure of inciden-
tal learning was added. Following administration of Digit Symbol-Coding, there now is an
optional procedure for evaluating an examinee's recall of the nine number-symbol pairs that
constitute the subtest. A second optional procedure requires the examinee to copy symbols as
quickly as possible without the need to form number-symbol associations. By comparing the
results of these optional tasks, it is often possible to ascertain the reasons for poor perfor-
mance (e.g., symbol recall inaccuracy or copying difficulties) on the Digit Symbol-Coding
subtest.
The WAIS-Ill was co-normed with the Wechsler Memory Scale-III (wAIs-m) for all age
groups in the standardization sample and linked with the Wechsler Individual Achievement
Test (WIAT) for persons 16-19 years of age. Co-norming involved the administration of both
the WAIS-Ill and WMS-III in counterbalanced order to 1,250 individuals in the WAIS-Ill stan-
dardization sample. Because of the relatively high correlation between intellectual ability and
memory functions, this procedure allows the clinician to determine if an examinee's memory
skills are commensurate with his or her level of intelligence. The linking process was based
on administration of both the WAIS-III and WIAT to a supplementary sample of 142 individuals
in the age range 16 to 19 years. This provides a method of comparing an examinee's intellec-
tual ability to his or her academic proficiency, a frequently important piece of information
when there is a need to identify the presence of a specific learning disability.
To strengthen the theoretical basis of the third edition, TPC made a number of changes.
First, a subtest that purportedly measures fluid intelligence was added to the scale. The Matrix
Reasoning subtest provides information that should assist examiners who wish to interpret
WAIS-Ill scores according to the Horn-Cattell theory of fluid and crystallized intelligence.
Also, by the addition of two supplementary subtests, Letter-Number Sequencing and Symbol
Search, examiners now have the option to regroup the various subtests and calculate theory-
based summary values that appear to measure the constructs of Working Memory and Pro-
cessing Speed. These dimensions are important for understanding cognitive functions since
they are thought to contribute to information acquisition and problem solving. Despite these
changes, the WAIS-III remains theoretically grounded in David Wechsler's pragmatic view of
intelligence.
Extensive Reliability and Validity Data

Internal Consistency
A series of studies conducted on the WAIS-I1I standardization sample indicated that it has
exemplary internal consistency. This type of reliability measures consistency with regard to
content sampling, as it is derived from a single administration of the test. For IQs and indexes,
it is derived by a formula for determining the reliability of a composite group of tests; for the
subtests, it is calculated by the split -half technique. The Full Scale reliability coefficients are
.97 or above across the entire age range in the standardization sample; Verbal Scale coeffi-
cients are .96 or higher; Performance Scale coefficients are .93 or higher. For the Verbal Com-
prehension, Perceptual Organization, Working Memory, and Processing Speed indexes, the
reliabilities are at least .94, .90, .91, and .86, respectively. Of the 14 subtests, Vocabulary has
the highest reliabilities (ranging from .90 to .95), whereas Object Assembly has the lowest
(ranging from .50 to .78). The Verbal Scale subtests are generally more reliable than the Per-
formance Scale subtests.
Standard Error of Measurement

Standard Error of Measurement (SEM) statistic reflects the reliability of the individual
score. It is the standard deviation of the distribution of error scores and is expressed in terms
of confidence limits that are placed around the obtained composite score or individual subtest
score. The larger the SEM, the larger is the band of error and the less precise the score. The
SEM is obtained by taking the square root of one minus the reliability coefficient and then
multiplying by the standard deviation of the test (i.e., SD -..J J-rxx). The average SEM for the
Full Scale IQ across all age groups is 2.30. The average SEM is larger for the Performance IQ
(3.67) than for the Verbal IQ (2.55). Because the indexes are based on fewer sub tests than the
traditional IQs, they yield scores that have slightly lower reliabilities and are somewhat less
precise. The average SEM is 3.01 for the Verbal Comprehension Index, 3.95 for the Percep-
tual Organization Index, 3.84 for the Working Memory Index, and 5.13 for the Processing
Speed Index. Of the seven Verbal Scale subtests, Vocabulary (0.79) has the smallest average
SEM, whereas Letter-Number Sequencing (1.30) has the largest. Within the Performance
Scale, the smallest average SEM is on the Matrix Reasoning subtest (0.97) and the largest is
on the Object Assembly subtest (1.66).
Test-Retest Stability
The WAIS-III has excellent temporal stability which was demonstrated by evaluating 394
individuals on two occasions with an average test-retest interval of thirty-five days. Uncor-
rected stability coefficients for the Verbal, Performance, and Full Scale IQs were, respec-
tively, .91, .83, and .91 for persons in the age range 16-29 years; .95, .88, and .96 for 30-54
years; .97, .91, and .96 for those 55-74 years; and .95, .93, and .96 for 75-89 years. Uncor-
rected stability coefficients for the Verbal Comprehension, Perceptual Organization, Working
Memory, and Processing Speed indexes were, respectively, .89, .79, .82, and .83 for persons
in the age range 16-29 years; .95, .86, .90, and .87 for 30-54 years; .96, .89, .90, and .89 for
55-74 years; and .93, .89, .85, and .92 for 75-89 years. Of the seven Verbal Scale subtests,
Information (.94) is, on the average, the most stable, whereas Letter-Number Sequencing
(.75) is the least stable. Within the Performance Scale, the Digit Symbol-Coding sub test (.86)
demonstrated the highest stability. The poorest stability was noted on the Picture Arrange-
ment subtest (.69).
The WAlS-lI/-WMS-lI/ Technical Manual reports changes in test scores by individuals in
the four age groups who were tested on two different occasions. The average Verbal IQ in-
creased by 3.2, 2.0, 2.1, and 2.4 points, respectively, for individuals in the age ranges 16-29
years, 30-54 years, 55-74 years, and 75-89 years. The average Performance IQ gains for each
of the four age groups were 8.2, 8.3,5.7, and 3.7 points. On the Full Scale IQ, increases from
test to retest averaged 5.7, 5.13.9, and 3.2 points. These retest changes, which are greater on
the Performance Scale than on the Verbal Scale, are likely the result of practice effects. How-
ever, the trend for retest gains on the Performance and Full Scale IQs to decrease in magnitude
as one moves from the youngest to the oldest groups is noteworthy. Perhaps they reflect age-
related cognitive changes in visual-spatial problem solving or incidental learning or both.
Retest changes on the four indexes are similar to those associated with the IQs. The average
practice effects across the four age groups was 2.4 points for the Verbal Comprehension In-
dex, 5.4 points for the Perceptual Organization Index, 2.4 points for the Working Memory
Index, and 3.9 points for the Processing Speed Index. The largest increases from test to retest
were on the Perceptual Organization and Processing Speed indexes. Of the seven Verbal Scale
subtests, the Vocabulary and Comprehension subtests demonstrated the smallest average re-
test gains across the four age groups (0.23 points), whereas Information had the largest mean
retest gain (0.58 points). Within the Performance Scale, the smallest retest gain was on the
Symbol Search Subtest (0.45) and the largest retest gain was on the Picture Completion Subtest
( 1.80).
Content Validity
A test that purports to measure intelligence must demonstrate appropriate content cover-
age and content relevance. To meet these requirements, test developers conducted extensive
literature reviews to identify any problems with items on previous editions of the scale. Next,
experts in the field of intellectual assessment reviewed potential WArS-III items, designed new
ones, and helped update others. Data from item tryout and pilot studies were also reviewed.
Based on these efforts, TPC felt that the third edition had achieved an appropriate level of
content validity.
Concurrent Validity
High correlations have been found between the WArS-III and other measures of intelli-
gence, achievement, and memory. These findings support concurrent validity because the
WArS-III and other tests were administered at approximately the same time. Concurrent valid-
ity studies reported in the WAlS-llI-WMS-llI Technical Manual indicate that the WArS-III Full
Scale IQ correlates .93 with the WAIS-R Full Scale IQ, .88 with the Wechsler Intelligence
Scale for Children-III Full Scale IQ, and .88 with the Stanford-Binet Intelligence Scale-IV
Composite Score. Correlations between the wrAT composites for Reading, Mathematics, Lan-
guage, and Writing and the WAIS-III Full Scale IQ are high and range from .68 to .81. The
magnitude of these associations allows for the prediction of achievement test performance of
an examinee based on knowledge of his or her WAIS-III scores. Another example of concurrent
validity is the relationship between various scores from the WMS-III and the WAIS-III. Correla-
tions between the eight memory scores and the Full Scale IQ are moderately high and range
from .36 to .68. These findings indicate that the WAIS-III and WMS-III, although significantly
related, are measures of different constructs.
Construct Validity
The pattern of intercorrelations among the WAIS-III subtests conform with expectations
based on practical knowledge and psychological theory. For example, the Verbal Scale subtests
intercorrelate more highly with each other than they do with the Performance Scale subtests,
whereas the Performance Scale subtests tend to intercorrelate more highly among themselves
than they do with the Verbal Scale subtests. Also, the 14 subtests are intercorrelated to a
statistically significant degree. This observation supports the premise that a pervasive, general
intelligence factor underlies the WAIS-III.
Exploratory factor-analytic studies ofthe WAIS-Ill standardization sample at five age ranges
(16-19,20-34,35-54,55-74, and 75-89) using 13 subtests (Object Assembly was excluded)
were reported in the WAIS-I/l-WMS-I/l Technical Manual. These procedures identified four fac-
tors, as mentioned previously: Verbal Comprehension, Perceptual Organization, Working
Memory, and Processing Speed. These factors were supported by a confirmatory factor analy-
sis of data from the same five age groups in the standardization sample. A separate explor-
atory factor analysis using a diagnostically heterogeneous patient sample (n = 152) was con-
ducted by the present authors. Using the same 13 subtests and identical statistical methodology,
highly similar results emerged. However, the Working Memory factor was composed of only
Digit Span and Letter-Number Sequencing. For the patient sample, the Arithmetic subtest
could not be uniquely allocated to any of the factors. Finally, a series of exploratory factor
analyses was conducted on the standardization sample separately for the total group and each
of the 13 age groups by Sattler and Ryan (1999). All 14 subtests (Object Assembly was added)
were used in the analyses. A four-factor solution characterized the WAIS-III, but the subtests
that contributed to each factor varied somewhat across age ranges.
The Verbal Comprehension and Perceptual Organization factors bear a close resemblance
to the Verbal and Performance scales. This provides construct validity support for interpreta-
tion of the Verbal and Performance IQs as meaningful cognitive dimensions. The significant
intercorrelations among the 14 subtests and the magnitude of the subtest loadings on the first
unrotated factor reported by Sattler and Ryan support the construct validity of the Full Scale
IQ. Overall, 50% of the variance in the WAIS-Ill may be attributed to the construct of general
intelligence.
Another way to assess construct validity is to ascertain whether scores on the WAIS-Ill
conform to the expectations of a credible theory of intelligence. For example, the Horn-Cattell
theory postulates that general intelligence can be divided into two separate components that
represent fluid and crystallized ability. The former involves the capacity to solve novel prob-
lems and think abstractly, whereas the latter involves the retrieval of previously learned and
stored verbal information. Fluid ability tends to decline with advancing age and is disrupted
by damage to the central nervous system. Crystallized ability is relatively resistant to age-
related deterioration and, in the absence of aphasia, severe brain dysfunction, or both, shows
less decline secondary to brain damage than do fluid abilities.
lfthe WAIS-III is viewed within the Horn-Cattell framework, the majority of Verbal Scale
subtests may be considered measures of crystallized intelligence, whereas most of the Perfor-
mance Scale subtests may be designated as measures of fluid intelligence. Inspection of the
WAIS-1II norms indicates that there are few differences between the older and younger stan-
dardization participants in verbal ability. Thus, on subtests such as Vocabulary and Informa-
tion, it takes about the same number of raw score points to earn a scaled score of 10 (50th
percentile) regardless of age. The situation is dramatically different for older and younger
standardization participants on the Performance Scale subtests, as the norm tables indicate
that older individuals have less adequate fluid abilities than do younger persons. This is clearly
demonstrated by the observation that, in the age range 20-34 years, it takes 76 raw score
points on the Digit Symbol-Coding Subtest to earn a scaled score of 10. However, in the age
range 85-89 years, it takes only 33 raw score points to earn the same scaled score. Likewise,
on the Symbol Search subtest, persons in the age range 20-34 need at least 33 raw score
points to earn a scaled score of 10, whereas individuals in the age range 85-89 years need to
earn only 14 raw score points to obtain a scaled score of 10.
ADMINISTRATION
Applications of the WAIS-III

The WAIS-III is a standardized instrument developed for use with persons in the age range
16-89 years. Examinees must have adequate vision and motor functions to permit a valid
administration of the Performance Scale subtests and should also be fluent in English and
possess adequate hearing to permit a valid administration of the Verbal Scale subtests. How-
ever, administration and interpretation of the measure to English-speaking persons from coun-
tries other than the United States should be done cautiously. In such cases, the examiner must
be alert to the possibility that failure on certain test items (e.g., Information subtest items 10
and II) reflects cultural loading, not an intellectual limitation. Interpretation of performance
data of individuals over 89 years of age should be done cautiously as norms are not available.
The use of the WAIS-III with elderly Hispanics and members of ethnic groups such as Asian
Americans and Native Americans should be carefully considered given that minimal data
were collected from these subgroups of the population.
As a measure of intellectual functioning, the WAIS-III can be used with individuals de-
scribed above. General applications of the measure include obtaining IQ estimates for both
adolescents with above-average functioning and adults, for determining examinees' cognitive
strengths and weaknesses, and for examining the influence of psychiatric difficulties on cog-
nitive functioning. Specific uses of the WAIS-III include diagnosing giftedness and mental
retardation and determining the extent of neuropsychological impairments and cognitive de-
cline due to age-related and non-age-related factors.
Test User Qualifications

Examiners should have training and experience in the administration of standardized
psychological instruments. Formal graduate courses in psychometrics and clinical assessment
(e.g., interviewing, mental status examination) are highly desirable. The examiner must be
thoroughly familiar with the instructions and procedures contained in the WAIS-l/I Administra-
tion and Scoring Manual, including the proper sequence of subtest administration, subtest
starting points and discontinuance rules, and the specific time limits required for administra-
tion of individual items. The last requirement is significant because some of the WAIS-I1I subtests
emphasize speed of performance and require precise timing to insure valid administration.
The examiner should also know when to utilize the supplementary and optional subtests.
Finally, some of the subtests are complex and require the examiner to simultaneously
manipulate the test manual, stimulus materials, and a stopwatch. Therefore, it is imperative
that examiners practice scale administration numerous times prior to using the WAIS-III with a
client or patient. The Sattler and Ryan (1999) Administrative Checklist for the WAfS-Ill pro-
vides a useful guide for judging the adequacy of an examiner's administrative performance.
Establishing and Maintaining Rapport

It is important for the examiner to establish and maintain rapport throughout each testing
session. Characteristics of empathy, genuineness, and warmth help communicate a desire that
each examinee perform at his or her best. The rapport-developing process includes an appro-
priate greeting of the examinee and explanation of the reason for the assessment and the
nature of the testing procedures. Next, rapport building involves the establishment of a con-
versational manner, an interactive relationship, and a non-threatening environment. The skilled
examiner will encourage the client to work efficiently, will respond to his or her questions,
and will attend to obvious needs for encouragement, redirection, or a break. Verbal and non-
verbal reinforcement should be given freely to maintain a high degree of effort on the part of
the examinee, but reinforcement should not be contingent on whether a response is correct or
incorrect. The WAfS-III Administration and Scoring Manual provides additional hints for es-
tablishing and maintaining rapport with individual examinees.
Standardized Administration
Reliability and validity of test scores are directly affected by test administration tech-
niques. The WAfS-III Administration and Scoring Manual provides the detailed instructions for
administration which should be read verbatim. Order of administration along with starting,
reverse, and discontinuation rules should be followed carefully. The manual also provides
information about some of the nuances of administration that are described below.
Repetition of Items
Repetition of subtest instructions is allowed. Use judgment when deciding to repeat items
from Verbal subtests. Digit Span and Letter-Number Sequencing items cannot be repeated as
that would invalidate the test. Repetition or practice of Performance subtest items should only
be done when indicated by the standardized instructions. Excessive requests for repetition of
items should be discussed and a reevaluation of the examinee's physical ability to complete
the WAIS-III should be considered.
Querying Responses
During scale administration, the use of some probing questions or queries is necessary to
help clarify vague or incomplete responses. The decision to query a response is made by the
examiner, but the WAfS-ill Administration and Scoring Manual provides examples and guide-
lines concerning when to initiate the process. During a typical WAIS-III administration, the
Vocabulary, Comprehension, and Similarities sub tests require the most frequent queries.
Administration Time
Wechsler (1997) reported that administration of the 11 standard subtests required to yield
the three IQs takes 75 minutes (range from 60 to 90 minutes), whereas administration time for
the 11 subtests that produce the four indexes is approximately 60 minutes (range from 45 to
75 minutes). Measurement of both the IQs and Indexes requires administration of 13 subtests
and takes 80 minutes (range from 65 to 95 minutes).
These estimates reflect the time needed to administer the WAIS-Ill to healthy individuals.
Assessment of persons with physical, cognitive, psychiatric, or a combination of these dis-
abilities is more time consuming and individual examiners must be aware of this fact so that
they can plan their testing sessions accordingly. For instance, the present authors tested 62
patients with a variety of medical and psychiatric conditions and found that average adminis-
tration time for the 11 standard subtests exceeded 90 minutes and for 13 subtests was almost
100 minutes.
Testing of Limits
Testing of limits is a term that describes the practice of going beyond standard testing
procedures to gather additional performance data from the examinee. Limits testing should
only be conducted after the entire test has been administered via standard procedures. Then,
additional procedures could be utilized to determine if an examinee's abilities surpass the
ceiling established by standard administration of the test or if elimination of time limits im-
proves performance. Additionally, probing questions could be used to establish the methods
by which examinees reach solutions. This extension of the standard testing procedure gener-
ates qualitative information that could shape clinical impressions of an examinee's abilities.
Scoring the WAIS-III

Scoring the WAIS-Ill subtests, scales, indexes and optional procedures is no simple mat-
ter. The literature indicates that graduate students, and even experienced psychologists, some-
times convert scores incorrectly, give inappropriate credit to individual items, and make mis-
takes when adding raw score points and scaled scores. On occasion, credit may be assigned
incorrectly to the number-symbol pairs on the Digit Symbol-Coding subtest and on optional
procedures. These scoring errors can result in misleading and inaccurate results. Examiners
need to be particularly cautious when scoring the Comprehension, Similarities, and Vocabu-
lary subtests. Personal experience indicates that examiners sometimes fail to credit responses
that appear in the test manual or do give credit to spoiled responses (Le., examinee's elabora-
tion reveals a misconception of the item). To guard against error, examiners need to have a full
appreciation of the scoring criteria, scoring guidelines, and scoring examples in the WAfS-ill
Administration and Scoring Manual.
INTERPRETATIVE GUIDELINES
Extracting the meaning of WAIS-III results can be a difficult task for clinicians of all
experience levels. Thus, rules of thumb for interpreting quantitative and qualitative perfor-
mance data are provided. The rules should be applied within the framework of a hypothesis-
driven interpretive approach, which will be outlined and discussed below.
Analysis of WAI5-m Results

Examining Quantitative Data: Analyzing IQs and Index Scores
The analysis of WAIS-I1I results begins by assessing the degree of intersubtest scatter (i.e.,
variability between the subtests) in the profile. This is done to determine whether the various
composites (e.g., IQs and indexes) hang together and provide adequate reflections ofthe con-
structs under consideration. For example, it is appropriate to interpret the Verbal IQ as a
measure of verbal comprehension and expression and the Performance IQ as a measure of
perceptual organization and nonverbal problem solving when the component subtests of the
scales cluster around their respective scaled score means. It is inappropriate to interpret the
constructs if marked variability is present in one or both scales.
It is important to note that a considerable degree of intersubtest scatter characterizes the
WAIS-I1I. For the standardization sample, the median and mean scatter ranges were both 7
scaled score points when 13 subtests were administered. Therefore, the present authors ques-
tion whether a composite group of subtests provides a meaningful dimension of ability when-
ever the degree of scatter within the composite equals or exceeds that for 95% of standardiza-
tion participants. If the magnitude of scatter is below that of 95% of the standardization sample,
the composite is usually accepted as a unitary reflection of mental ability.
Table 2.3 provides an illustrative example of a WAIS-Ill protocol in which all the standard
composites hang together according to the rule of thumb provided above. On the Verbal Scale,
the largest deviation from the mean (M = 10) is 2.0 points. By consulting Table B.3 in the
WAlS-l/l Administration and Scoring Manual, it can be seen that when seven subtests are ad-
ministered from the Verbal Scale, deviations greater than or equal to those found in 95% of
persons in the standardization sample range from 3.00 points on Vocabulary to 4.43 points on
Digit Span. On the Performance scale, when six subtests are administered the largest devia-
tion from the mean (M = 8.2) is 3.8 points. From Table 2.4 it is apparent that when six subtests
are administered, deviations from the mean that equal or exceed those of 95 % of the standard-
ization participants range from 3.57 points on Symbol Search to 4.14 points on Digit Symbol-
Coding. None of the subtests scores deviate meaningfully from the mean. Even the large
deviation noted for the Picture Arrangement subtest is not considered meaningful because a
difference of 4.03 points is required to equal or exceed 95% of the standardization sample.
Base rate tables that report cumulative scatter values for the indexes within the standard-
ization sample are not yet available. However, the authors' preliminary calculations suggest
that the reasonable threshold (i.e., scatter that equals or exceeds that of 95% of the standard-
ization sample) for questioning whether a subtest composite reflects a meaningful dimension
of ability is when one or more of its component subtests deviates from the mean by three or
more points. As demonstrated in Table 2.3, each of the Indexes provides a unitary estimate of
the construct it purportedly measures.
In the present example, none of the subtest scores deviate meaningfully from their re-
spective scale or index means. Therefore, it is feasible to interpret the various WAIS-lll com-
posites. From Table 3.3 it is apparent that the Verbal and Performance IQs differ by 11 points.
By consulting Table B.I in the WAlS-l/l Administration and Scoring Manual, it can be seen that
a difference of this magnitude is reliable at the 95% level. Thus, it probably is not the result of
measurement error and likely reflects a real difference in how the examinee expresses his or
her intelligence. However, inspection of Table D.3 in the WAlS-l/l Technical Manual indicates
that a difference of 11 points occurred in approximately 32% of persons in the standardization
sample whose Full Scale IQ fell within the ability range 90-109 when the direction of the
discrepancy is ignored. The frequency is 16% when the direction of the difference is consid-
Table 2.3. Illustrative WAIS-III Results

Subtests, Scales, Scaled Scale
and Indexes Score Mean' Difference
Verbal Scale
Vocabulary 10 10 0.0
Similarities 9 10 -1.0
Arithmetic 10 10 0.0
Digit Span 11 10 1.0 VerballQ = 101
Information 10 10 0.0
Comprehension 12 10 2.0
letter-Number 8 10 -2.0
Sequencing
~!:rformance Scale
Picture Completion 8 8.2 -0.2

Digit Symbol-Coding 7 8.2 -1.2
Block Design 6 8.2 -2.2
Matrix Reasoning 10 8.2 -1.8 Performance IQ = 90
Picture Arrangement 12 8.2 3.8
Symbol Search 6 8.2 -2.2
Full Scale IQ = 97
Verbal ComRrehension
Information 10 9.7 0.3
Vocabulary 10 9.7 0.3 Index = 98
Similarities 9 9.7 -0.7
PerceRtualOrganization
Picture Completion 8 8.0 0.0
Block Design 6 8.0 -2.0 Index = 88
Matrix Reasoning 10 8.0 2.0
Working Memor},
Arithmetic 10 9.7 0.3
Digit Span 11 9.7 1.3 Index = 97
letter-Number 8 9.7 -1.7
Sequencing
Processing SReed
Digit Symbol Coding 7 6.5 0.5 Index = 81
'Scale means are Ihe average respectively of the seven verbal subtests and six. performance sublests. Index
means are composed of only the sublests that contribule to the composite scores thai contribute to the Scalel
Index.
ered (divide the cumulative percentage from Table D.3 in half). This indicates that the 11-
point Verbal IQ-Performance IQ discrepancy is not unusual (the authors consider a difference
to be unusual if it occurred in 5% or fewer of the standardization participants) and that the
Full Scale IQ provides a reasonable explanation of the examinee's cognitive functioning.
Once the decision has been made to interpret the Full Scale IQ, it is necessary to convert
the score to an ability level using Table 2.2 in the WAfS-Ill Administration and Scoring Manual.
Next, the IQ is assigned a percentile rank and banded by either the 90% or 95% confidence
limits. Tables A.3 through A.5 in the WAfS-Ill Administration and Scoring Manual list the
percentile ranks and confidence limits for the Verbal, Performance, and Full Scale IQs. These
Table 2.4_ Deviations from the Mean of Six

Performance Scale Subtests that Occurred
in ~ 5% of the Standardization Sample'
Deviation from
Subtest the Mean
Picture Completion 3.93

Digit Symbol-Coding 4.14
Block Design 3.65
Matrix Reasoning 3.79
Picture Arrangement 4.03
Symbol Search 3.57
'Adopted from LoBello et al. (1998)
confidence limits are based on the standard error of estimation (SEE), not the SEM discussed
previously and presented in Table 3.4 of the WAfS-l/l-WMS-l/l Technical Manual. Confidence
limits based on the SEM for all age groups in the standardization sample are provided in Table
0-1 of Sattler and Ryan (1999). The SEE and the SEM are interpreted in exactly the same
manner and yield highly similar confidence limits for IQs within the middle portion of the
ability distribution. However, at the extremes of the distribution the SEE provides confidence
limits that are asymmetrical around the obtained score. The SEM provides symmetrical con-
fidence limits at all intelligence levels. The reason for this difference is that the SEE takes into
consideration the phenomenon of regression to the mean and assumes that examinees with
high IQs benefit from positive chance error and those with low IQs are compromised by
negative chance error.
As mentioned above, the present example has an II-point difference which is statisti-
cally significant and therefore reliable, but not unusual because a Verbal-Performance IQ
discrepancy of this magnitude occurred in 16% of the standardization sample. Also, because
the subtest scores within the Verbal Scale do not differ meaningfully from their mean, evi-
dence that the Verbal IQ reflects the unitary dimension of verbal comprehension and expres-
sion is provided. Similarly, the absence of meaningful scatter within the Performance Scale
indicates that the IQ may be interpreted as a unitary measure of the perceptual organization
and clerical speed construct.
The next step in the interpretative process is to examine the relationships between and
among the index scores by consulting Table B.I in the WAfS-ill Administration and Scoring
Manual. It can be seen that the Verbal Comprehension Index is significantly larger than both
the Perceptual Organization and Processing Speed Indexes, whereas the Working Memory
index is superior to the Processing Speed Index. On the other hand, the Working Memory
Index does not differ from either the Verbal Comprehension or Perceptual Organization in-
dexes. Finally, the Perceptual Organization and Processing Speed indexes are essentially the
same.
The Verbal Scale of the WArS-III reflects two major abilities, the first associated with the
Verbal Comprehension Index and the second with the Working Memory Index. In the present
example, these composites are comparable and indicate that the Verbal IQ provides an effec-
tive measure of the global construct of interest. The Performance Scale subtests measure two
principal abilities, with Perceptual Organization corresponding to the first and Processing
Speed to the second. For the current example, these composites are comparable and indicate
that the Performance IQ likely provides a valid measure of the major construct of interest.
Examining Qualitative Information: Detecting Errors

Numerical scpres are not always sufficient for detecting and understanding a client's
problems. Qualitative information, such as that gained by behavioral observations, may pro-
vide important insights into an examinee's cultural background, cognitive deficits, thought
processes, psychiatric status, or some combination of these. Likewise, it is often helpful to
document the nature and effectiveness of the strategies that an examinee employs en route to
either passing or failing individual items. Table 2.5 provides examples of qualitative errors on
selected WAIS-III subtests that, if encountered during a WAIS-III administration, should not be
overlooked.
Hypothesis-Driven Approach to Interpreting the WAIS-III
To fully utilize the quantitative and qualitative information provided by a WAIS-III exami-
nation, an organized approach that attempts to integrate historical information, behavioral
observations, and test results should be followed. The approach used in this chapter is summa-
rized in Table 2.6. It begins with a thorough review of available information from the client's
medical, educational, and social histories, the reason for referral, and the behavioral observa-
tions made during test administration. Prior to calculating the WAIS-IIJ scores, the examiner
formulates a few hypotheses and expectations about the test results.
Table 2.5. Examples of Qualitative Errors Associated with Selected WAIS-JII Subtests
Subtest Errors and Hypothesized Interpretations
Similarities "Not alike" response: May reflect concrete thinking, low IQ, or a psychological
characteristic such as negativism.
Stimulus bound response: Provides common associations to the word pairs, not
similarities. For example, in response to item 8, the examinee states, "Ear you hear
out of and eye you see out of." Sometimes given by concrete thinkers, persons with
low IQ, or patients with brain damage.
Digit Span Incorrect sequencing: The examinee recalls all the elements, but fails to maintain the
order of presentation. Sometimes associated with a learning disability and occasionally
noted in the records of individuals with significant depression or anxiety.
Initial errors: Incorrectly recalls the first or second digit in a sequence containing five
or fewer digits. When made by an examinee who is fully alert, it is likely that he or
she deliberately failed the item
Block Design Broken gestalts: Failure to maintain the 2x2 or 3x3 configuration of one or more
designs. Seen occasionally in the records of persons with documented brain damage.
Reflects a marked impairment of visuospatial ability.
Stacking: Blocks piled one on top of the other. An unusual error that suggests the
presence of brain dysfunction.
Digit Symbol-Coding Shape completion: Symbols are converted into perceptually similar forms. For instance,
the symbol A is converted into an "A." Seen occasionally in the records of patients
with brain damage.
Associating wrong symbol with a number: Suggests failure to comprehend instructions,
inattention, or visual scanning problems.
Table 2.6. Hypothesis-Driven Approach to Interpreting the WAIS-lII'

Phase I: Generating Hypotheses
1. Prior to calculation of the WAIS~II scores, gather data about the client's background information (e.g., medical,
educational, and social histories) and review the behavioral observations made during test administration.
Formulate a priori expectations or hypotheses about the examinee's test performance.
2. Translate a priori expectations or hypotheses into specific test results using a sequential process as follows:
a. Full Scale IQ
b. Verbal and Performance IQs
c. Index Scores
d. Subtest Scores
e. Individual Item Responses
Phase 1/: Profile Analysis and Hypothesis Testing
3. Evaluate the degree of intersubtest scatter in the Profile and determine the feasibility of interpreting the
various WAIS-III composites.
4. Test a priori expectations or hypotheses against the obtained scores.
5. If a priori expectations and hypotheses do not account for the data, formulate a posteriori explanations and
hypotheses for the findings.
Phase III: Gathering More Data and Drawing Concfusions
6. Collect additional data to test or support the a priori and a posteriori explanations and hypotheses.
7. Draw conclusions, but place an emphasis on explanations and hypotheses that are supported by at least two
pieces of evidence.
'When implementing this framework. clinicians should employ clinical decision-making strategies that minimize bias and misuse
of heuristics.
CASE EXAMPLE
Hypothesis-Driven Interpretive Approach

This section describes the WAIS-I1I results of EM., a 23-year-old woman with a high
school diploma who attended junior college for approximately two years. She had recently
transferred to a small, four-year liberal arts college in hope of obtaining a bachelor's degree in
political science. Her career goal was to become "._.a paralegal and maybe even a trial law-
yer."
Phase I. She is socially active and well liked by fellow students, but she has experienced
difficulty with most of her courses. The teaching faculty have given her extra attention during
class and also have provided one-on-one tutoring in mathematics and science. Because EM.
has not benefited academically from these interventions, she was referred by her academic
advisor for an evaluation of cognitive, academic, and general adaptive capacities.
EM. denies the presence of medical or psychiatric problems and takes no prescription
medications. However, the history is positive for multiple head injuries. The first occurred
when she was three years old and was associated with a brief loss of consciousness. A second
head injury was sustained in a motor vehicle accident when she was 18 years old. She remem-
bers hitting her head on the windshield and receiving one or more lacerations that required
stitches. This trauma was not associated with loss of consciousness, although she was con-
fused for a brief period and then fell asleep shortly after the injury. According to EM_, during
the second grade she was identified as having a learning disability, but she is unable to provide
further details. This problem was again identified during attendance at junior college, where
she received special assistance. (Please note that there is no way of knowing ifthe head injury
as a toddler caused the learning problems. A cause-and-effect relationship cannot be inferred.)

After two years of judicious course selection, her grade point average was sufficient for trans-
fer to a four-year college.
F.M. was seen for evaluation at 9:00 a.m. and was then administered an extensive battery
of tests. However, the present discussion will focus exclusively on the WAIS-III. Behavioral
observation revealed a friendly, talkative, and cooperative young lady. Affect was stable and
mood was mildly anxious, but generally appropriate to the situation. There was no evidence
of formal thought disorder or unusual thought content, but she was somewhat unsure of her-
self and giggled excessively. The latter behavior seemed immature for a 23-year-old woman,
but it may have been her way of handling test-related anxiety. Although the session took over
five hours, she worked without evidencing significant fatigue and her approach to the tasks
was slow and deliberate. Finally, she was frequently aware of her failures, but appeared dis-
couraged and frustrated on only one occasion. She was visibly uncomfortable and tense dur-
ing a task that required the solution of basic arithmetic problems without the aid of pencil and
paper. During this task she commented, "I've always been dumb in math, just don't get it'"
As Table 2.6 recommends, available background information and behavioral observa-
tion data were reviewed and reasonable a priori expectations about F.M.'s WAIS-Ill perfor-
mance were generated. Listed below are some expectations and hypotheses presented in a
sequential manner starting with the most global score:
I. The history of head injuries and possible learning disability suggests that her general
cognitive functions (i.e., Full Scale IQ) may be lower than expected, even through she
has completed high school and two years of junior college.
2. Comparable achievement is expected on the Verbal and Performance IQs. Persons
who have fully recovered from a closed head injury do not usually exhibit marked IQ
discrepancies. Also, data on groups with either traumatic brain injury or learning
disabilities reported in the WAlS-J/I-WMS-1/I Technical Manual did not produce large
Verbal-Performance IQ differences.
3. The Working Memory and Processing Speed indexes may reveal selective weaknesses.
Data in the WAlS-lI/-WMS-1/I Technical Manual indicate that learning-disabled indi-
viduals score poorly on the former index. Also, it was readily apparent during the
examination that she was unable to deal effectively with basic arithmetic calcula-
tions, a component of the Working Memory Index. The WAlS-I/I-WMS-1/I Technical
Manual also indicates that persons with a history of traumatic brain injury demon-
strate a relative weakness on the Processing Speed Index.
4. F.M. had received tutoring in mathematics and during administration of the Arith-
metic subtest she was unable to cope with task requirements. Therefore, her scaled
score on the Arithmetic subtest should emerge as a relative weakness in the overall
profile. Memory problems sometimes contribute to poor academic performance. Thus,
it is reasonable to expect EM. to experience some difficulty on the supplementary
Digit Symbol-Incidental Learning procedure.
5. The history oflearning disability suggests that auditory sequencing problems may be
seen on the Digit Span items and that answers to the Arithmetic items may reveal
specific types of errors (e.g., difficulty in carrying or borrowing).
All of the hypotheses offered in this example may not be supported by the calculated WAIS-III
scores. However, they constitute an attempt to integrate external information with the test
results and force the examiner to consider the complete array of information provided by the
WAIS-III.
Phase II. The next step in WArS-III interpretation is to evaluate the degree of intersubtest
scatter in the profile and thereby determine whether the summary scores represent valid esti-
mates of their respective constructs. Table 2.7 summarizes the quantitative and qualitative
information gleaned from F.M.'s evaluation. Analysis begins with an examination of the val-
ues provided in the "Difference" column of Table 2.7. For the Verbal subtests, the numbers are
compared to the values in Table B.3 of the WAfS-ill Administration and Scoring Manual. As
can be seen, the Vocabulary (4.3 points) and Arithmetic (-6.7 points) scores deviate meaning-
fully from the Verbal Scale mean. These are considered meaningful deviations because only
5% of the WArS-III standardization sample had scores that equaled or exceeded 3.00 points on
Vocabulary and 3.57 points on Arithmetic. These findings suggest that the Verbal IQ does not
provide a unitary estimate of the verbal comprehension construct and that a comparison of
this value with the Performance IQ will yield misleading information.
Examination of scatter within the Performance Scale is accomplished by comparing the
deviation scores under the "Difference" column in Table 2.7 with the values provided in Table
2.4. This indicates that none of the subtest scores deviate meaningfully (i.e., equal or exceed
that of 95% of the standardization sample) from the mean. Therefore, the Performance subtests
represent a unitary construct and the resulting IQ may be interpreted as a valid estimate of
F.M.'s processing speed and perceptual organization abilities.
Next, the cohesiveness of the four indexes is evaluated using the rule of thumb presented
previously. That is, an index is considered cohesive if none of its component subtests deviate
from the mean by 3.0 or more points. As can be gleaned from Table 2.7, the subtests that
comprise the Verbal Comprehension, Perceptual Organization, and Processing Speed indexes
demonstrate minimal scatter from their respective means and appear to represent unitary abili-
ties. On the other hand, F.M. performed inconsistently on subtests comprising the Working
Memory Index. Two of the three subtest scores deviate from the mean by 3.0 or more points,
indicating that this composite is not unitary in what it measures.
The fourth step in WArS-III interpretation involves testing the a priori expectations and
hypotheses against the obtained scores. As can be seen from Table 2.7, the Verbal, Perfor-
mance, and Full Scale IQs are significantly below expectations for someone with 14 years of
education, which confirms the first hypothesis. Based on F.M.'s demographic characteristics
(e.g., age, sex, education, race, social economic status, and urban vs. rural residence) it is
reasonable to expect average to above average IQs. The second hypothesis is also confirmed
since there is no Verbal-Performance IQ discrepancy.
To address the third hypothesis, it is necessary to consult Table B.l in the WAfS-ill Ad-
ministration and Scoring Manual. By reading down the appropriate columns in the table, one
can see that the Working Memory Index is significantly lower than each of the other indexes.
By following the same procedure, it is can be seen that the Processing Speed Index is signifi-
cantly below the Verbal Comprehension Index, but statistically equivalent to the Perceptual
Organization Index. These findings confirm the third hypothesis since the Working Memory
and Processing Speed indexes represent weaknesses in F.M.'s WArS-III profile.
The Arithmetic scaled score emerges as the single most prominent weakness in the pro-
file. Examination of the Digit Symbol-Incidental Learning scores indicates that F.M. freely
recalled six of the nine individual symbols and six of the associated number-symbol pairs.
These scores are below the 10th percentile and suggest that F.M. may experience learning and
memory problems. These findings confirm the hypothesis that addressed individual subtest
performance.
Finally, examination of item responses reveals errors on the Digit Span subtest that are
suggestive of a deficit in auditory sequencing. Analysis of F.M.'s answers to items on the
Table 2.7. EM.'s WAIS-III Results
~ Indexes
Verbal 86 Verbal Comprehension (VC) 101
Performance 86 Perceptual Organization (PO) 80
Full Scale 85 Working Memory (WM) 69
Processing Speed (PS) 88
Verbal Subtests SCi!h,~d SI;Qrl: ~ Difference
Vocabulary 12 7.7 4.3

Similarities 9 7.7 1.3
Arithmetic 7.7 -6.7
Digit Span 6 7.7 -1.7
Information 10 7.7 2.3
Comprehension 8 7.7 0.3
Letter-Number Sequencing 8 7.7 0.3
Perform anS;I: Subtl:~l

Digit Symbol-Coding 9 7.8 1.2
Block Design 7 7.8 -0.8
Matrix Reasoning 5 7.8 -2.8
Picture Arrangement 11 7.8 3.2
VC Subtests
Vocabulary 12 10.3 1.7
Similarities 9 10.3 -1.3
Information 10 10.3 -1.7
PO Sublests
Block Design 7 6.7 0.3
Matrix Reasoning 5 6.7 -1.7
WM Subtests
Arithmetic 1 5.0 -4.0
Digit Span 6 5.0 1.0
Letter-Number Sequencing 8 5.0 3.0
PS Subtests
Digit Symbol-Coding 9 B.O 1.0
Sym bol Search 7 8.0 1.0
Digit S~m!;1QI-!:oding. OpliQnal Prol;edures

Incidental learning - Pairing 6
Incidental learning - Recall 6
Oualitative Features
Digit Span (Forward)
-4-2-7-3-1 repeated as "4-3-2-7-1"
-6-1-9-4-7-3 repeated as "6-9-1-4-7-3"
-5-9-1-7-4-2-8 repeated as "5-9-7-4-2-1-8"
Digit Span (Backward)
-6-2-9 repeated as "9_6_2"
Arithmetic - careless errors such as $4.00 plus $5.00 = "$8.00"
- apparently lacks mastery of basic multiplication and division
Arithmetic subtest suggest that she lacks mastery of basic mathematical operations and con-
cepts. These findings confirm the last two a priori hypotheses.
EM.'s WAIS-III protocol contains one result that was not predicted. The Verbal Compre-
hension Index is 21 points higher than the Perceptual Organization Index. The practical impli-
cation of this finding is that EM. demonstrates her intelligence more effectively via the ex-
pression of previously learned and stored verbal information than by perceptual organization
and nonverbal problem solving. Obviously, it would be incorrect to interpret the absence of a
Verbal-Performance IQ difference as evidence of similar verbal and nonverbal cognitive func-
tioning. To explain this fact, it is necessary to formulate an a posteriori hypothesis. For in-
stance, it is reasonable to assume that EM.'s fund of information, word knowledge, and un-
derstanding of verbal relationships has been artificially increased beyond the expected baseline
because of her almost three-years exposure to college courses and associated educational
experiences.
Phase Ill. The sixth and seventh interpretative steps listed in Table 2.6 address the need
to collect additional data to support the interpretative hypotheses. To accomplish this goal,
EM. was administered a battery of tests that included measures of language, memory, aca-
demic proficiency, executive functions, and constructional ability. When the battery results
are considered in conjunction with the history of two head injuries and learning problems, it
appears reasonable to assume that EM. experiences a mild cerebral dysfunction and that this
factor is responsible, at least in part, for her lower than expected IQs and the various problems
related to academic achievement (please note that this is only an inference, cannot be proven,
and should not appear in a formal report). These conclusions are further reinforced by the
presence of subtle problems on tasks involving language (e.g., naming difficulties) and men-
tal flexibility. Memory evaluation indicates that she is significantly better at retaining infor-
mation presented via the auditory modality than through the visual modality, a finding that
makes sense in light of the superiority of the Verbal Comprehension Index over the Perceptual
Organization Index. She also displayed a significant retrieval deficit during a word list memo-
rization task. Academic proficiencies are low average in reading and factual knowledge, whereas
marked deficits are noted in writing and mathematics.
Preparing the WAI5-111 Report

The WAIS-III was administered to help answer the referral question and to provide in-
sights into EM.'s cognitive and adaptive abilities. However, the scores have little meaning in
and of themselves and must be translated so that EM. and the referral source can utilize the
information. This requires drawing conclusions and the preparation of a formal report. The
following paragraphs attempt to demonstrate this process with respect to the WAIS-III results:
Reporting the IQs

On the WAIS-lII, F.M. achieved a VerballQ of 86, a Performance IQ of 86, and a Full Scale IQ
of 85. The chances that the range of scores from 81-89 includes her true Full Scale IQ are
about 95 out of a 100. Her overall achievement is classified in the low average range and is
ranked at the 16th percentile. These results appear to be reliable and valid.
Comment. It is often important to provide the sophisticated reader with traditional

IQs. EM. was referred by her academic advisor and there was a clear expectation that such
information would be forthcoming. Also, it is recognized that the Verbal IQ does not consti-
tute an accurate measure of the verbal comprehension construct and that the Verbal-Perfor-
mance IQ discrepancy is misleading. However, for practical reasons this information is pre-
sented in another section of the report. The goal of this paragraph is to communicate informa-
tion concerning EM.'s general intellectual level. Finally, the results appear valid and reliable
because EM. was cooperative, and motivated to perform and because most of the a priori
hypotheses and expectations were addressed by reviewing data from multiple sources exter-
nal to the WAIS-III (e.g., history, behavioral observations, other test results).
Subtest Analysis (Verbal Scale)

Analysis of the Verbal Scale subtests revealed average achievement on measures ofver-
bal comprehension and expression (53rd percentile). She effectively handled tasks that re-
quire the ability to orally define words. recall facts about common events. people. and places.
and identify similarities between word pairs that describe objects and concepts. She also
demonstrated an average understanding and appreciation of social conventionalities and
performed adequately on a test of auditory tracking and information processing. With respect
to the latter observation. she listened to a series of random letters and numbers and then
repeated the numbers in ascending order and the letters in alphabetical order. However; when
asked to repeat strings of orally presented digits. her score fell within the low average to
borderline range (9th percentile). This was not the result of an attention problem since behav-
ioral evidence of distractibility or fluctuating motivation was absent during task administra-
tion. Instead, it appeared to reflect auditory sequencing problems since F.M.'s repetition was
limited to five digits forward in the prescribed order. F.M.'s most impaired score emerged
during mental arithmetic calculations (lst percentile) when she became frustrated and tense
and made errors on basic problems that involved addition ($4.00 + $5.00 = "$8.00"). sub-
traction ($10.00 - $6.00 = "$5.00"). and multiplication (25 x 6 = "70").
Comment. This paragraph describes EM.'s achievement levels and, where appropri-
ate, attempts to integrate behavioral observations and subjective impressions gleaned from
item analysis and the various subtest scores. It should also be apparent that whenever possible
the emphasis is put on the composite index (Verbal Comprehension Index) over single subtest
scores. However, is this case the Working Memory Index was not unitary and it was necessary
to focus on results from the individual subtests to adequately describe E M.'s performance.
Also, the Comprehension subtest is not officially a part of the Verbal Comprehension Index,
leaving the examiner to decide whether to ignore it, report it, or subsume it under the Verbal
Comprehension Index (The latter option is justified by factor analytic research.)
Subtest Analysis (Performance Scale)

Analysis of the Performance Scale subtests revealed a high average capacity to plan and
anticipate within a social context. When confronted with sets of pictures in mixed-up order;
she quickly and efficiently rearranged them into logical stories. Low average achievement
was evidenced on measures of visual processing speed and clerical efficiency (21st percen-
tile). She was able to work under time pressures. visually scan arrays of numbers and sym-
bols, and rapidly copy stimuli or place marks in designated places on the protocol. Low
average scores were also obtained on tests ofperceptual organization and nonverbal problem
solving (9th percentile) since she demonstrated an adequate capacity to identify omitted parts
in pictures of common objects and settings, reproduce abstract patterns using colored blocks,
and recognize missing elements in a series of gridded patterns. On a measure of incidental
learning for rehearsed material, her scores were again low average. Thus, she repeatedly
copied 9 number symbol pairs for 120 seconds and later remembered 6 individual symbols
and 6 associated pairs when a surprise recall trial was conducted.
Comment. This paragraph describes EM.'s achievement levels and puts emphasis on
the Processing Speed and Perceptual Organization Indexes. However, specific interpretations
are made for the Picture Arrangement subtest and the Digit Symbol-Incidental Learning pro-
cedure. These single subtest interpretations are necessary because neither measure contrib-
utes to the composite index scores within the Performance Scale.
Overall Conclusions
Overall, F.M. performed within the low average range of intelligence as evidenced by a Full
Scale IQ of85 (16th percentile). Equalfunctioning on the Verbal and Performance Scales was
suggested, but analysis of subtest scores lead to a different conclusion. Specifically, scores on
measures of verbal comprehension and expression (53rd percentile) were superior to those on
tasks ofprocessing speed and clerical efficiency (21st percentile) as well as those that reflect
perceptual organization and nonverbal problem solving (9th percentile).
F.M. displayed a satisfactory appreciation of social conventionalities and her ability to
plan and anticipate within a social context was solidly average. She also demonstrated nor-
mal attention-concentration and auditory tracking skills. However, difficulty with number
processing was noted during a digit repetition task when she displayed sequencing errors.
Problems were also documented on a test of mental arithmetic when she became tense and
frustrated by an apparent inability to successfully carry out even the most basic calculations.
Finally, on tests of incidental learning of rehearsed material, her scores fell within the low
average range.
Comment. These paragraphs summarize the WAIS-I1I results and provide a clarifica-
tion statement concerning the misleading Verbal-Performance IQ relationship. Many implica-
tions could be drawn from EM.'s results, but the fact that her general intelligence is in the low
average range should not be overlooked. EM. is attempting to succeed at a competitive col-
lege among students whose average entrance examination score suggests at least high average
intelligence. As a transfer student, EM. was not required to take an entrance examination.
Moreover, because of her normal verbal skills and appreciation of social conventionalities, it
is postulated that she was able to present a facade of adequacy during required pre-enrollment
and orientation activities. Apparently her problems went undetected until she was well into
her first semester of academic courses.
In most assessment situations, as with the present case, the WAIS-I1I is administered as
part of an extensive test battery. The findings for EM. were presented orally and in writing to
involved faculty, the student, and her significant others. It was decided that F.M. would leave
college at the end of the semester and in the interim she would be referred for therapeutic and
vocational counseling. Every effort was made to insure that EM. had a smooth transition from
the traditional college program to a more appropriate educational setting that emphasized job
skill acquisition. Tutoring in college level courses was terminated and remediation efforts
were focused on the development of proficiencies in basic, everyday mathematics.
SUMMARY
The WAIS-III is a psychometrically sound measure of general intelligence for individuals

who are 16 to 89 years of age. Updated norms and enhanced clinical utility make the measure
appropriate for use in a wide range of clinical situations.
Improvements in the stimulus material and test protocol make the measure more user-
friendly for both the examinee and the examiner. However, a strong foundation in psychologi-
cal assessment remains a prerequisite for test use and examiner precision and accuracy are
necessary for proper administration and scoring.
As suggested in this chapter, interpreting the WAIS-III is a challenging task given the
volume of client data generated. Nevertheless, with implementation of the hypothesis-driven
interpretive framework, meaningful information that is or is not consistent with other client
characteristics can be extracted.
SUGGESTED READINGS
Kamphaus, R. W. (1993). Clinical assessment of children's intelligence: A handbook for professional practice.
Boston: Allyn and Bacon.
LoBello, S. G., Thompson, A. P., & Evani. V. (1998). Supplementary WAIS-lII tables for determining subtest strengths
and weaknesses. Journal of Psychoeducational Assessment. 16, 196-200.
Sattler, J. M., & Ryan, J. J. (1999). Assessment of children. revised and updated third edition: WAlS·/ll supplement.
San Diego, CA: Sattler.
The Psychological Corporation. (1997). WAlS-/ll-WMs-/lltechnical manual. San Antonio, TX: Author.
Wechsler, D. (1997). WAls-J1I Administration and scoring manual. San Antonio, TX: The Psychological Association.

Wechsler Adult Intelligence Scale-III

Uploaded by

Copyright:

Available Formats

Wechsler Adult Intelligence Scale-III

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Wechsler Adult Intelligence Scale-III

Uploaded by

Copyright:

Available Formats

2

Wechsler Adult Intelligence Scale-III

Joseph J. Ryan and Shane J. Lopez

TEST CONSTRUCTION AND DEVELOPMENT

Structure of the WAI5-111

Enhancing Clinical Utility/Strengthening the Theoretical Basis

Table 2.1. WAIS-III Subtests, Optional Procedures, IQs, and Indexes

Table 2.1. Continued

Table 2.2. Demographic Characteristics of the WAIS-II1 Standardization Sample

Extensive Reliability and Validity Data

Standard Error of Measurement

Applications of the WAIS-III

Test User Qualifications

Establishing and Maintaining Rapport

Scoring the WAIS-III

Analysis of WAI5-m Results

Table 2.3. Illustrative WAIS-III Results

Picture Completion 8 8.2 -0.2

Table 2.4_ Deviations from the Mean of Six

Picture Completion 3.93

Examining Qualitative Information: Detecting Errors

Hypothesis-Driven Approach to Interpreting the WAIS-III

Table 2.6. Hypothesis-Driven Approach to Interpreting the WAIS-lII'

Hypothesis-Driven Interpretive Approach

as a toddler caused the learning problems. A cause-and-effect relationship cannot be inferred.)

Table 2.7. EM.'s WAIS-III Results

Verbal Subtests SCi!h,~d SI;Qrl: ~ Difference

Vocabulary 12 7.7 4.3

Perform anS;I: Subtl:~l

Digit S~m!;1QI-!:oding. OpliQnal Prol;edures

Preparing the WAI5-111 Report

Reporting the IQs

Comment. It is often important to provide the sophisticated reader with traditional

Subtest Analysis (Verbal Scale)

Subtest Analysis (Performance Scale)

The WAIS-III is a psychometrically sound measure of general intelligence for individuals

You might also like