Vineland Adaptive Behavior Scales

Dona G. Icabone, Western Michigan University

Background Information
Authors
Sarah S. Sparrow, Associate Professor of Psychology and Chief
Psychologist at the Child Study Center at Yale University, David A.
Balla, (dec.), Research Associate in Department of Psychology and
Associate Professor of Psychology at the Child Study Center at Yale
University and Domenic V. Cicchetti, Senior Research Psychologist and
Biostatistician in the Psychology Service at Veterans Administration
Medical Center in West Haven Connecticut and Research Scientist,
Department of Psychiatry at Yale University School of Medicine.

Publisher
American Guidance Service, Circle Pines, MN 55014-1796

Price
Complete Starter Set for Survey Form, Expanded $179.95
Form, and Classroom Edition

Copyright Dates and Revisions

The Vineland Adaptive Behavior Scales (VABS) is comprised of the
Interview Edition (Survey and Expanded Forms) (Sparrow, Balla, &
Cicchetti, 1984a,b) and Classroom Edition (Sparrow, Balla, & Cicchetti,
1985). The VABS is a revision of the Vineland Social Maturity Scale
(VSMS). The VSMS was first published in 1935 (Doll, 1935) and
revised in 1965 (Doll, 1963).

Groups for Whom the Instrument is Intended

The VABS are for use with individuals from birth to 18 years, 11
months or low-functioning adults. Both forms of the Interview Edition
cover this age span while the Classroom Edition is for children between
the ages of 3 years and 12 years, 11 months. The VABS are applicable to
those with or without disabilities.

DIAGNOSTIQUE 24(1-4),257-273 (1998-1999)

257
258 DIAGNOSTIQUE VOL. 24, NO. 1-4, 1998-1999

Forms of the Instrument

Three non-equivalent versions comprise the Vineland Adaptive
Behavior Scales. These are the two forms of the Interview Edition: (1)
Interview Edition-Survey Form and (2) Interview Edition-Expanded
Form; and, the Classroom Edition. Although all require questions to be
answered by a person most familiar with the child, the Interview Edition
uses a semi-structured interview format with the parent or primary care-
giver with no direct observation of the child by the interviewer required.
The Classroom Edition, on the other hand, is a questionnaire to be
answered by the teacher.

Purpose and Recommended Uses

According to the authors, the purpose of the VABS is to provide a
general assessment of "personal and social sufficiency of individuals from
birth through adulthood" (Sparrow et a1., 1984a, p.l l) for determining
areas of strength and weakness.
The Interview Edition provides an assessment of adaptive
behavior in a semi-structured interview format. The Survey
form provides a general assessment while The Expanded Form
offers a more comprehensive one. The results of the Expanded
Form can be used to prepare individual educational, habilita-
tive, or treatment programs. It can be administered indepen-
dently or following the Survey Form to provide more in-depth
information about deficits suggested in the Survey Form.
The Classroom Edition provides an assessment of adaptive
behavior in the classroom and includes items from the Survey
and Expanded Forms plus items related to academic function-
ing.
The Vineland Adaptive Behavior Scales are applicable when-
ever an assessment of an individual's daily functioning is
required. The scales serve uses in a variety of clinical, educa-
tion, or research settings. Perhaps the major clinical use to
which the Vineland Adaptive Behavior Scales will be applied is
as a major or ancillary diagnostic tool. [since]...deficits in
adaptive behavior, as well as intelligence, must be substantiat-
ed before an individual is classified as mentally retarded. [The
VABS]...is not limited to use with the mentally retarded...[It
may be used] ...with individuals who have other disabilities,
to determine levels of adaptive behavior and the extent to
VINELAND ADAPTIVE BEHAVIOR SCALES 259

which the affect daily functioning (Sparrow, et al., 1984a,

p.4).
The VABS can also be used for program planning.
The Survey Form, Expanded Form and Classroom Edition
indicate strengths and weakness in specific areas of adaptive
behavior and ...All three versions can be used to monitor
progress and evaluate ...success at completion. The
Expanded Form was specifically developed to provide detailed
information about the prerequisite skills of adaptive behaviors
and offers step-by-step guidelines for preparing an individual
program (Sparrow et al., 1984a, p.4).
A third use is research. "Because the Vineland does not require the
presence of the individual being assessed, it is useful for research about
mental and physical disabilities, infant development, and parent-child
relationships....specific research applications...are:
I. to assess the effects of various treatments or clinical interventions
upon levels of adaptive functioning.
II. to determine the relationship of adaptive behavior to levels of clin-
ical, cognitive, or educational functioning; and,
III. to gather information in longitudinal studies in which adaptive
functioning is a variable of interest. Because the Vineland can be
used with individuals from birth to adulthood, information gath-
ered...can be used throughout the duration of longitudinal studies"
(Sparrow, 1984a, p. 5).
Note: In addition, The Vineland Social-Emotional Early
Childhood Scales assess social and emotional functioning in children
from birth to 5 years, 11 months. Normative scores are provided for three
scales, Interpersonal Relationships, Play and Leisure TIme, and Coping
Skills. Scores may be used to develop educational and treatment pro-
grams.

Dimensions Measured
The domains of Communication, Daily Living Skills, Socialization,
and Motor Skills are measured in all three versions. In the
Communication Domain, subdomains of receptive, expressive and writ-
ten language are measured. The Daily Living Skills Domain includes sub-
domains of personal, domestic and community skills. The Socialization
Domain addresses interpersonal relationships, play and leisure time, and
coping skills subdomains. The Motor Skills domain measures both gross
260 DIAGNOSTIQUE VOL. 24, NO. 1-4, 1998-1999

and fine motor skill subdomains. In addition, the Survey and Expanded
Forms include a Maladaptive Behavior domain (with no subdomains),
the administration of which is optional.

Administration
The average administration time for the Survey Form is 20-60 min-
utes: for the Expanded form is 60-90 minutes; and for the Classroom
Edition is 20 minutes.
For the Survey and Expanded Forms, when used with individuals
without disabilities, the starting point is the item keyed in the record
booklet to the individual's chronological age. Chronological ages are not
rounded UPj thus, for individuals between 5-0-0 and 5-11-30, the inter-
viewer would use the starting point of age 5. If, for this population,
deficits are suspected in one or more domains, a lower starting is suggest-
ed. (Note: no guidelines are given for determining this lower starting
point.) All domains administered subsequent to a lower starting point
use this lower starting point. When used with individuals with mental
retardation or other disabilities, the starting point is based on the indi-
vidual's mental or social age as determined by other tests, or a best esti-
mate of the individual's functioning level. Generally, one can review the
items around the starting point and start the interview with a general
question which might include the starting point item, but not necessari-
ly be specific to it.
For each adaptive behavior domain, a basal is established consisting
of the highest seven consecutive items scored 2 while a ceiling is deter-
mined by the lowest of seven consecutive items scored O. In some cases,
no basal or ceiling will be established. Items before the starting point may
be used to establish a basal. In the case of two basals and two ceilings, the
higher basal and lower ceiling are used for scoring purposes. Because the
interviewer uses a general questioning technique rather than asking the
respondent the exact items, in order, the determination of a basal and
ceiling does not follow the linear path of other standardized tests. Note
that all items are administered in the Maladaptive Behavior Domain
In the Classroom Edition, every item must be scored by the teacher,
therefore, there are no basals and ceilings to be determined.
For all three versions, item scores reflect whether or not the indi-
vidual performs the activity described i.e., a score of 2 ="yes, usually", 1
= "sometimes or partially", and 0 = "no, never." For the Survey and
Expanded Forms, scores of N ( indicating whether the individual has no
 VINELAND ADAPTIVE BEHAVIOR SCALES 261

opportunity to perform the activity) and DK (indicating that the respon-

dent has no knowledge if the individual can perform the activity), are
also provided.
Responses are recorded on different forms for each of the three ver-
sions. In the Survey Form, responses and informal observations by the
interviewer are recorded in the Survey Form Record Booklet. This booklet
also contains a score summary page for recording and profiling derived
scores. In the Expanded Form, the interviewer records item scores in the
Expanded Form Item Booklet. After completing the interview, the inter-
viewer completes the Expanded Form Score Summary and Profile Booklet.
For the Classroom Edition, teachers respond to items printed in the
Classroom Edition Questionnaire Booklet which contains the items and a
score summary and profile page.

Summation of Data
Procedures
After the items on the Survey Form are administered, the inter-
viewer assigns a score of "2" to all items below the highest basal and a
score of "0" to all items above the lowest ceiling. The sum of item scores
for each subdomain page is then recorded and the subdomain scores are
totaled. The number of "Ns" and "DK's" are likewise totaled. If five or
more "OK's" are recorded for a domain, the test is not scored and it is sug-
gested that the domain be readministered with a respondent who has
more information of the individual's activities.
After the items on the Expanded Form are administered, cluster raw
scores are summed and converted into subdomain raw scores with all
"Ns" and DK's" assigned a value of 1. The sum of Domain standard scores
may be prorated to achieve the Adaptive Behavior Composite. Prorated
scores should be noted and only used for good reason (e.g., the language
domain may be prorated for a child with severe hearing impairments.)

Types of Scores Available

A variety of derived scores replace the Social Age and social
Quotient scores found in the original Vineland Social maturity Scale. For
the adaptive behavior domains and Adaptive Behavior Composite, stan-
dard scores (mean of 100, standard deviation of 15), percentile ranks, sta-
nines, adaptive levels and age equivalents may be derived. For the sub-
domains, adaptive levels and age equivalents are available. For the
262 DIAGNOSTIQUE Val. 24, NO. 1-4, 1998-1999

Maladaptive Behavior Domain, maladaptive levels are derived for Part 1

only.
For the seven supplementary norm groups of children with disabil-
ities, the following derived scores are supplied: supplementary norm
group percentile ranks and supplementary norm group adaptive levels for
the adaptive behavior domains and Adaptive Behavior Composite. For
the subdomains, supplementary norm group adaptive levels are derived.
for the Maladaptive Behavior domain, maladaptive levels for part 1 and
for part 1 and 2 combined are available.

Interpretation of Scores
The major basis for interpretation of performance on the
Survey Form and the Expanded Form lies in first determining
the derived scores for the Adaptive Behavior Composite,
domains, and subdomains, and then comparing performance
on the domains. The Expanded Form provides, in addition, a
method for interpreting performance at a more specific level
by indicating deficits in performance in clusters of items
(Sparrow et al., 1984a, p.109).
Three major means of investigating domain standard scores are pro-
vided. These include (a) differences between individual domain standard
scores and the mean domain standard score; (b) comparisons of pairs of
domain standard scores; and, (c) difference between highest and lowest
domain standard scores. Tables are provided for these investigations.
Statistically significant differences are provided as are unusual differ-
ences, which are defined as "...those observed to occur in an infrequent
degree in the standardization sample or supplementary norm groups
(Sparrow et al, 1984a, p. 109).
The Expanded Form includes a Summary and Profile Booklet
which allows for the empirical identification of clusters of items and for
easy transfer of information to the Program Planning Profile Form.
In addition, case study examples are provided for scoring and inter-
preting data based on both the standardized sample and the supplernen-
tary norms. These case studies also cover interpreting the VABS
Extended Form scores as part of a battery of other (often intelligence)
tests.
Attention is also paid to interpreting the scores of the Classroom
Edition in relation to the same child's scores on the Survey and
Expanded Forms. A form called Report to Parents is also available to pro-
VINELAND ADAPTIVE BEHAVIOR SCALES 263

vide graphic representations and non-technical interpretations of a

child's scores.

Standardization
Sampling Procedures and Sample Characteristics
For the Survey and Expanded Forms, the national sample con,
tained 3000 individuals with about one hundred in each of 30 age groups
between birth and 18 years, 11 months. The same 3000 individuals pro'
vided the norms for both forms. These individuals represented stratifica-
tion variables of the 1980 U.S. Census: sex, race or ethnic group, com,
munity size, geographic region, and parental level of education. Several
additional national samples provided supplementary norms for the
Survey and Expanded Forms: 1,050 ambulatory and nonambulatory indi-
viduals with mental retardation, 18 years or older, who were clients of
residential institutions; 800 ambulatory individuals with mental retarda-
cion, 18 years of age or older, who were clients of residential facilities;
250 nonambulatory individuals with mental retardation, 18 years of age
or older, who were clients of residential facilities; 100 individuals with
mental retardation, 18 years of age or older, who were clients of nonres-
idential facilities; 150 children with emotional impairments, 9 years of
age through 15 years, 6 months, who were clients of residential facilities;
200 children with visual impairments, 6 years of age through 12 years, 11
months, who were clients of residential facilities; and 300 children with
hearing impairments, 6 years of age through 12 years, 11 months, who
were clients of residential facilities.
For the Classroom Edition, a national sample of 3000 students ages
3 years through 12 years, 11 months provided the norms. The same strat-
ification variables used with the Survey and Expanded forms were used
with the Classroom Edition. It is unclear from the manual if there is
overlap between the original 3000 individuals used in the norming of the
Classroom Edition and those used in norming the Survey and Expanded
Forms.

Sample selection
For the national standardization, only the Survey form was admin-
istered. Subjects were selected from four geographic regions defined by
the 1980 U.S. census: Northeast North Central, South, and West with
the percentage from each region closely matching the percentage of the
264 DIAGNOSTIQUE VOL. 24, NO. 1-4, 1998-1999

U.S. population from birth to 18~11 living in each region. Within each
region, a standardization coordinator was hired for each community. This
coordinator nominated schools for inclusion in the program, making sure
to have representatives of preschools and day-care centers along with
school age children from public, parochial, and private schools.
Consent forms were distributed to parents/caregivers and they were
asked to supply relevant information about themselves and the indivtd-
ual to be rated. A total of 21,876 forms were returned and the sample of
3,000 was randomly selected by computer to correspond to 1980 census
figures for age, sex, community size, geographic location, parental educa-
tion level, and racial or ethnic category. For each "slot", one primary sub-
ject and two alternatives were chosen. When a particular "class" of indi-
viduals was still not represented, coordinators were requested to locate
individual children who fit the description.
The standardization coordinator at each site selected interviewers
who met the following requirements: backgrounds in psychology, child
development, education, or social work; basic knowledge of testing and
experience in the administration of individual assessment devices; and,
personality and interest in children necessary to conduct an assessment.
Extensive training sessions were conducted by the senior author or psy~
chologists personally trained by the senior author. Training consisted of
demonstrating and practicing the interview method to be used and learn-
ing how to apply the scoring criteria.
The Survey Form was administered twice to parents/caregiver of
484 individuals who were between 6 months and 18 years, 11 months.
The same interviewer readministered the entire scale with the same par-
ent/caregiver after a 2~4 week interval (mean of 17 days). The sample
was divided into six age groups: 0~6 through 2~11; 3~O through 4~11; 5~0
through 6~ 11i 7.o through 9~ 11i 1O~O through 13~ 11i 14~0 through 18~ 11

Reliability
According to the authors, all reliability data for both the Survey
and Expanded Forms are based on the Survey Form which was the only
form administered during the standardization program. Rasch-Wright
item calibration estimates permitted the generation of normative data for
the Expanded Form. Reliability data for the Classroom Edition was
reported only for interitem consistency. Coefficient alpha was used to
estimate this and was obtained by the administration of the test and
based on the consistency of responses to all items on the test. Thus, the
VINELAND ADAPTIVE BEHAVIOR SCALES 265

information below is based only on the Survey Form and applied also to
the Expanded form

Test-Retest Reliability
The majority of test-retest scores for the adaptive behavior domain
and Adaptive Behavior Composite standard scores fall in the .80's and
.90's across the six age groups. Across all age groups, except 5-0 through
6-II, the socialization domain had the lowest coefficients, ranging from
.76-.8. Across the total group, the lowest coefficients were socialization
(.81) and Motor Skills (.81). For the total sample, average differences for
the domains and the Adaptive Behavior Composite ranged from -0.9 to
2.0 standard score units. Or from one-sixteenth to one-eighth of a stan-
dard deviation. Estimated or stepped-up coefficients were not computed
for the Expanded Form.
The coefficients for the Maladaptive Behavior Domain are in the
.80s. These apply to both the Survey and Expanded Forms since the
domain is identical in each Form.

Internal Consistency Reliability (Based on Survey Form

Standardization data)
Correlations between odd and even items of each adaptive behav-
ior domain for each of 15 age groups plus seven supplementary groups
were computed. The Rasch-Wright latent-trait item calibration proce-
dure was used to correct for artificial inflation which can occur when
basals and ceilings are used. Rasch-Wright W-ability scale scores were
computed for raw scores in each domain and subdomain with the
Spearman-Brown applied to correct for full test length. Split-half relia-
bility coefficients for the Adaptive Behavior Composite were computed
using Guilford's formula for the reliability of a composite.
As reported by the authors, the median domain coefficients and the
range across age groups were: Communication (Median=.89; range= .73-
.94); Daily Living (Median=.90; range =.83-.92); Socialization
(Medtan-.Sc: range=. 78-.94)j Motor Skills (Median=.83j range=.70-
.95); Adaptive Behavior Composite (Median=.94j range=.89-.98)j
Maladaptive Behavior (Medians.So: range =.77-.88)
The authors indicate that the split-half reliability for the domains
are "quite satisfactory", for the Composite "excellent", and for
Maladaptive Behavior "quite adequate." The subdomain coefficients are
not as high, with the authors calling them "adequate."
266 DIAGNOSTIQUE VOL. 24, NO. 1-4, 1998-1999

Stepped-up estimates of the split-half reliability coefficients for the

Expanded Form as computed for the domains, Adaptive Behavior
Composite and subdomains across the 15 age groups were:
Communication (Median=.94; range= .84-.97); Daily Living
(Median=.95; range =.92-.96); Socialization (Median=.92; range=.88-
.97); Motor Skills (Median=.91; range=.83-.97); Adaptive Behavior
Composite (Median=.97; range=.94-.99); Maladaptive Behavior coeffi-
cients are the same as for the Survey Form (Median=.86; range =.77-.88)
since the items are the same in each form.
The authors indicate that the split-half reliability coefficients for
the domains on the Expanded Form are "very good", for the Composite
"excellent", and for Maladaptive Behavior "quite adequate." The subdo-
main coefficients again are not as high as the others.

Standard Errors of Measurement

Standard errors of measurement (SEM) were obtained from split-
half reliability coefficients determined in the national standardization
process. For the Survey Form, SEM were computed for the four domains
and the Adaptive Behavior Composite using standard score units with a
standard deviation of 15 across 15 age groupings. The SEM for the
Communication domain range from 3.6-7.8 (Mean=5.4); for the Daily
Living Skills domain, Mean =5.0; range 4.1-6.1; for the Socialization
domain. Mean=5.6; range 3.7-7.0' for the Motor Skills domain,
Mean=6.l; range 3.4-8.2; and for the Adaptive Behavior Composite,
Mean=3.6; range 2.2-4.9.
These measurements were used to determine the bands of error, sta-
tistically significant strengths and weaknesses, and statistically signifi-
cant pairwise differences. The authors encourage use of confidence lev-
els at 90 percent, but also provide confidence levels of 68,85,90,95, and
99 percent for the Adaptive Behavior Composite and the four domains.
SEM of subdomain raw scores are also reported for the same 15 age
groups by using raw score standard deviations and split half reliability
coefficients.
The same types of data reported for the Survey Form are reported
for the Expanded Form. Summary data for the Expanded Form SEM
across 15 age groups are: Communication (Mean-vl.O, range 2.6-6.0);
Daily Living Skills (Mean=3.5; range 2.8-4.3); Socialization (Mean=4.1;
range 2.6-5.2); Motor Skills (Mean=4.5; range 2.4-6.2); Adaptive
Behavior Composite (Mean=2.6; range 1.5-3.6)As might be expected
VINELAND ADAPTIVE BEHAVIOR SCALES 267

with the increased number of items in the Expanded Form, these SEM
are smaller, with narrower ranges.
For the Classroom Edition, SEM were computed for Domain and
Adaptive Behavior Composite scores. They are based on a standard devi-
ation of 15 for standard scores and used with coefficient alpha reliabili-
ties. They were computed across nine age groups and are:
Communication (Mean=4.0; range 3.2.5.2); Daily Living Skills
(Mean=3.4; range 3.0-4.2); Socialization (Mean=3.8; range 3.1-4.4);
Motor Skills (Mean=6.6; range 6.0-7.2); Adaptive Behavior Composite
(Mean=2.4; range 2.1-2.8)

Validity
Construct Validity
In relation to the Survey and Expanded Forms, the authors present
three kinds of construct validity: 1) developmental progression of scores,
2) factor analyses of domains and subdomains, and 3) profiles of scores
for seven supplementary norm groups. The mean raw scores of 15 age
groups in the national standardization sample (on Survey Form) indicate
developmental progression of the scores and provide support for the first
kind of construct validity. Principal components analyses were conduct-
ed for domain standard scores and subdomain raw scores. For domains,
each analysis produced one significant factor across each of eight age
groups which accounted for 55.4-69.8 percent of the variance, indicating
the Adaptive Behavior Composite is an appropriate index. For subdo-
main raw scores, "The results of factor analyses, in general confirm the
organization of the subdomains into their respective domains" (Sparrow
et al., 1984a, p.44)
For supplementary norm groups, all groups obtained mean standard
scores on the domains and Adaptive Behavior Composite which were
well below the mean (l00) of the national standardized sample. For stu-
dents labeled as hearing impaired and emotionally disturbed, scores in
the Motor domain were in the average range. Standard deviations scores
between the supplementary groups in total and the standardization group
showed variability in that the standard deviations of some supplementary
groups were lower than the standardization mean score of 15 while other
supplementary groups had higher standard deviations.
268 DIAGNOSTIQUE VOL. 24, NO. 1-4, 1998-1999

Content Validity
The authors purport that the content validity "...is sup-
ported by the thorough procedures used in the development of
the items. [These included ]...a careful review of other adap-
tive behavior scalesand the literature concerning child devel-
opment. Application of strict criteria to the original pool of
[3000] items, together with subsequent field testing, national
item tryout, and national standardization, produced a set of
items that assess adaptive behavior according to our defini-
tion" (Sparrow et al., 1984a, p.46).

Criterion Related Validity

The validity of the VABS was established by correlating its scores
with those of other adaptive behavior and intelligence measures. In rela-
tion to other adaptive measures, correlations between the original
Vineland Social Maturity Scale (Social Quotient and Deviation Social
Quotient) and the current Adaptive Behavior Composite were .55 for
each quotient, indicating a moderate degree of relationship. Similar,
moderate relationships were reported by the authors between the VABS
and the Adaptive Behavior Inventory for Children (Mercer & Lewis,
1978), the VABS and the AAMD Adaptive Behavior Scale (Nihira,
Foster, Shellhaus, & Leland, 1974).

Summary and Conclusions

Major Questions and Problems
A major problem of this test is its age. It was normed in 1981 and
1982, based on the 1980 census. Technological advancements have ren-
dered some of the items irrelevant (e.g, No items in communication
domain refer to assistive or alternative communication devices for assess-
ing expressive language; shoe tying and clothes buttoning items do not
acknowledge the increased use of velcro closures which allow for task
completion.). In addition, the students currently receiving special edu-
cation services have been in more inclusionary programs than those stu-
dents upon whom the test was normed. Specific problems are addressed
under "Undesirable Features".

Desirable Features
For the Survey and Expanded Forms, a conversational format
VINELAND ADAPTIVE BEHAVIOR SCALES 269

between the examiner and a knowledgeable adult is used. Goldstein,

Smith, Waldrep, & Interbitzen, (1987) found the open,ended response
format provided more information than a rating scale format. In addi-
tion, Perry and Factor, (1989) found that parents preferred the interview
method as opposed to completing an adaptive behavior scale themselves
The inclusion of a specific report to parents using graphic and non-tech-
nicallanguage to report and interpret scores is an asset.

Standardization
The VABS has been normed on a larger number of respondents
than either the AAMR-ABS-R (Nihira, Leland, & Lambert, 1993) or
the SIB,R (Bruininks, Woodcock, Weatherman, & Hill, 1996), the two
other major tests of adaptive behavior. The Classroom Edition and
Survey Form contain about 100 students per age level, an adequate num-
ber (Evans & Bradley-lohnson, 1988).

Administration
The primary desirable feature is its reliance on responses from those
who are most familiar with the individual being assessed. The ease of
administration vis-a-vis the respondent is also a very desirable feature.
That the VABS must be administered in the form of an interview by a
qualified professional allows a rich source of clinical information (Perry
& Factor, 1989) to be gathered.

Scores
The Adaptive Behavior Composite score is most appropriate for
diagnostic and educational decision-making purposes (Roberts, McCoy,
Reiddy, & Crucitti, 1993). The authors do not attempt to infer sophisti-
cated statistical properties to maladaptive behavior scores but rather clas-
sify them into broad categories of significant, insignificant, and non,
significant. (Perry & Factor, 1989)

Reliability
Salvia and Ysseldyke (1995) suggest .9 as minimum for tests to
determine eligibility for program placement, but others suggest that .85
is sufficient. Since .90 is difficult to achieve, .85 is commonly used
(Evans & Bradley-Johnson, 1988). The interval of2Aweeks is adequate
for test-retest reliability. Regarding internal consistency, except for the
270 DIAGNOSTIQUE VOL. 24, NO. 1-4, 1998-1999

motor domain and the Adaptive Behavior Composite scores, all correla-
tion coefficients are above .85 for the Classroom Edition. For the Survey
form, approximately 25% of domain scores were below .85; for Adaptive
Behavior Composite scores, however, the .85 criterion was met or
exceeded across all age levels.

Validity
Construct validity is established by showing that a scale's scores
conform to the theory underlying the construct (Evans & Bradley~
johnson, 1988). Evidence is shown for the Classroom Edition by an
increase of scores with age and with measures of intelligence. Sparrow et
al., (1985), in reference to the Classroom Edition, state that factor analv-
ses "produced factors that were similar, in some age groups, to the orga-
nization of the subdomains (p.30). Evans & Bradley-johnson (1988),
however, noted "care is needed, when interpreting by domain and sub-
domain scores, as factors are not consistent across ages (e.g., the daily liv-
ing skills domain measures daily living skills at early age levels, but mea-
sures communication and socialization skills as well at older ages" (p,
281) The Survey Form also shows that scores increase with age and cor-
relate with results of various intelligence tests.

Undesirable features
Standardization Issues
The Classroom Edition has disproportional numbers from the
North Central Region and the South (Evans & Bradley-Johnson, 1988).
This edition defines SES according to parents' educational background
and parents with less than four years of high school are underrepresent-
ed (9.1% compared to 1980 census of 20.1%) and parents with four or
more years of college are overrepresented (30.6% compared to 1980 cen-
sus of 18.7%) (Evans & Bradley-johnson, 1988). It also has dispropor-
tionately high numbers of students from urban communities and low
numbers from rural communities (Evans & Bradley-johnson, 1988)
The fluctuation in means and SDs across age groups pose a problem,
especially with children with mental retardation (Sattler, 1989). No sep-
arate norms are presented for children with mental retardation (Perry &
Factor, 1989). Standardization samples for the Classroom Edition includ-
ed students with disabilities only if they were in a regular classroom
(Evans & Bradley-johnson, 1988), which makes these norms question-
VINELAND ADAPTIVE BEHAVIOR SCALES 271

able for use today, since so many more children with mild mental retar-
dation are receiving services in regular classrooms. Although the VABS
has supplemental norms for students with disabilities, those students
were from residential facilities.
In assessing the VABS for use with pre-school children with devel-
opmental disabilities, Roberts et a1. found, "The VABS domains ...[mea-
sure]...a single factor representing Personal Independence rather than
four separate areas of Adaptive Behavior" (1993, pg. 270).

Administration Issues
Depending solely on respondents who are familiar with the indi-
vidual being tested may cause an inflation of scores as those respondents
may feel it necessary to present the individual in his or her "best light."
Other potential difficulties are related to the examiner's skills in framing
questions, eliciting responses and scoring responses (Sattler, 1989). Thus,
the VABS requires experience and training in interview techniques for
Survey and Expanded Forms.

Scores
The means and SOs vary considerably from age to age, thus, stan-
dard errors are only rough approximations (Silverstein, 1986).
Silverstein also notes that it is also difficult to compare individuals across
ages or to perform longitudinal comparisons on the same individual.
With a sample of children with autistic impairments aged 8-18, Perry and
Factor (1989) found that one-third of their sample scored "off the scale".
They also noted that, at the low end of the distribution, "standard scores
and age equivalents are sometimes ordered differently, thus obscuring
interpretations of strengths and weaknesses" (p. 52).

Reliability
Test-retest data are not presented for the Classroom Edition (Evans
& Bradley-Johnson, 1988) and half of the coefficients for the domains on
the Survey Form were less than .85. No data are presented for interrater
reliability for the Classroom Edition, therefore caution should be used
when interpreting these results. Middleton, Keene, & Brown (1990) sug-
gest that, " The low convergent and discriminant reliability of some
scales may be a function of the relatively low reliability of the VABS" (p.
672).
272 DIAGNOSTIQUE VOl. 24, NO. 1-4, 1998-1999

Validity
No item analyses are presented for content validity. (Evans &
Bradley-johnson, 1988).

Overall Evaluation
The Vineland Adaptive Behavior Scales are adequate for use for
eligibility decisions when an adaptive behavior measure is called for. Its
major strengths, especially the richness of data collected by the semi-
structured interview method, are offset by its age.

Recommendations
If one were to use the Vineland as currently published for eligibili-
ty decisions, the Adaptive Composite Score should be used rather than
individual domain scores. It is better to base individual program goals
and objectives on observable behavior and criterion-based measurements
than on domain and subdomain results.
Several other instruments are more appropriate, based on their
more recent standardization. These include the AAMR: Adaptive
Behavior Scales-Revised: Residential and Community Scales (Nehira,
et.al., 1993), the AAMR: Adaptive Behavior Scales-Revised: School
(Lambert, et.al., 1993), and the Scales ofIndependent Behavior-Revised
(Bruininks, et.al., 1996).

References
Bruininks, R. H., Woodcock, R. W., Weatherman, R.E, & Hill, B. K. (1996). SIB-R·
Scales of Independent Behavior: Revised. Chicago: Riverside.
Doll, E. A. (1935). A genetic scale of social maturity. The American Journal of
Orthopsychiatry,S, 180-188.
Doll, E. A. (1963). Vineland Social Maturity Scale. Circle Pines: MN, American Guidance
Service.
Evans, L. O. & Bradley-Johnson, S. (1988). A review of recently developed measures of
adaptive behavior. Psychology in the Schools, 25, 276-287.
Goldsrein.D. J., Smith, K. B., Waldrep, E. L., & Inderbitzen, H. M. (1987) Comparison
of the Woodcock-Johnson Scales of Independent Behavior and the Vineland
Adaptive Behavior Scales I infant assessment. Journal of Psychoeducational Assessment,
5, 1-6.
Lambert, N., Nihira, K., & Leland, H. (1993).AAMR Adaptive Behavior Scale-Revised:
School, (2Dd ed.). Austin,TX: ProEd.
Mercer, J. R., & Lewis, J. E (1978) Adaptive Behavior Inventory for Children. New York:
The Psychological Corp.
Middleton, H. A., Keene, R. G., & Brown, G. W. (1990) Convergent and discriminant
VINELAND ADAPTIVE BEHAVIOR SCALES 273

validities of the Scales of Independent Behavior and the Revised Vineland Adaptive
Behavior Scales. American Joumal on Mental Retardation, 94(6),669-673.
Nihira, K., Foster, R., Shellhaas, N., & Leland, H. (1974). AAMD Adaptive Behavior Scale
Manual. Washington, OC: American Association on Mental Deficiency.
Nihira, K., Leland, H., & Lambert, N. (1993). AAMR Adaptive Behavior Scale-Revised:
Residential and Community, (2nd ed.). Austin, TX: ProEd.
Perry, A. & Factor, D. C (I989) Psychometric validity and clinical usefulness of the
Vineland Adaptive Behavior Scales and the AAMD Adaptive Behavior Scale for an
autistic sample. Journal of Autism and Developmental Disorders, 19( 1),41-55.
Roberts, C, McCoy, M., Reiddv, D., & Crucitti, F. (1993) A comparison of methods of
assessing adaptive behavior in pre-school children with developmental disabilities.
Australia and New Zealand]oumal of Developmental Disabilities 18(4)" 261-272.
(6th
Salvia, J., & Ysseldyke, J. E. (1995). Assessment ed.). Boston: Houghton-Mifflin Co.
Sattler, J. M. (1989) Review of the Vineland Adaptive Behavior Scale. In Conoley, J. C
Close & Kramer, J. J. The Tenth Mental Measurements Yearbook. Lincoln NB:
University of Nebraska Press.
Silverstein, A. B. (1986). Nonstandard standard scores on the Vineland Adaptive
Behavior Scales: A cautionary note. American Journal of Mental Deficiency, 9 I, 1-4.
Sparrow, S. S., Balla, D.A., & Cicchetti, D.V. (1984). Vineland Adaptive Behavior Scales:
Interview Edition, Survey Form , Circle Pines, MN: American Guidance Service.
Sparrow, S. S., Balla, D.A., & Cicchetti, D.V. (1984) Vineland Adaptive Behavior Scales:
Interview Edition, Expanded Form, Circle Pines, MN: American Guidance Service.
Sparrow, S. S., Balla, D.A., & Cicchetti, D.V. (1985) Vineland Adaptive Behavior Scales:
Classroom Edition" Circle Pines, MN: American Guidance Service.