Adaptive Behavior Assessment System (2nd Edition) by

Harrison, P. L., & Oakland, T. (2002). San Antonio, TX:

Psychological Corporation
Robert D. Richardson
Salt Lake City School District
Matthew K. Burns
University of Minnesota

Purpose and Available Scores ABAS-II generates scaled scores for each of
these skill areas. Further, the AAMR (2002)
The Adaptive Behavior Assessment
concluded that adaptive behavior clusters into
System - Second Edition (ABAS-II) provides three general factors: conceptual, social, and
a comprehensive, norm referenced assessment

of adaptive behavior for individuals from practical skills that help people function in
birth to 89 years. Rating scales are provided
everyday living, and the ABAS-II has compos-
ite scores in each of these areas in addition to a
for parents, teacher/daycare providers, and
adult self-report, so an individual’s adaptive
general adaptive composite (GAC) score.

Thus, the ABAS-II seems well suited for diag-

functioning can be surveyed from multiple nosis/classification of mental retardation for
perspectives. Authors report that it may be these standard diagnostic rubrics.
used to assess adaptive skills for diagnosis and
classification of disabilities, for identification Format
of strengths and limitations, and for docu- includes five forms:
The ABAS-11
mentation of progress over time.
Estimates of adaptive behavior are perhaps
Parent/Primary Caregiver (ages 0-5),
most frequently used to assist in diagnosing
Teacher/Daycare Provider (ages 0-5 ), Parent
(ages 5-21 ), Teacher (ages 5-21 ), and Adult
mental retardation. ABAS-II is a theory driven
(ages 16 to 89). The first three of the above
test which conforms to the three prominent
mentioned forms are also available in
definitions used in the identification of mental
Spanish. The Infant and Preschool forms
retardation: the American Association on (ages 0 to 5 years) are new to the second edi-
Mental Retardation (AAMR, 1992 & 2002), tion. The items on the parent and teacher
the Diagnostic and Statistical Manual of Mental forms (ages 5-21 years) are the same as they
Disorders (4th ed. Text Revision; DSM,IV,TR, were on the previous edition; however, the
American Psychological Association, 2000), directions printed on the forms have been
and the 1997 amendments to the Individuals expanded and revised to make the rating sys-
with Disabilities Education Act. According to tem more clear and better operationalized.
the first two of these definitions (AMMR, 1992 Furthermore, the various forms are now color
& DSM-IV TR) adaptive behavior is com- coded so that an examiner is less likely to
prised of ten specific skill areas: communica- confuse information from various sources.
tion, community use, functional academics, The format of the ABAS-II follows that of
home/school living, health and safety, leisure, typical measures of adaptive behavior in that
self-care, self-direction, social, and work. the respondent rates the observed behavioral

51
Downloaded from aei.sagepub.com at Yale University Library on May 16, 2015
52

frequency of various skills, using a four-point nosed with various disabilities included in the
rubric of not able, never, sometimes, and always, normative samples range from 2.93%
grouped together into skill areas and domains. (Teacher/Daycare Provider form) to 7.17%
In addition, respondents are provided an (Adult Self-Report form). Given that esti-
option of marking the item as guessed and can mates of total school-aged population with

provide additional comments as desired. The disabilities are approximately 11% (U.S.
number of items marked as guessed is then Department of Education, 2001), these num-
compared to a normative sample to determine bers appear somewhat low.
the validity of the results. Scales in which too The administration procedures are general-
many items were guessed at, as determined by ly simple and well-defined, so standardization
exceeding the mean number of guessed items should be easily attainable, but several con-
for the norm group, result in questionable cerns remain. First, when the age of the indi-

data and it is recommended that a different vidual being rated overlaps with more than
one form, little guidance is given for which
respondent complete the scale.
Unlike many measures of adaptive behav- form to use. There is more than one appropri-
ate form for individuals who are 5 years old
ior there are no basals and ceilings to limit
testing time. All items are to be answered to (school age or preschool forms), or who are 16
to 21 years old (school age or adult).
obtain a score. One advantage of this
approach is that some individuals (such as Presumably the form for the lower age group
those with autism spectrum disorders) may would be used with lower function individu-
have uneven profiles such that a rank order- als. Fortunately, correlations between alter-
nate forms is relatively high for the general
ing of difficulty for the normative population
may have little or no bearing. Further, com-
adaptive composite (corrected r .71 to .75). =

However, correlations among skill group areas

pleting all the items also permits respondents
to complete forms independently rather than varyconsiderably (corrected r .19 to .78). =

More specific decision-making procedures

in the presence of the evaluator, which could
and data are needed.
increase freedom to deliberate and possibly
A second concern is conceptual awkward-
decrease effects from social interaction (such
ness due to the level of abstraction found in
as halo effects). However, for other cases
certain items. When rating the frequency of
(such as individuals who are non-ambulatory, an adaptive behavior, such as shoe tying, one
non-verbal, or severely mentally retarded) is not generally interested in how frequently
asking the respondent to complete the entire the behavior occurs, but rather how often it
form may be an unnecessary waste of time. occurs when the behavior is called for (or

Standardization appropriate to the circumstance). It would

not be adaptive to literally always be tying
Extensive data are provided regarding the one’s shoes. The ABAS-II solves this problem
standardization sample. The total number of by adding the phrase &dquo;as needed&dquo;. This works
participants was 7,370 and ranged from 750 smoothly for some items (like &dquo;ties shoes
(Teacher/Day Care Provider form of the sometimes as needed&dquo;). However, it can be
Infant-Preschool Form) to 1,690 (Teacher quite awkward in other cases. For example,
form of the School-Age forms). Every age &dquo;Has more than one friend?&dquo; a. &dquo;is not able&dquo;,
grouping included at least 100 participants. b. &dquo;never when needed&dquo;, c. &dquo;sometimes when
The sample appears to closely match United needed&dquo;, or d. &dquo;always when needed&dquo;. First
States census data (1999) regarding age, gen- one has to evaluate whether or not the indi-

der, race/ethnicity, education, and geographic vidual is able to have more than one friend
location. The percentage of children diag- given the right circumstances. Next one has
Downloaded from aei.sagepub.com at Yale University Library on May 16, 2015
 53

to think about when one might need more were also near, at, or exceed .90. Thus, those
than one friend (a decision which is bound to scores can be used with confidence. Some

be quite subjective). Then one has to evaluate variance in coefficients occurred for the indi-
all of the cases in which the individual need- vidual skill areas, which implied that inter-
ed a friend and determine whether or not the pretation of those scores should be done cau-
individual had filled this need (more subjec- tiously. Interrater
reliability estimates are
tive decisions). This item is abstract and con- lower, but generally exceed .80, and cross-
ceptually difficult. informant coefficients generally exceed .70.
A third concem relates to readability of the Although these scores are lower than might
items. This is particularly true for the parent be hoped for, they compare very favorably to
forms given that the educational level of par- other measures of adaptive behavior.
ents can vary considerably. The manual indi- The content validity of the ABAS-II is
cated that the forms were written at a 5th grade adequately argued through expert ratings of
level, but did not state how this reading level items, respondent comments on ease of com-
was derived. Using the Flesch-Kincaid pleting the scales, and consistency with
AAMR and DSM-IV-TR definitions. Score
Readability Formula (Kincaid & McDaniel,
differences between age groups are also
1974) the ABAS-II parent rating for preschool
had a reading level of mid-fifth grade (5.6), demonstrated and support the validity of the
with skill area subscales varying in readability data. Skill, domain, and GAC scores were
from end of third ( 3 . 7 ) to end of seventh (7.9). intercorrelated and result in coefficients that
The readability of the items on the school age range from .47 to .93. Further, confirmatory
Parent Form and the Adult Form were rated
factor analyses supported the current factor
structure, and data were provided to support
approximately one grade higher (6.7 and 6.8 the convergent and divergent validity of the
respectively) and have skill areas varying from data using measures of achievement, intelli-
mid-third grade to end of eight grade (3.7 to
gence, and behavior.
8.9). Fortunately, an altemative administration It would be difficult to find a measure of
procedure is available in that items may be read adaptive behavior with more convincing data
to the respondent, and standardized directions
for doing so are provided in the manual. It regarding clinical validity. Diagnostic utility
was examined with several studies that com-
should be noted that the readability for the
ABAS-II is not significantly higher than other
pared scores of individuals diagnosed with
various disorders including mental retarda-
adaptive behavior inventories. Nevertheless, tion, developmental delays, known biological
the examiner should be aware that this is an risk factors, physical impairments, language
issue that may need to be addressed.
disorders, pervasive developmental disorder
not otherwise specified, autism, and atten-
Reliability and Validity
tion-deficit/hyperactivity disorder (AD/HD)
The ABAS-II manual provides extensive to a matched group. Significant differences
reliability and validity data that are impres- and moderate to strong effects were found.
sive in both scope and quality. Estimates of For example, the mean score for children with
reliability are provided for internal consisten- AD/HD was particularly low in the Self-
cy, test-retest, interrater, and cross-form con- Direction skill area, with 73% having scores
sistency. Coefficients are provided for the dif greater than two standard deviations below
ferent forms, for various proficiency levels and the mean, compared to 17% in the matched
for each age grouping. Generally, coefficients control group.
for the GAC exceed .90, except for children While not typically included in standard,
less than 1 year old, and the domain scores ized measures of adaptive behavior, it could be

Downloaded from aei.sagepub.com at Yale University Library on May 16, 2015

54

valuable to examine whether or not statistical should be interpreted cautiously. Data to sup-
biases exist for population sub-groups based port the tool’s reliability and validity are
on gender, culture, ethnicity, or language. impressive and the authors should be com-
People from different cultural backgrounds mended for exploring the usefulness of the
have different expectations and pragmatic data for intervention planning and progress
needs for individuals to fulfill. Further, certain monitoring. Although the assessment of
items appear on face value that they would adaptive behavior should not be limited to
have cultural bias (e.g., &dquo;sings the alphabet norm-referenced testing, the ABAS-II pro-
song,&dquo; and &dquo;recites nursery rhymes from mem- vides data that could strengthen most com-
ory&dquo;). These issues could have practical impli- prehensive assessments of adaptive behavior.
cations for interpreting scores. Even within a The ABAS-II appears technically superior to
most of its competitors and it is recommend-
given culture expectations might be different
between males and females. For example, the ed for use. However, care must be taken to
assure that respondents are able to read and
Comprehensive Test of Adaptive Behavior
understand items when administered in writ-
(CTAB, Adams, 1984) provides two sets of
ten format. Additional research is needed to
norms, one for males and one for females. Test
establish that there is no statistical bias with
development research for the CTAB indicat-
ed that males and females acquired adaptive regard to gender or various subgroup popula-
tions.
skills at different rates and disproportionately
’ ’
in different domains (Adams, 1984). ,t ’

References
The ABAS, II manual does not report to
American Association on Mental Retardation (1992).
what degree score differences may be attribut- Mental retardation: Definition, classification, and sys-
able to gender or cultural differences. The tems of support th
9
(
ed.). Washington, DC: Author.
examiner must rely solely on common sense, American Association on Mental Retardation (2002).
knowledge of other cultures, good family his- Mental retardation: Definition, classification, and sys-
tems of support. Washington, DC: Author.
tories, observations, and clinical judgment to
American Psychological Association (2000). Diagnostic
make decisions about appropriateness of and statistical manual of mental disorders (4th ed., text
applications of ABAS,II scores. revision). Washington, DC: Author.
Adams, G. L. (1984). Comprehensive Test of Adaptive
Conclusions Behavior. San Antonio, TX: The Psychological
The ABAS-II was developed from sound Corporation.
Kincaid, J., & McDaniel, W. (1974). An inexpensive way
theory and empirical methodology, and pro- of calculating Flesch Reading Ease scores (Patent
vides a norm group that is sufficiently large Disclosure Document No. 031350, 215).
and representative of the current U.S. popu- Washington, DC: Patent Office.
lation. The GAC is adequately reliable to U.S. Department of Education (2001). To assure the free
appropriate public education of all children with disabili-
make eligibility and entitlement decisions, ties: Twenty-third annual report to Congress on the
and domain scores are stable enough for clin- implementation of the Individuals with Disabilities Act.
ical and intervention utility. Skill area scores Washington, DC: U.S. Government Printing Office.

Downloaded from aei.sagepub.com at Yale University Library on May 16, 2015