This article was downloaded by: [University of Manitoba Libraries]

On: 16 February 2015, At: 09:34

Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

Applied Neuropsychology: Adult

Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/hapn21

A Validated Seven-Subtest Short Form for the WAIS-IV

a b c d
John E. Meyers , Margaret M. Zellinger , Tim Kockler , Mark Wagner & Ronald Mellado
e
Miller
a
Private Practice , Mililani , Hawaii
b
Private Practice , Brunswick , Maine
c
Private Practice, St. George , Utah
d
Private Practice, Philadelphia and Jenkintown, Pennsylvania
e
Psychology, Brigham Young University Hawaii , Laie , Hawaii
Published online: 26 Mar 2013.

To cite this article: John E. Meyers , Margaret M. Zellinger , Tim Kockler , Mark Wagner & Ronald Mellado Miller
(2013) A Validated Seven-Subtest Short Form for the WAIS-IV, Applied Neuropsychology: Adult, 20:4, 249-256, DOI:
10.1080/09084282.2012.710180

To link to this article: http://dx.doi.org/10.1080/09084282.2012.710180

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
 APPLIED NEUROPSYCHOLOGY: ADULT, 20: 249–256, 2013
Copyright # Taylor & Francis Group, LLC
ISSN: 2327-9095 print=2327-9109 online
DOI: 10.1080/09084282.2012.710180

A Validated Seven-Subtest Short Form for the WAIS-IV

John E. Meyers
Private Practice, Mililani, Hawaii

Margaret M. Zellinger
Private Practice, Brunswick, Maine
Downloaded by [University of Manitoba Libraries] at 09:34 16 February 2015

Tim Kockler
Private Practice, St. George, Utah

Mark Wagner
Private Practice, Philadelphia and Jenkintown, Pennsylvania

Ronald Mellado Miller

Psychology, Brigham Young University Hawaii, Laie, Hawaii

This study presents a short form of the Wechsler Adult Intelligence Scale-Fourth Edition
(WAIS-IV; Wechsler, 2008) using the subtests (Block Design, Similarities, Digit Span,
Arithmetic, Information, Coding, and Picture Completion) suggested by Ward (1990).
These seven subtests were used to predict the full WAIS-IV Full-Scale IQ, as well as
the Verbal Comprehension, Perceptual Reasoning, Working Memory, and Processing
Speed Index scores. Two different data sets were used: the first consisted of 70 subjects
and the second consisted of 32 subjects. The first data set was used to create a linear
regression and the second data set was used to validate the results and compare to the
prorated score method from the WAIS-IV manual. The prorated estimated scores corre-
lated significantly with their counterparts and proved to be a better method of estimating
the Full-Scale IQ and most of the index scores, but the regression equation was better at
predicting the Processing Speed Index. The current study is consistent with the Ward
(1990) and Pilgrim, Meyers, Bayless, & Whetstone (1999) studies and represents a reliable
and valid way of assessing intellectual functioning in an abbreviated format.

Key words: intelligence tests, neuropsychological tests, psychological tests, psychometrics,

Wechsler Adult Intelligence Scale

INTRODUCTION opposed to the trends for reimbursement, which require

increasing efficiency in test selection, administration,
As the field of neuropsychology becomes more sophisti- and reporting. Although an understanding of the
cated, the number and complexity of available tests have patient’s intellectual ability is helpful for interpreting
expanded exponentially. This trend is diametrically their performance across a battery of neuropsychologi-
cal tests, efforts have been made over the years to
Address correspondence to John E. Meyers, Meyers Neuropsycho-
shorten the intelligence measures so that the limited
logical Services, 94-553 Alapoai St. #162, Mililani, HI 96789. E-mail: testing time and patient energy could be spent on tests
[email protected] that are more relevant to diagnosing and measuring
 250 MEYERS ET AL.

cognitive functions related to central nervous system the assessment of brain injury (Callahan, Schopp, &
disorders. Johnstone, 1997) and in psychiatric evaluations
Strauss, Sherman, and Spreen (2006, p. 286) indicate (Benedict, Schretlen, & Bobholz, 1992). Due to the
that the seven-subtest short form (Ryan & Ward, 1999) popularity of the Ward seven-subtest model (Strauss
is popular in clinical use. Some limitations to the use of et al., 2006), it was decided to focus this study on the
any short-form IQ scores have been explored in other seven subtests identified by Ward.
publications, such as Kaufman and Kaufman (2001); The fourth edition of the Wechsler Adult Intelligence
they do not recommend the use of short-form IQ in Scale (WAIS-IV) was published in 2008. There are no
psychoeducational evaluations, and Pierson, Kilmer, longer VIQ or PIQ scores, but the Verbal Comprehen-
Rothlisberg and McIntosh (2012) do not recommend sion Index (VCI), Perceptual Reasoning Index (PRI;
short-form IQs be used in assessing talented and gifted previously Perceptual Organization Index), Working
individuals. In neuropsychological assessment, Lezak, Memory Index (WMI), and Processing Speed Index
Howieson, and Loring (2004) indicate that, ‘‘In sum, (PSI) from the WAIS-III were carried forward to the
‘IQ’ as a score is inherently meaningless and not WAIS-IV. The WAIS-IV technical manual (Wechsler,
infrequently misleading as well’’ (p. 22). Thus, an IQ Coalson, & Raiford, 2008, p. 75) indicates that the vari-
Downloaded by [University of Manitoba Libraries] at 09:34 16 February 2015

score in some cases may be helpful, and in other cases, ous index scores correlate r ¼ .84–.91 with their
it may not be useful; this determination is up to the WAIS-III counterparts, thus indicating good similarity
clinician to decide. The use of short-form measures, as between the WAIS-III and WAIS-IV. The core WAIS-IV
Strauss et al. point out, is popular in neuropsychology, subtests are: Block Design, Similarities, Digit Span,
where an IQ score is less important but can be informa- Matrix Reasoning, Vocabulary, Arithmetic, Symbol
tive in a clinical setting. IQ is considered to be generally a Search, Visual Puzzles, Information, and Coding. The
measure of academic potential (Neisser et al., 1996). manual indicates that the testing time is 60 to 90 min
In recognition of this need to occasionally use a for the core battery administration. There are also five
short-form IQ, Ward (1990) developed a seven-subtest supplemental subtests: Letter–Number Sequencing, Fig-
short form for the Wechsler Adult Intelligence Scale- ure Weights, Comprehension, Cancellation, and Picture
Revised (WAIS-R), which had correlations of .97, .96, Completion.
and .98 with the Verbal IQ (VIQ), Performance IQ With the introduction of the WAIS-IV and the popu-
(PIQ), and Full-Scale IQ (FSIQ), respectively, of their larity of the seven-subtest short form in the previous
full-length WAIS-R counterparts. The seven-subtest versions of the WAIS-R and WAIS-III, it is a natural
form consists of the Block Design, Similarities, Digit extension of the previous research to asses this previously
Span, Arithmetic, Information, Coding, and Picture established short-form version with the new WAIS-IV.
Completion subtests administered under standard ins- The current study involved comparison of a regression-
tructions. The scaled scores were then weighted: VIQ based method and a prorated scores method for estimat-
raw ¼ 2
(Information þ Similarities) þ (Digit Span þ ing full WAIS-IV scores from the Ward (1990)
Arithmetic). PIQ raw ¼ 2
(Picture Completion þ Block seven-subtest short form. Because the WAIS-IV does
Design) þ (Coding). FSIQ raw was the sum of the VIQ not contain a VIQ and PIQ, but instead uses the factor
raw and the PIQ raw. Scores were then looked up in scales (VCI, PRI, WMI, and PSI), prediction of the
the manual as usual to obtain the IQ scores. Pilgrim, VIQ and PIQ scores is not possible. However, the factor
Meyers, Bayless, and Whetstone (1999) mirrored Ward’s scales and the FSIQ could be predicted. To date, there
research by validating the seven-subtest form with the have been no published studies regarding any short-form
WAIS-Third Edition (WAIS-III; Wechsler, 1997), show- version of the WAIS-IV.
ing nearly identical correlations of .97, .95, and .98 with The advantage of using a seven-subtest short form is a
the VIQ, PIQ, and FSIQ, respectively, with their simple matter of time allocation. In this time of limited
full-length WAIS-III counterparts. The validity of the insurance payment, time that would have been used to
use of the weighted algorithm for the short-form complete the full WAIS-IV could be used to directly assess
WAIS-R and WAIS-III has fared well when compared other cognitive concerns. The core WAIS-IV takes 60 to
with other methods used to estimate IQ scores from 90 min to complete, whereas the seven-subtest short form
the shortened Wechsler tests with both normal stan- takes 30 to 45 min to complete, thus saving half the time.
dardization samples and a variety of clinical samples The time savings of 30 to 45 min can then be put to use in
(Hilsabeck, Schrager, & Gouvier, 1999; Iverson, Myers, more specific neuropsychological assessment without add-
& Adams, 1997; Kulas & Axelrod, 2002). Other research ing additional time and fatigue to the assessment. This
(e.g., Axelrod, Ryan, & Ward, 2001) employed a pro- may be particularly useful for individuals with more
rated score to estimate IQ and index scores. The Ward severe injuries who may not have the stamina for a lengthy
seven-subtest short form has been studied by other assessment. The seven-subtest short form has already been
researchers, who found it to be particularly useful in shown to be effective in assessing brain injury (Callahan
 VALIDATED SEVEN-SUBTEST SHORT FORM FOR THE WAIS-IV 251

et al., 1997) and in assessing psychiatric populations Central to Midwest region; Author 2 contributed 64 from
(Benedict et al., 1992). By establishing the usefulness of the East coast; Author 3 contributed 18 from the Western
the seven-subtest method, this would pave the way region; and Author 4 contributed 6 from the Eastern
for replication of the Callahan et al. and Benedict et al. United States. The data from Author 2 and Author 4
studies. Thus, the development of a seven-subtest form were combined as Group 1, and the remaining subjects
of the WAIS-IV has many potential advantages. from Authors 1 and 3 were combined as Group 2.
Several models have been used to assess ‘‘goodness of
fit’’ for various prediction methods used with the
WAIS-III seven-subtest short form: (1) the correlation Group 1. Seventy participants were included in Group
between the full-test score and the short-form score; (2) 1 (Mage ¼ 39.2 years, SD ¼ 19.1; Meducation ¼ 13.5
the discrepancy between the scores (measured in stan- years, SD ¼ 3.3). Fifty-four (77.1%) of the individuals
dard errors of measurement [SEM]); and (3) whether were right-handed, and 16 (22.9%) were left-handed; 66
the original score and the short-form score are in the (94.3%) were Caucasian and 4 (5.7%) were not; 39
same diagnostic classification range. Iverson et al. (55.7%) were female and 31 (44.3%) were male. Twelve
(1997) commented that if the predicted score was not individuals lived in a rural setting and 58 in an urban set-
Downloaded by [University of Manitoba Libraries] at 09:34 16 February 2015

within 2 SEM of the full-test score and if the estimated ting. The diagnostic makeup of this group included 24
score was in a different performance classification from with cognitive disorder-not otherwise specified (NOS),
the full-test score, the estimated score would be con- 10 with a learning disability, 3 with pervasive developmen-
sidered clinically inaccurate. Therefore, the combination tal disorder, 8 with attention-deficit disorder=attention-
of SEM and diagnostic classification change was used to deficit hyperactivity disorder (ADD=ADHD), 7 with a
determine the best model in this study. vascular=cerebrovascular accident (CVA), 1 with Lewy
body dementia, 2 with posttraumatic stress disorder, 1
with mild retardation, 1 with autoimmune disorder, 1 with
METHODS sleep disorder, 1 with Char syndrome, 1 with polysub-
stance abuse, 1 with a nonverbal learning disability, 1 with
Participants/Procedures antiphospholipid antibody syndrome, 1 with vasovagal
syncope, and 7 with primary mental health diagnosis.
Consecutive cases meeting the following criteria were
collected from the authors: Individuals had completed
the core WAIS-IV subtests plus the Picture Completion Group 2. Thirty-two participants were included in
subtest as part of a full neuropsychological assessment Group 2 (Mage ¼ 31.5 years, SD ¼ 16.8; Meducation ¼ 12.0
using the Meyers Neuropsychological Battery (MNB); years, SD ¼ 2.3). Twenty-seven (84.4%) individuals were
and there were no missing data on the WAIS-IV sub- right-handed and 5 (15.6%) were left-handed.
tests. One case was dropped due to cultural considera- Twenty-eight (87.5%) were Caucasian and 4 (12.5%)
tions (foreign-born, English as a second language, were not; 15 (46.9%) were female and 17 (53.1%) were
artificially lowering WAIS-IV). The data were collected male. Five lived in a rural setting and 27 in an urban set-
from the general clinical practices of the authors ting. The diagnostic makeup of this group was 9 with
between January 2009 and February 2012, across the cognitive disorder-NOS, 7 with a learning disorder, 4
United States, until 102 eligible cases were obtained. with ADD=ADHD, 3 with CVA, 3 with posttraumatic
In this study, the core WAIS-IV subtests plus the stress disorder, 2 with pervasive developmental disorder,
Picture Completion subtest were given as part of a full and 4 with primary behavior health diagnoses.
neuropsychological assessment using the MNB (Meyers The diagnostic makeup of these groups included com-
& Rohling, 2004; Volbrect, Meyers, & Kaster-Bundgaard, mon neuropsychological clinical cases that were referred
2000). The MNB has built-in performance validity mea- for clinical evaluation. All had been administered the
sures (PVMs, Meyers, 2011). Failure on two or more of WAIS-IV core subtests (Block Design, Similarities, Digit
these PVMs is considered an invalid data set (Meyers, Span, Matrix Reasoning, Vocabulary, Arithmetic,
Miller, Rohling, & Axelrod, 2011; Meyers & Volbrecht, Symbol Search, Visual Puzzles, Information, and
2003; Meyers, Volbrecht, Axelrod, & Reinsch-Boothby, Coding). One additional subtest was also given, Picture
2011). The Meyers Validity Index (MVI) is the total num- Completion. It was necessary to give Picture Completion
ber of failed PVMs. Only valid cases were included in the so that all the original Ward (1990) seven-subtest short-
sample. A case was excluded if the MVI was 2 or more. form subtests could be given. All subtests were adminis-
Data sets that met the above criteria were contributed tered and scored as indicated in the test manual, by
by the authors until a total of 102 data sets were achieved. doctoral-level clinicians or technicians trained and super-
Data were contributed as deidentified, preexisting clinical vised by the neuropsychologist. The FSIQ, VCI, PRI,
case data. Author 1 contributed 14 data sets from the WMI, and PSI scores were calculated as per manual
 252 MEYERS ET AL.

instructions. Data were then entered into Predictive Ana- TABLE 2

lytics Software (PASW) Statistics 18 (Statistical Package Frequency Count of the Band of Error Actual – Estimated,
Group 2 Data
for the Social Sciences), which was used for all analyses.
> þ 2 þ2 1 – 2 >2
SEM SEM SEM SEM SEM
Data Analysis R_Full-Scale IQ (FSIQ) 0 3 20 6 3
R_Verbal Comprehension (VCI) 1 5 13 5 8
The demographic makeup of each group was calculated;
R_Perceptual Reasoning (PRI) 1 8 13 3 7
the results of this demographic description were presented R_Working Memory (WMI) 0 0 32 0 0
in the ‘‘Participants=Procedure’’ section. An v2 compari- R_Processing Speed (PSI) 2 9 16 5 0
son of the urban and rural status of the participants was P_Full-Scale IQ (FSIQ) 1 6 23 2 0
conducted and showed no significant difference between P_Verbal Comprehension (VCI) 4 3 13 9 3
P_Perceptual Reasoning (PRI) 3 5 13 7 4
the two study groups, v2(1) ¼ 0.04, p ¼ .85.
P_Working Memory (WMI) 0 0 32 0 0
Next, Group 1 was selected; with this data set, a lin- P_Processing Speed (PSI) 5 14 7 5 1
ear regression was calculated using the age-scaled scores
from the Ward (1990) seven subtests to predict the FSIQ Note. SEM from page 45, Table 4.3 of Wechsler et al. (2008) for
Downloaded by [University of Manitoba Libraries] at 09:34 16 February 2015

FSIQ ¼ 2.16, VCI ¼ 2.85, PRI ¼ 3.48, WMI ¼ 3.67, PSI ¼ 4.78.
and the other index scores with 70 cases selected to help
R_ ¼ Estimate based on the regression equation; P_ ¼ Estimate based
achieve stable regression beta weights for the seven vari- on prorated score equation.
ables (Hill & Lewicki, 2007). Then, using the regression
equations developed for Group 1, the regression scores
were calculated using Group 2 data. as having the highest number of cases within þ2 SEM
Next, using the usual prorated scores method from of the full WAIS-IV score.
the WAIS-IV manual, the FSIQ and index scores were The best-methods estimates were then compared to
calculated. The formulas for the regression equations the original scores to determine the significance of the dif-
and the prorated scores are presented in Table 1. The ference between the actual and estimated scores, and the
scores were then calculated for the Group 2 data for correlations of the actual and estimated scores. Finally,
both the regression equations and the prorated scores the clinical accuracy of these scores was evaluated by
methods. Using the SEM from page 45, Table 4.3 of determining the percent of cases where the diagnostic
the WAIS-IV technical manual (Wechsler et al., 2008), classification of the score would change based on the full
the band of error was calculated for each individual in WAIS-IV versus the estimated score, the percent of cases
Group 2, using both the regression and prorated score where the short-form estimate was more than 2 SEM dif-
methods. The results of the band-of-error calculation ferent from the full-test score, and the percent of cases
are presented in Table 2. The band of error was then where there was both a change in diagnostic classification
visually inspected to determine which methods produced and the predicted and actual scores differed by more
the best ‘‘hit rate’’ for correct estimation of the FSIQ than 2 SEM (i.e., the short-form score was ‘‘clinically
and index scores. The ‘‘best methods’’ were identified inaccurate’’).

TABLE 1
Regression Beta Weights and Constants for the FSIQ and Index Scores (Derived From Group 1)

Subtest FSIQ VCI PRI WMI PSI

Similarities (SI) 0.941 2.937

Block Design (BD) 1.144 2.794 0.921
Information (IN) 1.148 2.835
Digit Span (DS) 0.859 2.835
Coding (CD) 0.997 4.063
Arithmetic (AR) 0.877 0.778 2.783
Picture Completion (PC) 0.868 1.913
Constant 32.148 43.530 45.279 43.700 49.402
R 0.987 0.981 0.930 1.000 0.936
Prorated Formulas:
Prorated, using scale scores, and applying usual ((IN þ SI þ BD ((IN þ SI)
3)=2 ((BD þ PC)
3)=2 DS þ AR CD
2
computation of index scores per the þ PC þ DS þ AR þ CD)
10)=7
WAIS-IV Manual
R .986 .980 .915 1.000 .921

FSIQ ¼ Full-Scale IQ; VCI ¼ Verbal Comprehension; PRI ¼ Perceptual Reasoning; WMI ¼ Working Memory; PSI ¼ Processing Speed.

All p < .001.
 VALIDATED SEVEN-SUBTEST SHORT FORM FOR THE WAIS-IV 253

RESULTS TABLE 3
Paired-Samples Statistics Using Group 2 Data
The equations and the regression beta weights resulting Std. Cohen’s
from the stepwise linear regression with Group 1 data, Std. Error d (Effect
using the age-scaled scores from the Ward (1990) seven Mean N Deviation Mean Size)
subtests to predict the FSIQ and the other index scores Pair Full-Scale IQ 92.6 32 16.0 2.9 0.06
are presented in the top half of Table 1. The final model 1 R_Full-Scale IQ 93.6 32 16.9 3.0
using seven subtests showed a correlation of .99 between Pair Verbal 94.3 32 17.1 3.0 0.09
the FSIQ and the regression FSIQ; the correlations 2 Comprehension
R_Verbal 96.0 32 18.2 3.2
between the index scores and their respective prorated Comprehension
versions were also quite strong (.93 for PRI to 1.0 for Pair Perceptual Reasoning 95.5 32 14.0 2.5 0.07
WMI). The results of the initial regression analysis 3 R_Perceptual 96.5 32 13.8 2.4
(Group 1 data) show that the Ward seven subtests are Reasoning
able to predict fairly accurately the FSIQ and the vari- Pair Processing Speed 92.2 32 14.2 2.5 0.14
4 R_Processing Speed 90.3 32 13.5 2.4
ous index scores of the WAIS-IV.
Downloaded by [University of Manitoba Libraries] at 09:34 16 February 2015

The bottom of Table 1 shows the formulas and results Pair Full-Scale IQ 93.9 32 16.0 2.9 0.11
of the prorated score method, which also produced a cor- 1 P_Full-Scale IQ 92.1 32 16.3 2.9
Pair Verbal 94.3 32 17.1 3.0 0.03
relation of .99 between the FSIQ and the prorated FS IQ; 2 Comprehension
the correlations between the index scores and their P_Verbal 94.8 32 17.1 3.0
respective prorated versions were also quite strong Comprehension
(.92–.98; the WMI correlation was of course 1.0 because Pair Perceptual Reasoning 95.5 32 14.0 2.5 0.04
both subtests comprising the WMI are included in the 3 P_Perceptual 96.1 32 14.5 2.6
Reasoning
Ward [1990] seven-subtest short form. The WMI is not Pair Processing Speed 92.2 32 14.2 2.5 0.19
included in further discussion). 5 P_Processing Speed 89.3 32 15.7 2.8
Both methods produced strong correlations and
would be considered reliable methods of calculating a Note. Guidelines used in interpreting effect size: negligible .19;
small ¼.20  .49; medium ¼ .50–.79; large ¼.80 (Cohen, 1988).
Ward (1990) seven-subtest IQ or composite score with
R_ ¼ Estimate based on the regression equation; P_ ¼ Estimated based
correlations within 1 point using either method. How- on prorated score equation.
ever, when examining Table 2, it can be seen that the pro-
rated score method more consistently performed within Finally, considering Iverson et al.’s (1997) criteria for
the 2 SEM range, with the exception of the PSI. For clinical inaccuracy, a visual inspection was made of
the PSI, the regression equation seems to perform better the frequency with which the original score would be
than the prorated score method.
To compare differences between regression and
prorated methods, dependent-sample t-tests were run TABLE 4
for the FSIQ, Mprorated ¼ .47 and Mregression ¼ 0.80, Paired-Samples Correlations Between Full WAIS-IV Score and
Short-Form Estimated Score Using Group 2 Data
t(31) ¼ 4.95, p < .01; VCI, Mprorated ¼ 0.56 and
Mregression ¼ 1.71, t(31) ¼ 3.82, p < .01; PRI, Weighted Regression Estimates N Correlation Sig.
Mprorated ¼ 0.63 and Mregression ¼ 1.02, t(31) ¼ 0.53, Pair 1 Full-Scale IQ & R_Full-Scale IQ 32 .99 .000
p > .05; WMI, Mprorated ¼ 0.00 and Mregression ¼ 0.20, Pair 2 Verbal Comprehension & 32 .97 .000
t(31) ¼ 2.51, p < .05; and PSI, Mprorated ¼ 2.97 and R_Verbal Comprehension
Mregression ¼ 1.95, t(31) ¼ 1.33, p > .05. Pair 3 Perceptual Reasoning & 32 .93 .000
Scores from both methods were then compared to R_Perceptual Reasoning
Pair 4 Processing Speed & 32 .93 .000
the original full WAIS-IV scores, and a Cohen’s d was R_Processing Speed
calculated to check if the difference between the actual
scores and the estimated scores was meaningful. The Prorated Scores Method N Correlation Sig.
results of this analysis (Table 3) showed that all Cohen’s
Pair 1 Full-Scale IQ & P_Full-Scale IQ 32 .99 .000
d scores for both methods were in the negligible range,
Pair 2 Verbal Comprehension & 32 .97 .000
indicating that the difference between the full WAIS-IV P_Verbal Comprehension
scores and the estimated scores based on either method Pair 3 Perceptual Reasoning & 32 .95 .000
using the seven-subtest short form were not statistically P_Perceptual Reasoning
significant nor clinically meaningful. As shown in Pair 4 Processing Speed & 32 .92 .000
P_Processing Speed
Table 4, the estimated scores using the regression
method and the prorated method showed nearly ident- R_ ¼ Estimate based on the regression equation; P_ ¼ Estimated
ical results. based on the prorated score.
 254 MEYERS ET AL.

TABLE 5 those reported for the Ward seven-subtest short form

Percent of Estimates Considered ‘Clinically Inaccurate’ for of the WAIS-III. These findings indicate that the pro-
Regression-Weighted Versus Prorated Methods of Estimation
rated scores and regression equations do appear to be
% Outside stable and are adequately predictive of their full-version
% % 2 SEM counterparts. For all accuracy measures, the most
Mean Outside 2 Changed AND Changed
accurate prediction for the short form was the prorated
Score Diff (SD) SEM Classification Classification
method of estimating the FSIQ score. The prorated
R_FSIQ –0.79 (2.92) 9 16 3 FSIQ compares favorably with Iverson et al. (1997),
P_FSIQ 0.46 (2.15) 3 9 3 who found 9% of FSIQ, 6% of VIQ, and 17% of PIQ
R_VCI –1.70 (4.44) 28 28 22
P_VCI –0.56 (4.46) 22 31 13
scores from the WAIS-R estimates would be clinically
R_PRI 1.01 (5.35) 25 31 16 inaccurate. Lange, Iverson, Viljoen, and Brink (2007)
P_PRI –0.62 (4.72) 22 34 19 found 2% to 5% of VCI, 5% to 10% of POI (the
R_PSI 1.95 (5.32) 6 38 3 WAIS-III version of PRI), and 3% to 8% of WMI scores
P_PSI 2.96 (6.11) 19 53 19 from prorated WAIS-III data met this criteria. Given
R_ ¼ Estimate based on the regression equation; P_ ¼ Estimated the generally higher correlations using the prorated
Downloaded by [University of Manitoba Libraries] at 09:34 16 February 2015

based on the prorated score. method and the low proportion of scores meeting the

Clinically inaccurate. clinically inaccurate criteria, there is very strong support
for the continued use of this short form with the pro-
classified differently using the short-form methods, rated scores method for the calculation of the FSIQ
combined with the frequency with which the original score. If index scores are also needed, the prorated score
score differed from the estimated score by more than 2 is both simpler and more clinically accurate for the VCI,
SEM (i.e., >4 points for FSIQ, >5 points for VCI, >6 and it is nearly as accurate for the PRI. For the PSI, the
points for PRI, or >9 points for PSI). As can be seen weighted regression showed clear superiority in clinical
in Table 5, the FSIQ score would be clinically inaccurate accuracy; although the correlations between the pro-
just 3% of the time using either method. The VCI would rated PSI and the actual PSI were nearly identical, the
be inaccurate 22% of the time using the regression model additional variables included in the weighted regression
but only 13% with the prorated method, and the PRI cal- made this method superior to the prorated model which
culations would be 16% inaccurate with the regression used just one variable.
model compared with 19% with the prorated method. Although the group data showed very high correla-
The PSI, however, had an inaccuracy rate of 3% with tions comparable to those found in previous research,
the weighted regression but a 19% inaccuracy rate with there was more variability in the predictive accuracy at
the prorated scores method. Neither model was consist- the individual level, especially for the VCI, PRI, and
ently superior to the other for all scores and indexes. PSI. This variability at the individual-case level has also
The prorated method showed similar or better perfor- been observed in previous research using short forms
mance on all but PSI. Given that the correlation between with the WAIS-R (Iverson et al., 1997) and the WAIS-III
the actual PSI and the regression is .93 and the prorated (Kulas & Axelrod, 2002), especially for the component
method is .92, the difference between the two methods is scores (VIQ, PIQ, and the index scores). In the current
actually small; but the prorated method is simpler, is the study, the greatest differences were observed when there
method presented in the WAIS-IV manual, and is well was a spread of 6 or more points in the age-scaled scores
known in the field of neuropsychology. for the subtests in the short form. When clinicians
observe such a spread in the individual patient’s data,
it is recommended the remainder of the subtests in the
DISCUSSION relevant indices be administered before making conclu-
sions based on profile analysis.
The results of the current study showed that it is possible Using a short form of the WAIS-IV provides many
to use the seven subtests identified by Ward (1990) as a advantages for a clinician to consider. First, abbreviated
short form for the WAIS-IV. The original Group 1 data testing is both efficient and less strenuous on the exam-
were used to create the regression equations that were inee. In situations where the referral is not related to
then validated against a second set of data (Group 2). qualification for educational services or documentation
Group 2 data were also used in the prorated score of mental retardation, the authors find the abbreviated
method for a direct comparison with the regression form to be an acceptable method to use when there are
equation method. There was a strong correlation (as more pressing clinical questions to answer. Moreover,
seen in Table 4) between the original index scores and the Ward (1990) seven-subtest scales are often part of
the estimated index scores using both regression weight- other testing batteries, such as the MNB. The MNB uses
ing and the prorated scores method, comparable with the short form as its preferred form over the long form of
 VALIDATED SEVEN-SUBTEST SHORT FORM FOR THE WAIS-IV 255

the WAIS-III=WAIS-IV. The use of a short form pro- future research. Although the authors believe the sample
vides a reliable estimate of IQ and, in most cases, index size for both the creation and validation samples was
performance; although the potential information from sufficient for their purpose in this study, certainly repli-
some subtests is lost, the time savings allows the clinician cation with a larger sample size would be appropriate.
to include a variety of more specific neuropsychological Additionally, clinical validation of this short form with
assessment instruments, thus focusing the assessment different population and clinical groups should be made.
more on neuropsychological factors. In cases where there The results of this study indicate that the Ward (1990)
is atypical variability in the short-form subtest scores, seven-subtest short form of the WAIS-IV, whether
administration of additional WAIS-IV subtests may be applied with the regression-derived scoring algorithm
clinically justified. Another obvious advantage is the presented here or a simpler prorated score method,
time savings: The seven-subtest form takes 30 to 45 min appears to retain psychometric properties similar to
to administer, which provides a 50% time savings (using the full version of the WAIS-IV.
the WAIS-IV manual’s estimate of 60 to 90 min to
complete the core WAIS-IV tests).
In the current study, the seven-subtest form was admi-
REFERENCES
Downloaded by [University of Manitoba Libraries] at 09:34 16 February 2015

nistered in the same order as in the full WAIS-IV, skipping

only the unused subtests and administering the Picture
Axelrod, B. N., Ryan, J. J., & Ward, L. C. (2001). Evaluation of
Completion as the final subtest. By using this order, the seven-subtest short forms of the Wechsler Adult Intelligence
Block Design subtest is administered first, in both the short Scale-III in a referred sample. Archives of Clinical Neuropsychology,
form and the regular form. Thus, the problem with order 16, 1–8.
effect as postulated by Thompson, Howard, and Anderson Benedict, R. H., Schretlen, D., & Bobholz, J. H. (1992). Concurrent
(1986) is not applicable to the study findings and to validity of three WAIS-R short forms in psychiatric inpatients.
Psychological Assessment, 4, 322–328.
enhancing the generalizability of the study results. There Callahan, C. D., Schopp, L., & Johnstone, B. (1997). Clinical utility
obviously are other short-form versions possible (Donnell, of the seven subtest WAIS-R short form in neuropsychological
Pliskin, Holdnack, Axelrod, & Randolph, 2007; assessment of traumatic brain injury. Archives of Clinical Neuropsy-
Reid-Arndt, Allen, & Schopp, 2011; Thompson et al.), chology, 12, 133–138.
and although these may be interesting to pursue in future Cohen, J. (1988). Statiscal power analysis for the behavioral sciences
(2nd ed.). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc.
studies, they are not the purpose of the current study. Donnell, A. J., Pliskin, N., Holdnack, J., Axelrod, B., & Randolph, C.
In this study, we used a clinical sample, although this (2007). Rapidly-administered short forms of the Wechsler Adult
sample was not the same as the standardization sample Intelligence Scale-3rd edition. Archives of Clinical Neuropsychology,
for the WAIS-IV; however, it is a sample consistent with 22(8), 917–924.
the types of cases with which this short form could be Hill, T., & Lewicki, P. (2007). STATISTICS: Methods and applications.
Tulsa, OK: StatSoft.
used. Another limitation is that the sample was mostly Hilsabeck, R. C., Schrager, D. A., & Gouvier, W. D. (1999). External
Caucasian. However, ‘‘demographically adjusted norms validation of the equivalence of two computational formulas for a
are not intended for use in psychoeducational assessment, WAIS-R seven subtest short form. Applied Neuropsychology, 6(4),
determination of intellectual disability, vocational assess- 239–242.
ment, or any other context in which the assessment goal Iverson, G. A., Myers, B., & Adams, R. L. (1997). Comparison of two
computational formulas for a WAIS-R seven-subtest short form.
is to determine absolute functional level (IQ or memory)’’ Journal of Clinical Psychology, 53, 465–470.
(Overview of Advanced Clinical Solutions for WAIS-IV and Kaufman, J. C., & Kaufman, A. S. (2001). Time for the changing of
WMS-IV, 2009, p. 7). Thus, any demographic difference the guard: A farewell to short forms of intelligence tests. Journal
is likely not of clinical significance. of Psychoeducational Assessment, 19, 245–267.
Another limitation of the use of any short form is the Kulas, S. F., & Axelrod, B. N. (2002). Comparison of seven-subtest
and Satz-Mogel short forms of the WAIS-III. Journal of Clinical
loss of some information from the original full form. Psychology, 58, 773–782.
However, in neuropsychology specifically, and in other Lange, R. T., Iverson, G. L., Viljoen, H., & Brink, J. (2007). Clinical
situations where a short form is used, the loss of infor- validation of Canadian WAIS-III index short forms in inpatient
mation is minimal given the strong correlations with neuropsychiatry and forensic psychiatry. The Clincial Neuropsychol-
the original long form. The clinician must weigh the loss ogist, 21, 434–441.
Lezak, M. D., Howieson, D. B., & Loring, D. W. (2004). Neuropsycho-
of information and the gain of time that can be more logical assessment (4th ed.). New York, NY: Oxford University
productively spent assessing specific referral questions. Press.
Further research is needed to replicate the work of Meyers, J. E. (2011). Electronic manual for the Meyers Neuropsycholo-
Callahan et al. (1997) in a brain injury sample and the gical Battery (MNB). Retrieved from http://www.meyersneuro
work of Benedict et al. (1992) in a psychiatric popu- psychological.com
Meyers, J. E., Miller, R. M., Rohling, M. L., & Axelrod, B. N. (2011).
lation, to see if the seven-subtest short form retains suf- Does the source of a forensic referral affect neuropsychological
ficient diagnostic strength as it did with the previous test performance on a standardized battery of tests? The Clinical
versions of the WAIS. This should be the subject of Neuropsychologist, 25, 477–487.
 256 MEYERS ET AL.

Meyers, J. E., & Rohling, M. L. (2004). Validation of the Meyers Short Reid-Arndt, S. A., Allen, B. J., & Schopp, L. (2011). Validation of
Battery on mild TBI patients. Archives of Clinical Neuropsychology, WAIS-III four-subtest short forms in patients with traumatic brain
19, 637–651. injury. Applied Neuropsychology, 18(4), 291–297.
Meyers, J. E., & Volbrecht, M. E. (2003). A validation of multiple Ryan, J. J., & Ward, L. C. (1999). Validity, reliability, and standard
malingering detection methods in a large clinical sample. Archives errors of measurement for two seven-subtest short forms of
of Clinical Neuropsychology, 18, 261–276. the Wechsler Adult Intelligence Scale III. Psychological Assessment,
Meyers, J. E., Volbrecht, M. E., Axelrod, B. N., & Reinsch-Boothby, 11, 207–211.
L. (2011). Embedded symptom validity tests and overall neuropsy- Strauss, E., Sherman, E. M., & Spreen, O. (2006). A compendium of
chological test performance. Archives of Clinical Neuropsychology, neuropsychological tests: Administration, norms, and commentary
26, 8–15. (3rd ed.). New York, NY: Oxford University Press.
Neisser, U., Boodoo, G., Bouchard, T. J., Boykin, A. W., Brody, N., Thompson, A. P., Howard, D., & Anderson, J. (1986). Two- and four-
Ceci, S. J., . . . Urbina, S. (1996). Intelligence: Knowns and unknowns. subtest short forms of the WAIS-R: Validity in a psychiatric sample.
American Psychologist, 51, 77–101. Canadian Journal of Behavioral Science, 18, 287–293.
Overview of Advanced Clinical Solutions for WAIS-IV and WMS-IV. Volbrecht, M., Meyers, J., & Kaster-Bundgaard, J. (2000). Neuro-
(2009). Retrieved from http://www.pearsonassessments.com/NR/ psychological outcome of head injury using a short battery. Archives
rdonlyres/D4ACCA90-D61F-4C24-BA37-61F6AC068C38/0/ of Clinical Neuropsychology, 15, 251–265.
Overview_WAISIV_WMSIV.pdf Ward, L. C. (1990). Prediction of Verbal, Performance and Full
Pierson, E. E., Kilmer, L. M., Rothlisberg, B. A., & McIntosh, Scale IQs from seven subtests of the WAIS-R. Journal of Clinical
Downloaded by [University of Manitoba Libraries] at 09:34 16 February 2015

D. E. (2012). Use of brief intelligence tests in the identification Psychology, 46, 436–440.
of giftedness. Journal of Psychoeducational Assessment, 30, Wechsler, D. (1997). WAIS-III: Administration and scoring manual.
10–24. San Antonio, TX: Pearson.
Pilgrim, B. M., Meyers, J. E., Bayless, J., &Whetstone, M. M. (1999). Wechsler, D. (2008). WAIS-IV administration and scoring manual. San
Validation of the Ward seven-subtest WAIS-III short form in a Antonio, TX: Pearson.
neuropsychological population. Applied Neuropsychology, 6, Wechsler, D., Coalson, D. L., & Raiford, S. E. (2008). WAIS-IV
243–246. technical and interpretive manual. San Antonio, TX: Pearson.