Research in Developmental Disabilities, Vol. 12, pp. 435--451, 1991 0891-4222/91 $3.00 + .

00
Printed in the USA. All fights reserved. Copyright © 1991 Pergamon Press pie

Rating Problem Behaviors in Outpatients

With Mental Retardation: Use of the
Aberrant Behavior Checklist

Lisa S. Freund and Allan L. Reiss

Department of Psychiatry, .Division of Child Psychiatry, Johns Hopkins
University School of Medicine and The Kennedy Institute

Parent and teacher ratings of behavior problems of an outpatient sample of llO

children, adolescents, and young adults with IQs ranging from severe mental
retardation to borderline were obtained using a modified version of the Aberrant
Behavior Checklist (ABC). Using factor analytic techniques, the five-factor
structure of the parent data corresponded extremely well with the five factors
originally obtained from staff ratings o f mentally retarded inpatients (i.e.,
Irritability, Withdrawal, Hyperactivity, Stereotypies, and Inappropriate Speech).
Factor content was virtually identical between the parent and original ABC data
with differences involving only one or two items per scale. The teach-er data
also revealed a factor structure that corresponded to the same five fac-tors as
the parent and original data. Although the teacher and parent factors showed a
high degree o f similarity, the teacher data suggested that the Stereotypies and
Inappropriate Speech factors of the parent and original analy-ses were not the
same constructs for teacher respondents. Age was related to the withdrawal
factor for parent data; level of intellectual functioning was the only subject
characteristic related to factor scale scores in both parent and teacher data.
Test-retest reliabilities were adequate to excellent for all factors for both parent
and teacher data. Parent-teacher cross-informant reliabilities were adequate for
at least four of the factors. The results of the report indicate that the ABC is a
useful, reliable instrument for assessing maladaptive behav-iors in young,
developmentally disabled outpatients.

Thi s work was supported by grants MH0072 6 from the National Institutes o f Health and HD25806
from the National Institute of Child Health and Human Development and grants from the Andrew W.
Mellon Foundation and The John Merck Fund (Dr. Reiss). The authors wish to express their appreciation
to Drs. Michael Aman and Gary Chase for their helpful comments regarding the preparation of this
manuscript.
Requests for reprints should be sent to Lisa S. Freund, Ph.D., Associate Director, Behavioral Genetics
Unit, The Kennedy Institute, Room 507, 550 North Broadway, Baltimore, MD 21205.
435
436 L. S. Freund and A. L. Reiss

There has been a great deal of work in the construction of factor-analyzed

rating scales for assessment of childhood behavior problems. Such instru-
ments, however, rarely have been devised for specific use with mentally
retarded children and adolescents. This absence of useful rating instru-ments
exists despite the evidence that children with mental retardation are more likely
to exhibit deviant behaviors and are at a higher risk for various psychological
and neuropsychiatric disorders than children with average IQs (Corbett, 1985;
Eyman, Borthwick, & Miller, 1981; Menolascino & Swanson, 1982). The
practical need for specific rating instruments for mentally retarded youth
appears widespread. Such instruments would be of value in clinical settings as
research tools, aids in diagnosis, and a means of monitoring interventions.
Mental health professionals interested in the measurement of behavior
disorders in mentally retarded children often utilize instruments that are not
well established psychometrically for the mentally retarded population. Part II
of the AAMD Adaptive Behavior Scale (ABS) (Nihira, Foster, Shellhaas, &
Leland, 1974) assesses domains related to personality and behavior disorders,
including antisocial behavior, withdrawal, stereotyped behavior, and
hyperactive tendencies. The scale has been normed on chil-dren and adults
with mental retardation, but the norming process did not take into account
level of retardation of the sample. Reliability (both cross-informant and test-
retest) is limited, and construct validity has not been established sufficiently
(Sattler, 1988).
A limited number of scales established for children with average IQs have
been reported to be applicable for use with mentally retarded children. For
example, among scales assessing specific psychiatric behavioral disor-ders,
Matson, Barrett, and Helsel (1988) report that the Child Depression Inventory
(Helsel & Matson, 1984) and the Child Behavior Checklist (Achenbach &
Edelbrock, 1978) are usefulin assessing behaviors associat-ed with depression
in mentally retarded children. Two self-rating instru-ments, the Beck
Depression Inventory (Beck, Mendelson, Mock, & Erbaugh, 1961; Beck,
Carlson, Russell, & Brownfield, 1987) and the Depression Self-Rating Scale
(Birleson, 1981), were reported to be useful with developmentally delayed
adolescents. However, these reports were based on small samples in inpatient
units. In general, comprehensive evalu-ation of established scales assessing
behavior disorders for average-IQ chil-dren has not been undertaken for the
mentally retarded population.
Instruments designed specifically for mentally retarded individuals have
been limited in scope or often are restricted to the assessment of adults. The
maladaptive behavior section of the Vineland Scales (Sparrow, Balla,
& Cicchetti, 1984) is limited in that it is a general scale that does not differ-
entiate among domains of behavioral disorders. Other scales are limited in that
they focus on one category of behavioral or psychological disturbance.
 Rating Problem Behaviors 437

Examples are the Emotional Disorders Rating Scale (Feinstein, Kaminer,

Barrett, & Tylenda, 1988), which focuses on mood disturbances in devel-
opmentally disabled children, and the Beckwith et al. (1985) checklist for
assessing depressive symptoms in mentally retarded adults . The
Psychopathology Instrument for Mentally Retarded Adults (PIMRA) (Kazdin,
Matson, & Senatore, 1983) was designed to assess a wide range of symptoms
with scales for depression, anxiety, schizophrenia, psychosex-ual disorder, and
adjustment disorder. Psychometrically, however, the instrument demonstrated
only three broad factors (affective disorder, somotoform, psychosis), with
different clinical categories of psychopathol-ogy represented in each factor
(Matson, Kazdin, & Senatore, 1984; Watson, Aman, & Singh, 1988). In
addition, data were limited in that they were based only on staff or
professional ratings of adults with mental retardation and did not include
parent or teacher ratings of mentally retarded children and adolescents.

One of the more promising checklists for assessing a broad range of

behavior disorders in mentally retarded individuals is the Aberrant Behavior
Checklist (ABC) (Aman & Singh, 1986). The instrument was developed in
New Zealand to measure response to treatment of institution-alized, moderate
to profound mentally retarded individuals across a broad range of psychiatric
symptoms. The psychometrics of the scale are strong (e.g. Aman, Singh,
Stewart, & Field, 1985a; 1985b), reveal several distinct factors, and are based
on data that include mentally retarded children as well as adults. Data based on
staff ratings of a large number of mentally retarded inpatients (from age 5
through adult) were factor analyzed. Five factors were derived, including
Irritability, Lethargy, Hyperactivity, Stereotypic behavior, and Inappropriate
speech. Further research with the instrument has validated the factor structure
based on a U.S. sample of inpatient, mentally retarded adults (Aman,
Richmond, Stewart, Bell, & Kissel, 1987), and a British sample of inpatient,
mentally retarded adults (Newton & Sturmey, 1988). The factor structure of
the scale recently has been validated with a U.S. sample of inpatient children
and adolescents with dual diagnoses of mental retardation and psychiatric
disturbance (Rojahn & Helsel, in press).

Although the ABC is a potentially useful tool for assessing behavior dis-
orders in mentally retarded children and adolescents, more data from these age
groups is necessary in order to establish the instrument's utility. Also of
importance is the establishment of psychometric characteristics and factor
structure when ratings are obtained from parents and teachers of children and
adolescents with mental retardation referred for outpatient treatment. All
current reports on the psychometrics of this instrument are based on samples
of children with high levels of psychiatric disturbance, and the scale may or
may not be equally useful or as reliable for a less disturbed,
438 L. S. Freund and A. L. Reiss

community-based sample. Parent and teacher ratings are important because

these are the individuals who usually report on behavior in clinical outpa-tient
settings. Obtaining ratings on a mentally retarded outpatient sample is
important because the large majority of children and adolescents with
developmental disabilities are followed clinically and through special edu-
cation facilities on an outpatient basis. Thus, this is an important target group
for establishing a valid and reliable checklist for behaviors interfer-ing with
social, academic, and vocational functioning.
In this report, we present data collected from parents and teachers of 110
children, adolescents, and young adults with mental retardation using mod-
ified parent and teacher versions of the ABC (referred to as Parent-ABC and
Teacher-ABC, herein). All children were referred to a neuropsychiatlic clinic
on an outpatient basis. The psychometric characteristics of the instru-ment for
both parent and teacher responders were assessed by employing the same
statistical procedures used by the authors of the original instru-ment. Another
aim of this report was to compare factor characteristics based on parent and
teacher reports of this community-based sample.

METHOD
Subjects
One hundred ten children, adolescents, and young adults with mental
retardation were evaluated. Approximately 72% of these individuals were
patients seen in the outpatient department of a neuropsychiatric unit in a center
for developmentally disabled and handicapped children. These sub-jects were
all those mentally retarded children scheduled for appointments within a 2-
year period whose parents completed the Parent-ABC form. Approximately
10% of the parents with scheduled appointments did not return a Parent-ABC.
The remaining 28% of the sample were participants in an unrelated research
investigation of developmental disabilities utiliz-ing the Parent-ABC and
based at the same center. Ages ranged from 3 to 25 years, with the large
majority of the sample between the ages of 3 and 18 years. All levels of mental
retardation were represented, as was borderline intellectual functioning. More
complete characteristics of the group, including age, gender, and functional
level, are presented in Table 1.

Instrument

The original ABC consists of 58 items. For each item, the respondent
decides to what degree the behavior is a problem on a scale of 0 (not at all a
problem) to 3 (the problem is severe in degree). The modified versions
(Parent-ABC and Teacher-ABC) used in this study changed the wording of
 Rating Problem Behaviors 439

TABLE 1.
Characteristics of Subjects Rated by Parents Using the Aberrant
Behavior Checklist

Male Female Total

(n = 76) (n = 34) (n = 110)

Age
M (in years) 10.2 11.6 10.5
SD 4.6 4.1 5.0
IQ
M 53.5 51.9 53.0
SD 14.4 16.2 14.9
Severity of retardation (% o f sample)a
Borderline 10 4
Mild 24 13
Moderate 20 5
Severe 15 9
Age groups (% of sample)
3-5 years 12 5
6-12 years 32 15
13-17 years 20 5
18-25 years 5 6

aCategorization based on DSM-III-R criteria.

the original ABC items and the instructions for completing the form for the
purposes of clarity and reduced reading level. While the original ABC list-ed
elaborated item descriptors in a manual (Aman & Singh, 1986) separate from
the rating form, we added reworded item descriptors from the ABC manual to
the rating form itself under each item. Our modifications result-ed in a Parent-
ABC with an overall 6th grade reading level and a Teacher-ABC with an
overall 11 th grade reading level. Example item modifications for the parent
and teacher forms were as follows:
Original Form: "Stereotyped, repetitive movements." "Movements
Parent-ABC Form: made over and over. For example, head rolling,
hand waving, complex finger move-ments, tapping
hands on body and bouncing. "Seeks isolation
Original Form: from others."
Teacher-ABC Form: "Seeks isolation from others. Uncommunicative,
moody, hides or finds a quiet sFot by oneself. Sits
in comer or moves away when approached by
others."

Instructions to both parents and teachers requested ratings of the child's

current behaviors. Approximately 20% of the sample was on medication for
behavior. If the child was on medication for behavior some times and
440 L. S. Freund and A. L. Reiss

off medication other times, the parent or teacher was to rate the behavior as it
would be if the child were off medication. This instruction was given as a means
of understanding the severity of the child's behavior problems when not controlled
through medication. The respondents were asked to rate how the child behaves
with others, not just with the parents or teacher. In addi-tion, the respondent was to
consider whether a behavior interfered with the child's development.

Procedure

Upon the scheduling of clinical or research appointments, parents were mailed

Parent- and Teacher-ABCs. The two versions were clearly marked as "Parent" or
"Teacher." Parents were asked to complete the parent form and give the teacher
version to their child's teacher. The parents were asked to bring the completed
forms with them to their child's appointment.

RESULTS
Parent - ABC - Psychometric Characteristics

Factor analysis. A five-factor structure very similar to that found for the original
ABC was generated for the Parent-ABC. Analyses were carried out using
programs from StatView II for the Macintosh Computer. Factor analyses were
completed in a manner similar to that used by the originators of the instrument
(Aman & Singh, 1986). A principal factoring method with iteration was used and
followed by varimax rotation followed with promax rotation. The original ABC
analysis generated a five-factor solu-tion. A scree test on the present Parent-ABC
data confirmed that a five-fac-tor solution would include the maximum number of
meaningful factors for the data. Initial analyses of the Parent-ABC identified four
items from the original scale with low squared multiple correlations (.40 and
below) and low loadings on all five factors (below .30). The items were considered
unreliable and were subsequently dropped from the final analysis. The items
included "inactive, sluggish," "blank expression on face," "rolls head back and
forth," and "does not try to communicate by words or gestures." Thus the final
five-factor solution for the Parent-ABC included 54 items and was found to
account for 55% of the common variance. The greatest interfactor correlations
were .38 (Factors 1 and 3) and .34 (Factors 5 and 3). All other interfactor
correlations were below (+) . 14.

All items in the final solution of the Parent-ABC had loadings at .41 or higher.
Table 2 shows sample items and the mean factor loading for each factor for the
parent data. (Loadings for teacher data also are shown in Table 2 and are discussed
later on.)
 TABLE 2.
Five-Factor Structure Loading Matrix a Generated by a Principal Factoring Method With Iteration and Varimax Rotation Followed by Promax for
Parent b and Teacher c Respondent Ratings

Factor Loading
Subscale I II III IV V
Item Number Parent Teacher Parent Teacher Parent Teacher Parent Teacher Parent Teacher

I. Irritability

10. Temper tantrums .80 .72 - . 06 .06 .01 .05 -.01 .13 .18 .08
14. Irritable ("grumpy " o r "whiny") .60 .72 .06 .04 -.07 -.12 .03 -.09 .36 .19
29. Demands must be met right away .70 .78 -.08 -.08 .11 .09 -.01 -.04 .25 - . 16
52. Injures self .67 (d) .07 (d) -.21 (d) .22 (d) - . 29 (d)
57. Throws temper tantrums when she/he
4~ does not get own way .76 .83 - . 11 - . 02 .01 .05 .07 .10 .22 .00
Mean loading (parent n = 15;
teacher n = 16) .61 .67 .08 .04 .18 .15 .10 .11 .16 .10

II. Withdrawal

5. Tries to be alone, does not interact

with others - . 02 .12 .82 .75 .01 -.17 .05 .03 -.15 - . 04
16. Withdrawn, prefers to do things alone;
quiet to the exla'eme .02 - . 10 .75 .80 -.I1 - . 16 .06 .00 - . 20 - . 10
23. Does nothing but sit and watch others;
does not respond to social play - . 12 -.18 .79 .74 -.04 -.08 -.03 -.13 - . 02 .22
42. Prefers to be alone .02 .13 .87 .74 -.09 -.17 .14 .04 -.11 - . 17
Mean loading (parent n = 14;
teacher n = 18) .08 .12 .64 .65 .09 .15 .10 .06 .12 .09
(Continued on next page)
 TABLE 2. Continued

Factor Loading
Subscale I II III IV V
Item Number Parent Teacher Parent Teacher Parent Teacher Parent Teacher Parent Teacher

III. Hyperactivity

1. Overactive, does not sit still, pushes

chairs, runs or walks all over the place .04 -.04 -.01 -.16 .84 .88 .09 .00 -.06 .04
15. Restless, unable to sit still -.05 -.01 -.03 -.08 .86 .89 .21 .05 .04 .11
38. Will not stay in seat .04 -.05 -.01 .03 .79 .82 .15 .04 .04 -.04
48. Constantly runs or jumps
around the room .11 .02 .02 .00 .82 .74 .03 -.03 -.13 -.10
4~
to 54. Excessively active .01 -.01 -.18 -.14 .87 .80 .15 .09 -.14 .05
Mean loading (parent n = 15;
teacher n = 13) .14 .02 .11 .15 .66 .69 .16 .05 .11 .10
IV. Stereotypies
6. Meaningless body movements made
over and over .05 -.10 .06 .17 .00 .36 .70 .60 .01 .20
7. Injures self (d) .54 (d) -.37 (d) .49 (d) .62 (d) .14
11. Repetitive movements -.02 -.10 .08 .17 .05 .36 .72 .78 .29 .20
45. Waves or shakes arms, fingers, legs or
toes over and over in same way
for no reason -.08 -.09 -.13 -.03 .06 .18 .53 .68 .37 -.11
Mean loading (parent n = 5;
teacher n = 8) .08 .20 .11 .09 .06 .22 .63 .61 .15 .12
 V. Inappropriate Speech

9. Talks excessively; talks much

of the time; may or may not
make sense .15 .22 -.02 -.25 -.10 .31 .09 -.08 .60 .49
10. Repetitive words .00 (d) -.02 (d) .17 (~ .18 (ct) .62 (d)
33. Talks to self -.14 -.02 .21 .17 --.06 .32 .18 --.02 .53 .59
46. Says words or phrases
over and over, sounds
like a broken record -.03 (d) -.01 (d) .19 (d) .16 (d) .57 (~
Mean loading (parents n = 5;
teacher n = 2) .14 .12 .07 .21 .11 .32 .15 .05 .57 .54
aTable contains highest loading items for each subscale.
bFour items with low loadings were dropped from the original 58 ABC items.
4~ cOne item with a low loading was dropped from the original 58 ABC items.
dltem was not included in this factor for this data.
444 L. S. Freund and A. L. Reiss

The same general five factors derived for the original ABC were found for
the Parent-ABC. Based on item content, it appeared quite reasonable to retain
the labels from the original analyses of the ABC for the Parent-ABC. These
labels were Irritability, Withdrawal, Hyperactivity, Stereotypies and
Inappropriate Speech. In order to compare the Parent- and Original-ABC
(Aman et al., 1985a) factor structures, coefficients of congruence (Catell,
1978) were computed for the five factors. Highly similar factor structure was
indicated by coefficients of .88, .84, .86, .87, and .82 for factors 1 through 5,
respectively.
Only slight differences in item content were noted between the Parent and
original ABC factors. Forty - nine of the 54 items (91%) from the Parent-ABC
loaded on the same factors as the original factor derivation (Aman et al.,
1985a). Four items had been dropped from the Parent-ABC because of their
low loadings. When the loadings for these items were compared across
factors, all four items loaded most highly on the same Parent-ABC factors as
they did on the original ABC. The Withdrawal and Inappropriate Speech
Factors were replicated exactly. The remaining fac-tors differed by only one or
two items. These differences included the fol-lowing: (1) The original ABC
factor, Irritability, included "cries over little things" and "depressed mood . "
These two items were included in the Inappropriate Speech and Withdrawal
factors, respectively, of the Parent-ABC. (2) Two of the items in the original
Hyperactivity factor of the origi-nal ABC analysis were found in the
Irritability factor for the Parent-ABC (i.e.,"inappropriately noisy and rough"
and "disrupts group activities"). (3) The final discrepancy involved the
"strange behavior" item, which loaded highest on the Hyperactivity factor for
the Parent-ABC instead of the Stereotypies factor in the original analysis.

Composite scores were generated for the entire sample by the procedure
suggested by the authors of the ABC (Aman et al., 1985a). Composite scores
were derived by summing the obtained ratings (0 to 3) for the items loading
most highly for each factor. This procedure resulted in a total score for each of
the five subscales for each subject. Correlations among the five subscales of
the Parent-ABC were conducted via Pearson correlations. The two
externalizing behavior subscales, Irritability and Hyperactivity, showed a
moderate to strong correlation, r = .72 in the present analysis, whereas these
two subscales were only moderately correlated (r = .52) in the origi-nal
analysis. The remaining subscales showed little interdependence, with
coefficients ranging from . 19 to .34.
Although our sample was too small to generate norms for the composite
scores, multiple regression analyses were performed with the scores to
evaluate whether subject characteristics such as gender, age, and intellec-tual
functioning were related to composite subscale scores. Age, IQ, and the
produc t of age and IQ (carrying the age × IQ interaction) were
 Rating Problem Behaviors 445

regressed on subscale scores. The multiple regression coefficient was sig-n i f i c a

n t f o r the Irritabl e (R = .286; F ( 3 , 106) = 3 . 14, p < .05) and Withdrawal (R =
.31; F[3, 106] = 5.14, p < .05) subscales. Analysis of the individual contributions
of the age, IQ, and interaction variables showed t h a t as IQ d e c r e a s e d , th e I
r r i t a b i l i t y c o m p o s i t e s c o r e s i n c r e a s e d , F(1, 106) = 8.02, p <
.01, and higher withdrawal composite scores were associated with increased age,
F(1, 106) = 5.63, p < .01. No other effects associated with age, IQ, or their
interaction were significant. Multiple regression analyses of age, gender, and their
interaction with subscale scores revealed no further significant effects.

Reliability. Alpha coefficients for assessing the internal consistency among items
within each of the subscales were computed. Similar to analyses of the original
ABC, results indicated that the extent to which items within each subscale
measured a consistent attribute was quite high. Alpha coeffi-cients were .90 for
Irritability, .93 for Withdrawal, .90 for Hyperactivity, .88 for Stereotypic behavior,
and .83 for Inappropriate Speech.
Test-retest reliability indicated a good stability in the composite scores ove r
time. The test-retest reliability was assessed by having a rando m selection of 30
parents complete the Parent-ABC a second time, approxi-mately 1 month after
completion of the first Parent-ABC. Pearson corre-l a t i o n c o e f f i c i e n t s f o
r th e s u b s c a l e s w e r e .95 ( I r r i t a b i l i t y ) , .92 (Withdrawal), .88
(Hyperactivity), .88 (Stereotypic behavior), and .80 for (Inappropriate Speech).

Teacher-ABC Psychometric Characteristics

Ninety-four of the 110 respondents to the Parent-ABC had their child's teacher
complete a Teacher - ABC . This subgroup of children was com - prised of 65
males and 29 females. The mean age of the children was 11 y e a r s (SD = 5 . 0; r
a n g e : 3 - 2 6 y e a r s ) and the m e a n IQ wa s 5 2 . 2 9 (SD = 15.56; range: 20-
79).
Factor analyses. Factor analyses identical to those performe d with the Parent-
ABC were conducted with the Teacher-ABC data. A scree test indi-cated that a
five-factor solution included the largest number of meaningful factors. Only the
"rolls head back and forth" item was deemed unreliable and dropped for this
analysis. A set of five factors similar to those for the Parent- and original ABCs
was generated. The final five-factor solution for the Teacher-ABC accounted for
60% of the variance. Interfactor correla-tions were low, with the largest
correlations at .39 (factors 2 and 4) and
.32 (factors 1 and 3). All other correlations were below (+) .12. Table 2 shows
sample items and the mean factor loading for each factor for the parent data.
446 L. S. Freund and A. L. Reiss

Coefficients of congruence were computed for the five factors in order to

compare the Parent- and Teacher-ABC factor structures. The coefficients for
factors 1 through 5 were .91, .86, .81, ,70, and .65, respectively. This result
indicates that the factor structures were fairly similar between the Parent- and
Teacher-ABCs. Disregarding the four items dropped from the Parent-ABC
analysis, 44 (80%) of the remaining 54 items of the Teacher-ABC were
assigned to the same factors as in the Parent-ABC analysis and the original
ABC analysis. Major differences between the Teacher-ABC and both the
Parent- and original ABC analyses were with the Stereotypies and
Inappropriate Speech Factors. The Teacher-ABC analysis grouped self-injury
with stereotypies instead of the Irritability items as in the other two analyses.
This led to a renaming of the factor to Stereotypies/Self-injury for teacher data.
The teacher data also grouped repetitive speech items with the Hyperactivity
Factor, whereas the other two factor solutions grouped these items within the
Inappropriate Speech Factor.
Composite subscale scores were generated in the manner described for the
Parent-ABC analysis. Correlations among the five subscales of the Teacher-
ABC were computed using Pearson correlation coefficients. For these data,
the Hyperactivity composite scores were moderately correlated with both the
Irritability and Inappropriate Speech composite scores (r = .55 and .51,
respectively). Although the other intercorrelations were low, ranging between
.12 and .38, the Teacher-ABC composite scores showed more
interdependency than those generated for the Parent-ABC.
Multiple regression analyses assessing the relation between demographic
characteristics and the teacher subscale scores were conducted in the same
manner as for the parent data. The only significant relation observed was for
the multiple regression analysis for Stereotypies/Self-injury subscale (R =
.295; F[3, 106] = 3.26, p < .05). Analysis of individual variables showed that
as IQ decreased, the Stereotypies/Self-injury subscale scores increased
(F[1,106] = 5.0, p < .01).

Reliability. Internal consistency was high for each factor. Alpha coefficients
assessing the internal consistency among items within each of the subscales were
.88 for Irritability, .94 for Withdrawal, .89 for Hyperactivity, .90 for Stereotypic
behavior, and .79 for Inappropriate Speech.
Test-retest reliability was also assessed. Teachers of 25 children com-pleted
a second Teacher-ABC approximately 1 to 2 months after comple-tion of the
first Teacher-ABC. The test-retest reliability was not as high as that for the
Parent-ABC for each factor. Spearman correlation coefficients for the
subscales were .61 (p < .05) Irritability, .50 (p < .05) Withdrawal,
.61 (p < .05) Hyperactivity, .67 (p < .01) Stereotypic behavior, and .59 (p <
.05) Inappropriate Speech).
 Rating Problem Behaviors 447

In order to assess cross-informant reliability between parents and teach-ers

accurately, Teacher-ABC subscale scores were recomputed by imposing the
Parent-ABC subscale assignment on the teacher data. Reliability was then
computed by means of Pearson correlation coefficients between the Parent- and
Teacher-ABC subscale scores. The results indicated moderate reliability for four
of the scales: .49 (p = .0001) Irritability, .47 (p = .0001) H y p e r a c t i v i t y , .45
( p = . 0 0 0 1 ) S t e r e o t y p i e s , and .39 ( p = . 0 0 0 3 ) Inappropriate Speech.
The mean correlation across subscales was .45. The Withdrawal subscale was
deemed unreliable in terms of multiple raters since the correlation between parent
and teacher data on this factor was not significant (r = . 18).

DISCUSSION
The need for comprehensive and reliable instruments for the measure-ment o f
behavior disorders in young individuals with mental retardation motivated this
investigation. The results indicate that the ABC is such an instrument and that it
can be used for an outpatient, mentally retarded pop-ulation with parent and
teacher respondents. This instrument could be use-ful to clinicians as a quick
screening tool for identifying problem behavior areas and to both clinicians and
researchers as a means of monitoring the effects of therapeutic interventions in the
school and home environments.
Parent and teacher versions of the ABC were assessed for their psycho-metric
characteristics using the same factor analytic techniques as those performed on the
original ABC (Aman & Singh, 1986) data acquired from inpatient caretakers. In
the present analysis, parent and teacher data were obtained for an outpatient
sample of children, adolescents, and young adults ranging in intellectual level
from severe mental retardation to bor-derline intellectual functioning. The five-
factor structure using parent data corresponded extremely well with the original
ABC five factors. Factor content was virtually identical between the Parent-ABC
and the original analysis. Differences between the two analyses involved only one
or two items per scale. The parent analysis also revealed four items that could not
be included reliably in the final factor solution (i.e., "sluggish," "blank face," "n o
communication," "rolls head"). These items showed very low ratings across
parents, suggesting that these behaviors either do not occur frequentl y with
parents or are not p e r c e i v e d as significant problems . Among subject
characteristics such as gender, age, and IQ, level of mental retardation was
negatively related to Irritability and age was positively related to Withdrawal.
These results are similar to those reported by Aman, et al. (1987). That is, no
differences among composite scores based on gen-der were found, age differences
were apparent only when data from older
448 L. S. Freund and A. L. Reiss

individuals were included, and more severe mental retardation was related to
higher ratings on the Irritability and Stereotypies factors.
The data from the Teacher-ABC revealed a factor structure that corre-
sponded to the same five factors as the parent and original analyses. Although
the teacher and parent factors showed a high degree of similarity, the teacher
data suggest that the Stereotypies and Inappropriate Speech fac-tors of the
parent and original analyses are not the same constructs for teacher
respondents. The repetitive speech and self-injury items from the teacher data
consistently loaded on the Hyperactivity and Stereotypies fac-tors,
respectively, instead of on the Inappropriate Speech and Irritability Factors as
the parent data did. Only the Hyperactivity, Irritability and Withdrawn scales
appear to maintain a consistent structure between parent and teacher
respondents. Degree of mental retardation was again the only subject
characteristic related to the subscale scores such that teachers rated individuals
with lower IQs higher on the Stereotypies/Self-injury factor.
Reliability in terms of internal consistency was high for both parent and
teacher data sets. Whereas the parent test-retest reliability was excellent and
comparable to that reported for the original ABC (Aman, Singh & Turbott,
1987), teacher test-retest reliability was only moderate. It is not clear why
teacher reliability was not as strong as parent reliability, espe-cially since all
teachers were individuals who interacted with the subjects everyday, five days
a week, during the school year. Even so, teachers may have had less
opportunity to observe the subjects across a range of situa-tions and may have
more expertise in extinguishing problem behaviors than parents.

The cross-informant reliability between parent and teacher respondents was

moderate for the Irritability, Hyperactivity, Stereotypies, and Inappropriate
Speech scales. These results correspond fairly well with those reported for
multiple inpatient staff raters of the original ABC (Aman et al., 1987). The
present parent-teacher reliabilities are all somewhat lower than those in the
original report. However, the parent-teacher relia-bilities reported here are
comparable to or stronger than those reported for other behavior rating scales
assessing childhood psychopathology in aver-age IQ children (e.g.,
Achenbach, McConaughy, & Howell, 1987; The Behavior Problem Checklist;
Quay, 1979; and the Child Behavior Checklist; Jensen, Traylor, Zenakis, &
Davis, 1988). For example, Achenbach et al. (1987) report a mean correlation
of .27 for behavior rat-ings from teachers and parents of the same child across
a variety of scales and studies. Thus the mean correlation of .45 for parent-
teacher ratings using the ABC is impressive by comparison. As discussed by
Achenbach et al. (1987), the correlations between teacher-parent respondents
is not higher most likely because of the situational differences under which
par-ents and teachers observe the child.
 Rating Problem Behaviors 449

The Withdrawal scale was not found reliable between parent and teacher
respondents at all. Again, this is not unexpected given that Achenbach et al.
(1987) found significantly lower correlations between parent and teacher
respondents across rating scales for internalizing (overcontrolled) behaviors
than externalizing (undercontrolled) behaviors. Indeed, the internalizing
behaviors contained in the Withdrawal scale are more difficult to assess and
may not cause readily identifiable behavior management problems in the
classroom. Parents also may be more sensitive to subtle withdrawal behav-iors
than teachers. On the basis of this analysis, therefore, the Withdrawal factor
will be reported differently depending upon the respondent.
The ABC is a quick, straight-forward rating scale that is easy to adminis-ter
once expanded descriptors are added to the items on the scale itself. Parent
data obtained from the ABC offers information on a clinically important range
of problem behaviors. The present analysis found slight deviations in the
parent data factor structure from the original ABC factors based on inpatient
staff ratings. The results suggest fundamental agreement in factor structures
between institutional and community samples, although additional
community-derived data are needed to determine if the differ-ences noted here
are robust. If such differences are consistent, then a slight-ly different scoring
system may be appropriate for the outpatient, mentally retarded population.
Teacher respondents to the ABC are somewhat less reliable than parents and
do not perceive problem behaviors within the same conceptual structures as
parents. On the basis of the analyses reported here, the Teacher-ABC data are
perhaps best utilized in conjunction with parent data in order to assess child
behaviors across more than one situa-tional domain.

Future research involving larger samples at all age levels between preschool
and young adulthood is necessary. Analysis of parent and teacher data from
larger samples of children using confirmatory factor analysis (e.g., with
LISREL) would provide the most conclusive evidence for the factor structure
described in this and prior studies with the ABC. Larger samples would be
useful for establishing norms for the ABC based on par-ent and teacher
ratings. The availability of norms will make the instrument much more useful
in clinical settings. While adequate criterion validity has been shown for the
ABC subscales when compared with independent psy-chiatric diagnoses for a
sample of inpatient children with mental retardation (Rojahn & Helsel, in
press), it is also important to establish clinical validity using a sample of
children, adolescents, and young adults with mental retardation being followed
on an outpatient basis. The clinical validity of the ABC in relation to
psychiatric diagnoses for the sample assessed in this study will be presented in
a subsequent report. At present, this instrument appears well suited for
research purposes and for use as an aid in clinical evaluation of problem
behaviors of young, mentally retarded outpatients.
45 0 L. S. Freund and A . L. Reiss

REFERENCES
Achenbach, T. M., & Edelbrock, C. S. (1978). The classification of child psychopathology: A
review and analysis of empirical efforts. Psychological Bulletin, 85, 1275-1301.
Achenbach, T. M., McConaughy, S. H., & Howell, C. T. (1987). Child/adolescent behavioral and
emotional problems: Implications of cross-informant correlations for situational specificity.
Psychological Bulletin, 101,213-232.
Aman, M. G., Richmond, G., Stewart, A. W., Bell, J. C., & Kissel, R. C. (1987). The Aberrant
Behavior Checklist: Factor structure and the effect of subject variables in American and New
Zealand Facilities. American Journal of Mental Deficiency, 91,570-578.
Aman, M. G., & Singh, N. (1986). Aberrant Behavior Checklist. New York: Slosson Educational
Publications, Inc.
Aman, M. G., Singh, N., Stewart, A. W., & Field, C. (1985a). The Aberrant Behavior Checklist:
A behavior rating scale for the assessment of treatment effects. American Journal of Mental
Deficiency, 89, 485-491.
Aman, M. G., Singh, N., Stewart, A. W., & Field, C. (1985b). Psychometric characteristics of the
Aberrant Behavior Checklist American Journal of Mental Deficiency, 89,492-502.
Aman, M. G., Singh, N., & Turbott, S. (1987). Reliability of the Aberrant Behavior Checklist and
the effect of variations in instructions. American Journal of Mental Deficiency, 92, 237-240.
Beck, A. T., Mendelson, M., Mock, J., & Erbaugh, J. (1961). An inventory for measuring depres-
sion. The Archives of General Psychiatry, 4, 53-63.
Beck, D. C., Carlson, G. A., Russell, A. T., & Brownfield, E E. (1987). Use of depression rating
instruments in developmentally and educationally delayed adolescents. American Academy of
Child and Adolescent Psychiatry, 26, 97-100.
Beckwith, B. E., Parker, L., Pawlarczyk, D., Cook, D. I., Schumacker, K. S., & Yearwood, K.
(1985). The dexamethasone suppression test in depressed retarded adults: Preliminary findings.
Biological Psychiatry, 20,825-831.
Birleson, P. (1981). The validity of depressive disorders and the development of a self-rating
scale: A research report. Journal of Child Psychology and Psychiatry, 22, 73-88.
Catell, R. B. (1978). The scientific use of factor analysis in behavioral and life sciences. New
York: Plenum Press.
Corbett, J. A. (1985). Mental retardation Psychiatric aspects. In M. Rutter & L. Hersov (Eds.),
Child and adolescent psychiatry: Modern approaches (pp. 661-678). Oxford: Blackwell
Scientific Publications.
Eyman, R. K., Borthwick, S. A., & Miller, C. (1981). Trends in maladaptive behavior of mentally
retarded persons placed in community and institutional settings. American Journal of Mental
Deficiency, 85, 473~[77.
Feinstein, C., Kaminer, Y., Barrett, R. P., & Tylenda, B. (1988). The assessment of mood and
affect in developmentally disabled children and adolescents: The Emotional Disorders Rating
Scale. Research in Developmental Disabilities, 9, 109-121.
Helsel, W. J., & Matson J. L. (1984). The assessment of depression in children: The internal struc-
ture of the child depression inventory (CDI) and its relationship to the Matson Evaluation of
Social Skills with Youngsters 0VIESSY). Behavioral Research Therapy, 22,289-294.
Jensen, P. S., Traylor, J., Zenakis, S. N., & Davis, H. (1988). Child psychopathology rating scales
and interrater agreement: I. Parents' gender and psychiatric symptoms. Journal of the American
Academy of Child and Adolescent Psychiatry, 17, 442-450.
Kazdin, A. E., Matson, J. L., & Senatore, V. (1983). Assessment of depression in mentally retard-
ed adults. American Journal of Psychiatry, 140, 1040-1043.
Matson, J. L., Barrett, R. P., & Helsel, W. J. (1988). Depression in mentally retarded children.
Research in Developmental Disabilities, 9, 39-46.
Matson, J. L., Kazdin, A. E., & Senatore, V. (1984). Psychometric properties of the
Psychopathology Instrument for Mentally Retarded Adults. Applied Research in Mental
Retardation, 5, 81-89.
 Rating Problem Behaviors 451

Menolascino, F., & Swanson, D. (1982). Emotional disorders in the mentally retarded. British
Journal of Mental Subnormality, 28, 46-55.
Newton, J. T., & Sturmey, P. (1988). The Aberrant Behavior Checklist: A British replication and
extension of its psychometric properties. Journal of Mental Deficiency Research, 32, 87-92.
Nihira, K., Foster, R., SheUhass, M., & Leland, H. (1974). AAMD Adaptive Behavior Scale (rev.
ed.). Washington, DC: American Association of Mental Deficiency.
Quay, H. C. (1979). Classification. In H. C. Quay & J. S. Werry (Eds.). Psychopathological disor-
ders of childhood (2nd ed.). New York: Wiley.
Rojahn, J., & Helsel, W. J. (in press). The Aberrant Behavior Checklist in children and adoles-
cents with dual diagnosis. Journal of Autism and Developmental Disorders.
Sattler, J. M. (1988). Assessment of children. San Diego: Jermoe M. Sattler, Publisher.
Sparrow, S. S., Balla, D. A., & Cicchetti, D. V. (1984). Vineland Adaptive Behavior Scales. Circle
Pines, MN: American Guidance Service.
Watson, J. E., Aman, M. G., & Singh, N. (1988). The Psychopathology Instrument for Mentally
Retarded Adults: Psychometric characteristics, factor structure, and relationship to subject
characteristics. Research in Developmental Disabilities, 9, 277-290.