INTER-RATER SCORE AGREEMENT BETWEEN UNTRAINED AND TRAINED
 
PAIRS OF CHII~D CARE WORKERS WHO RATED A MENTALI~Y RETARDED
 
SAMPLE WITH THE AAMD ADAPTIVE BEHAVIOR SCALE
 
A Thesis
 
Presented to
 
the Department of Psychology
 
Emporia State University
 
In Partial Fulfillment
 
of the Reqllirements for the Degree
 
Master of Science
 
by
 
Westley E. Tatman
 
December 1977
 
AN ABSTRACT OF THE THESIS OF 
Westley E. Tatman for the Master of Science
 
(name of student) (degree)
 
in Psychology
 ---presented on December 23, 1977 (maJor) ( date) 
Title: INTER-RATER SCORE AGREEMENT BETWEEN UNTRAINED AND 
TRAINED PAIRS OF CHILD CARE WORKERS WHO RATED A MENTALLY 
RETARDED SAMPLE WITH THE AAMD ADAPTIVE BEHAVIOR SCALE 
Abstract approved: ~ ~,JJ4~U4 
A study was undertaken to test the reported inter-rater 
reliabilities on Part Two of the American Association 
on Mental Deficiency Adaptive Behavior Scale. In 
addition, this study trained raters in order to determine 
if trained raters would obtain higher reliabilities than 
untrained raters, since the AAMD claimed that untrained 
persons could accurately administer the scale. From a 
state institution, a sample of 16 child care workers from 
,;, 
three different residential units rated retardates on 
their respective units. The total number of retardates 
from each residential unit totaled 22 individuals. The 
\ 
;},?~~~~J: ;:~j~; 
results indicated that the trained raters achieved signif­
 icant relationships less often than did the untrained 
raters' when their domain scores were compared. In the 
first unit, the scores of all raters were found to have a 
significant relationship in five of the 14 domains, while 
only three of the raters' domain scores in the second 
unit were found to have significant relationships, and 
none of the domain scores of the raters in the third unit 
were found to have a significant relationship on a 
consistent basis that considered untrained/trained, 
untrained and trained rater combinations. Although the 
correlation coefficients of each rater type varied from 
one unit to another, only two of the domain correlations 
consistently supported the findings of the reliability 
study cited by the AAMD on Part Two of the revision of 
this scale. 

ACKNOWLEDGEMENTS 
I would like to thank Dr. Louise Hoover, Dr. 
Jeannie Frieman and Dr. Ray Foster for their assistance 
during the course of this study. 
Appreciation is also extended to the Child 
Developmentaljsts who volunteered to participate in this 
study. Special mention and appreciation is given to those 
developmentalists who participated as trained raters in 
the study. 
~ 
ii 
TABLE OF CONTENTS 
Page 
LIST OF TABLES ••••••••••••••••••••••••••••••••••• v 
Chapter 
1. INTRODUC TION ••••••••••••••••••••••••••••• 1 
THEORETIC AL FORMlJLATION ••••••••••••• 0 •• 0. 1 
TIlE PROBIJEM 0 ••••••••••••••••••••••••• 0 • 3 
Statement of the Problem •••••••••••• 0 4 
Statement of the Hypothesis •••••••••• 4 
Purpose of the Study •••••••.••••••••• 4 
Si.gnificance of the Study ••••••••••.• 5" 
DEFINITIONS OF TERMS ••••••••••••••••••••• 5" 
Adaptive Behavior •••••••••••••••••••• 6 
Reliability ••••••••••.••••••••••••• 0. 6 
o 0 • • • • • • • • •Intelligence Quotient (IQ) 6 
Mental Retardation 00 ••••••••••••••••• 6 
Mild Mental Retardation •••••••••••• 0. 6 
Moderate Mental Retardation •••••••••• 7 
Severe Mental Retardation •••••••••••• 7 
Profound Mental Retardation •••••••• 0. 7 
LIMITATIONS OF THE STOny •• 0 ••••••••••••• 0 7 
2. REVIEW OF RELATED LB'ERATURE .0 ••••••••••• 9 
3. METHODS AND PROCEDURES •••••••••••••••••• 0 20 
• • • • • • • • • • • • • • • • POPULATION AND SAMPLING 21 
iii 
• • • • • • • • • • • • • • • • • • • • • • • • • • 
• • • • • • • • • • • • • • • • • • • 
MATERIALS AND INSTRUMENTATION •••••••••• 
DESIGN ••••••••••••••••.••.••••••••••••• 
DATA COLLECTION •••••••••••••••••••••••• 
DATA ANALYSIS 
4.	 ANALYSIS OF DATA • • • • • • • • • • • • • • • • • • • • • • • • • • 
RESPONSE ANALYSIS •••••••••••••••••••••• 
STATISTICAL ANALYSIS 
5.	 SUMMARY, CONCLUSIONS AND RECOMMENDATIONS • • 
SUl4M"AR.Y •••••••••••••••••••••••••••••••• 
CONCLUSIONS • • • • • • • • • • • • • • • • • • • • • • • • • • • • 
RECOMMENDATIONS •••••••••••••••••••••••• 
REFERENCE NOTES ••••••••••••••••••••••••••• 
REFERENCES • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 
APPENDIXES • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 
A.	 Instructions for completing the 
AAl1D Adaptive Behavior Scale •••••• 
B.	 Data Summary Sheets • • • • • • • • • • • • • • • 
C.	 Profile Summary Sheets •••••••••••• 
D.	 Additional Instruction Sheet •••••• 
iv 
Page 
24
 
26
 
28
 
29
 
31
 
31
 
32
 
45
 
45
 
48
 
52
 
55
 
56
 
59
 
59
 
63
 
66
 
69
 
LIST OF TABLES 
Table Page 
1. Unit A Rater Score Comparisons • • • • • • • • • • • • • 34­
 2. Unit B Rater Score Comparisons • • • • • • • • • • • • • 37 
3. Unit C Rater Score Comparisons • • • • • • • • • • • • • 4-1 
v 
Chapter 1 
INTRODUC TION 
The American Association on Mental Deficiency 
(AAMD) reported that its newly revised Adaptive Behavior 
Scale showed good reliability. It reported high reliabil­
 ities between the rating scores of child care workers who 
independently rated the same subject (AAMD, 1974). This 
chapter has been devoted to presenting a similarly designed 
reliability study that compared the rating scores between 
child care workers who rated the same sUbject with the 
AAMD Adaptive Behavior Scale. 
THEORETICAL FORMULATION 
Within the past decade professionals have become 
critical of the IQ score as the sole criterion used in the 
assessment of the individual's intellectual processes. 
Professionals have been aware, and have long recognized 
that a person's social adaptation cannot be totally pre­
 dicted from the intellectual process. Two individuals with 
identical IQ scores do not necessarily cope with societal 
expectations the same way. Intelligence scores vary from 
one person to another, and so does the way in which a 
person will adapt and adjust to his social world (Nihira, 
1969). 
1 
2 
Professionals within the field of mental retarda­
 tion have noted and reported that the use of IQ alone, 
without assessing the individual's social functioning, 
was quite damaging. Frequently, the mentally retarded 
individual is not given recognition for acquired social 
skills; consequently, this individual is inappropriately 
placed with other retardates whose social skills are not 
as well developed (Nihira, 1969). 
In 1961 the AAMD specified that a diagnosis of 
mental retardation must include deficiencies within two 
dimensions, namely measured intelligence and adaptive be­
 havior (Nihira, 1969). In order to assist clinicians in 
the assessment of adaptive behavior, the AAMD developed 
the Adaptive Behavior Scale in 1969, which was revised in 
197~. Tests that assert accurate behavior measurement 
usually refer to completed studies which demonstrate that 
the test actually measures what it claims, i.e., has good 
validity. In addition, references are also made to com­
 pleted studies which demonstrate that the scores obtained 
by different administrators are relatively similar, i.e., 
have good reliability. The 197~ revision of the AAMD 
Adaptive Behavior Scale claimed high reliability between 
test administrators (AAMD, 197~). This claim was based on 
a single study, which has not been published. Considering 
that only one study has been completed on the new revision, 
and that no study has been published, further research into 
this scale's reliability is warranted. 
3 
THE PHOBIJJM 
The Adaptive Behavior Scale manna1 implied that the 
scale can be used by individuals with little or no train­
 ing, such as institutional aides, parents and teachers 
(AAMD,1974). If, in fact, untrained persons used this 
scale and obtained results of high reliability, the claims 
of its advocates concerning inter-rater reliabili. ty woul.d 
be supported. However, if the scores were found to be dis­
 similar, then the claims made concerning scorer reliability 
could be questioned. Since the reported AAMD reliability 
study used only one pair of raters for each individual 
rated, would their claims be substantiated if more than one 
pair of score comparisons were made? If more than one pair 
of raters was used would the attained scores between raters, 
who rated the same individual, be similar or dissimilar? 
Furthermore, would the scores obtained by trained raters be 
similar or dissi.milar to those obtained by untrained 
raters? 
The reliabi.11 ty study reported in the 1974 edition 
of the manual was based on a revision where only Part One 
items were changed. Because Part Two remained unchanged 
and because the reliability for Part Two was reported to be 
considerably lower than Part One, this study will be spe­
 cifi.cally concerned wi tb Part Two of the AAMD Adaptive 
Behavior Scale. 
4
 
Statement of the Problem 
Is there a significant relationship between the 
scores obtained by the untrained raters on Part Two of the 
AAMD Adaptive Behavior Scale? 
Is there a significant relationship between the 
scores obtained by the trained raters on Part Two of the 
AAMD Adaptive Behavior Scale? 
Is there a significant relationship between the 
scores obtained by the untrained and trained raters on 
Part Two of the AAMD Adaptive Behavior Scale? 
Statement of the Hypotheses
 (Null Form) 
There is no significant relationship between the 
scores obtained by the untrained raters on Part Two of the 
AAMD Adaptive Behavior Scale. 
There is no significant relationship between the 
scores obtained by the trained raters on Part Two of the 
AAMD Adaptive Behavior Scale. 
There is no significant relationship between the 
scores obtained by the untrained and trained raters on 
Part Two of the AAMD Adaptive Behavior Scale. 
Purpose of the Study 
The intent of this study was to investigate the 
inter-rater reliability claims of the AAMD Adaptive Behav­
 ior Scale. Specifically, the scores obtained by untrained 
and trained raters were compared. An additional reason 
for pursuing this investigation was that the AAMD study used 
5 
only one pair of raters from which their reliabiJjty data 
was obtained. However, the present study was designed to 
correct this deficiency by comparing the scores between 
three pairs of untrained and trained raters. 
Significance of th~_Study 
Since the recognition of the importance of adaptive 
behavior in determining social functioning, it became nec­
 essary to establish an instrlwent that could accurately 
measuxe adaptive behavior criteria. The AAMD Adaptive Be­
 havior Scale has claimed such accuracy. If this study re­
 produced the evidence indicating good inter-rater relia­
 bility, it would help to solidify these claims, and thus 
support the view that this instrl~ent is a good measuring 
tool for adaptive behavior. On the other hand, if this 
study did not reproduce the evidence indicating good inter­
 rater reliability, the claims made by adherents of this 
scale may not be supported. In either event, the results 
of this study can only add more information into the search 
for an appropriate means of identifying adaptive behavior 
indices. 
DEFINITIONS OF TERMS 
The terms to be used in this study consist of defi­
 nitions of adaptive behavior and reliability. In addition, 
terms relating to the levels of mental retardation are 
provided. 
6 
Ada~~i~e Behavior 
This term is defined as the ability of an individ­
 ual to adapt to the natural and societal demands within his 
environment. In addition, it means the degree to which a 
person can function and maintain himself independently, and 
the degree to which he satisfies the cultural dema~ds of 
social and personal responsibility (Nihira, 1969). 
Reliability 
Reliability in this study was defined operation­
 ally. Specifically, reli.ability was interpreted as the 
level of inter-rater agreement which was reflected by 
correlation coefficients derived from either the Pearson 
product-moment (r) or the Kendall Coefficient of 
Concordance (W). 
Intelligence Quotient (IQ] 
The IQ is defined as the score obtained from an 
individual's performance on either the Stanford-Binet or 
the Wechsler Intelligence Scales (AAMD, 1973). 
Mental Retardation 
Mental Retardation is defined as the presence of 
significantly subaverage intellectual functioning as well 
as eXisting deficiencies in adaptive behavior that is 
demonstrated during the developmental period (AAMD, 1973). 
Mild Mental Retardation 
The mildly retarded person is defined as one who 
7 
scores within the range of 68-52 on the Stanford-Binet 
Intelligence Scale, or 69-55 on the Wechsler Intelligence 
Scale (AAMD, 1973). 
Moderate Mental Retardation 
The moderately retarded person is defined as one who 
scores within the range of 51-36 on the Stanford-Binet 
Intelligence Scale, or 54-40 on the Wechsler Intelligence 
Scale (AAMD, 1973). 
Severe Mental Retardation 
The severely retarded person is defined as one who 
scores within the range of 35-20 on the Stanford-Binet 
Intelligence Scale, or 39-25 on the Wechsler Intelligence 
Scale (AAMD, 1973)0 
Profound Mental Retardation 
The profoundly retarded person is defined as one 
who scores within the range of 19 and below on the Stanford­
 Binet Intelligence Scale, or 24 and below on the Wechsler 
Intelligence Scale (AAMD, 1973). 
LIMITATIONS OF THE STlIDY 
This study was limited to the determination of 
inter-rater reliability, that is the comparison of two sets 
of scores in order to ascertain their level of agreement 
with each othero No attempt was made to generalize to any 
dimension other than inter-scorer reliabtlity. The 
8 
sUbjects in this study were limited to child care workers 
from the Kansas Neurological Institute, Topeka, Kansas. 
The limitation of three pairs of raters was imposed because 
the institution's residential living units, in most cases, 
contained only three "morning" shift workers and three 
"afternoon" shift workers. The retardates were selected so 
that each unit taking part in this study had retardates in 
each IQ classification, i.e., each liVing unit was composed 
of mild, moderate, severe and profound mentally retarded 
persons. Due to the unequal sex and age distribution in 
the selected retardates it was not possible to control for 
these variables. A final limitation was that only compar­
 isons with Part Two of the AAMD Adaptive Behavior Scale 
were performed. This limitation was imposed because 
according to the information presented in the 1974 manual 
of the AAMD Adaptive Behavior Scale, the reliability coef­
 ficients for Part One were reported to be considerably 
higher than those on Part Two of the scale, and because 
only Part One of the scale was revised. 
Chapter 2 
REVIEW OF RELATED LITERATURE 
The AAMD Adaptive Behavior Scale was published in 
1969, but in 1974 it was revised. The scale was designed as 
a behavior rating scale for mentalJy retarded, emotionally 
maladjusted, and developmentally disabled persons. Gen­
 erally, the term "adaptive behavior" refers to an individ­
 ual's effectiveness in coping with the natural and social 
demands of the environment. The IQ score was not able to 
measure the dimensions of adaptive behavior; therefore, the 
AAMD designed an instrument capable of measuring an individ­
 ual's level of adaptive behavior (Nihira, 1969). The devel­
 opment of this scale began in 1965 when the AAMD sponsored 
a project established at the Parsons State Hospital and 
Training Center. The aim of this project was to develop 
an understanding of adaptive behavior as it related to men­
 tal retardation and. emotional disturbance, and to develop a 
system of measuring adaptive behavior from infancy through 
adulthood. Furthermore, this project would facilitate an 
improved understanding of adaptive behavior, and would lead 
directly to improved methods of evaluation and treatment of 
mentalJy retarded persons in terms of present needs and in 
terms of long-range goal planning (Leland et al., 1967). 
Nihira and Shellhaas (1970) summarized the findings 
9 
10 
of the Adaptive Behavior Project which directly led to the 
formulation of the Adaptive Behavior Scale. They reported 
that Scott's (1966) adaptive strategies: accommodation, 
locomotion and construction were accepted as the concepts 
underlying adaptive behavior. Accommodation is the acqui­
 sition of patterns and traits that satisfy eXisting environ­
 mental requirements. Locomotion involves movement in the 
search for an environment congenial to the individual's 
present behavior patterns and traits. Construction is 
changing the environmental requirements so that the environ­
 ment becomes more congenial to the individual's resources. 
Three types of rehabilitation programs were conceptualized 
from these three adaptive strategies. The first and most 
frequently occurring. accommodation. is the ability to mod­
 ify or develop patterns of behaviors or traits that will 
enable the retarded person to cope with existing environ­
 mental demands. In locomotion the retarded person must find 
an environment that will accept his limitations. and, of 
course, he needs to be placed in that environment. Finally, 
construction requires that the environment be changed to the 
extent that it becomes more accepting of the retarded per­
 son's existing patterns of behaviors and traits. In order 
to utilize these three approaches information must be 
obtained concerning the individual's present behavior pat­
 terns and traits, and the demands and requirements of the 
environment. The development of the Adaptive Behavior Scale 
was an attempt to find this kind of information. 
11 
In 1961, when the AAMD declared that an individual 
had to have deficiencies in both measured intelligence and 
adaptive behavior before that person could be diagnosed 
mentally retarded, it became necessary to develop a special 
program of research (Leland et al., Note 1). Researchers 
realized that in order to develop norms of adaptive behav­
 ior, they had to make comparisons between individuals. 
According to Nihira, Foster and Spencer (1968) such a com­
 parison would require an analytical process. They indi­
 cated that a means was needed to find the "common denomi­
 nator" of such complex behavorial phenomena, dimensions of 
adaptive behavior. Multivariate analysis was presented as 
the means to describe such phenomena by identifying a set 
of dimensions that would provide a quantitative description 
of the nature and variation of human behavior by the use of 
systematically obtained observations of consistencies and 
patterns of behavior. The application of a multivariate 
research approach to the study of adaptive behavior 
required the discovery of factors that can be replicated 
in two or three different samples, to form hypotheses con­
 cerning the nature of these factors, and to continue 
experimentation with other pertinent factors relating to 
the adaptive behavior dimension. The major areas accepted 
by other rating scales were examined in order to obtain 
adequate samples. This examination included the behavior 
rating scales presently part of the public domain in both 
the United States and Great Britain. In addition, a list 
12 
of significant behaviors was obtained from institutional 
ward personnel via semistructured interviews of ward per­
 sonnel. These behaviors were behaviors that required day­
 to-day care and supervision. A preliminary behavior 
checklist yielded 325 specific behaviors representing 10 
behavorial domains. These domains were: Independent 
Functioning, Physical Development, Economic Activity, Num­
 ber and Time Concept, Occupation (Domestic), Language 
Development, Self-Direction, Occupation (General), Social­
 ization and Social Responsibility. From the list of 325 
items, 211 items in the checklist were found to have a 
significant correlation with adaptive behavior, independent 
of measured intelligence, on at least one adaptive behav­
 ior level. These 10 behavorial domains were again found 
to be salient features of adaptive behavior when Nihira 
(Note 2) administered the Adaptive Behavior Scale to 458 
adolescents in a residential setting for the retarded. 
Nihira (1969) reported that a factor analysis of 
the 325 items contained in the preliminary Adaptive Behav­
 ior Checklist produced six factors which accounted for 94.4 
percent of the total variance in the correlation matrix. 
The sample studied was 919 adult, ambulatory, institution­
 alized retardates. The six factors isolated were: 
Personal Independence, Social Maladaptation, Institutional 
Difference, Intra-Maladaptation, Sex Difference and Age 
Difference. Personal Independence consisted of those 
behavior items representing profound social or 
13 
psychological withdrawal and a vegetative behavior pattern. 
The Social Maladaptation factor included destructive, 
rebellious, untrustworthy and anti-social behaviors, and 
personality difficulties suggestive of various negative 
attitudes toward the social environment. Institutional 
Difference was interpreted merely as the factor of insti­
 tutional difference. There were significant differences 
between the two institutions used in respect to the vari­
 ables of Occupation (Domestic), Number and Time Concepts 
and Language Development. The items pertaining to the 
Intra-Maladaptation suggested the presence of a self-depre­
 ciating and intropunitive process in the adaptive behavior 
sphere. Sex Difference was merely defined as the factor of 
sex difference. Female subjects were rated lower than were 
male subjects on the domains of Independent Functioning and 
Physical Development. The factor of Age Difference was 
found to be related to the variable of Sexually Aberrant 
Behavior. This variable was found to be inversely related 
to age, for example these behaviors had been observed more 
frequently among younger residents than among older resi­
 dents. Out of the six obtained factors, Personal Independ­
 ence, Social Maladaptation and Intra-Maladaptation were 
felt to have importance upon the quest for general dimen­
 sions of adaptive behavior. The three remaining factors, 
Institutional Difference, Sex Difference and Age Difference 
were described as control variables. 
In another study by Nihira (1969) factors similar 
14­
 to those of his adult study were found. The results of 
studying 313 institutionalized children revealed three 
major dimensions or factors related to the adaptive behav­
 ior process; they were Personal Independence, Social Mal­
 adaptation and Personal Maladaptation. The Personal Mal­
 adaptation factor found in this study was paralleled, and 
considered quite similar to the factor Intra-Maladaptation 
reported in the adult study. These factors were also found 
to be quite stable across age ranges that spanned from 
preadolescence through adulthood. 
Tomiyasu et al. (1974-) administered the AAMD Adap­
 tive Behavior Scale (revised) (Japanese translation) to 
1,917 retarded children and 6,092 retarded adults, and 
found that the significant factors in Part One of the scale 
were: Personal Independence, Social Adjustment and Person­
 al and Social Responsibility. The three significant fac­
 tors found on Part Two of the scale for both children and 
adults were: Anti-Social and Aggressive Behavior, Self­
 Stimulating Behavior and Deficient Interpersonal Behavior. 
In another recent study Guarnaccia (1976) had the 
Adaptive Behavior Scale administered to 4-0 retarded adults 
by their counselors at a vocational training center. A 
factor analysis revealed the presence of the following fac­
 tors: Personal Independence, Personal Responsibility, Pro­
 ductivity and Social Responsibility. While controlling for 
sex, verbal IQ, performance IQ and maternal trust, they 
found that the predictors together accounted for 75 percent 
15 
of the variance in the factor of Personal Independence, 
but very little of the variance in the other three factors. 
Validity studies completed on the AAMD Adaptive 
Behavior Scale consisted of factorial validity-factor anal­
 ysis studies of the domain scores and practical validity. 
Factor analysis studies of the domain scores isolated three 
major dimensions: Personal Independence, Social Maladap­
 tation and Personal Maladaptation (Nihira, 1969,1969). 
The Social Maladaptation and Personal Maladaptation were 
found to be independent of one another. In fact, the 
delineation of these two factors suggested that a retarded 
person with behavior disorders usually exhibits one of 
these two categories of response patterns. The Personal 
Independence and Social Maladaptation factors accounted for 
approximately 70 percent of the total variance of the 
group studied. 
Leland et ale (Note 1) studied 41 institutionalized 
retarded persons between the ages of 10 and 13, and found 
that the scores on Part One domains of the scale discrim­
 inated significantly between those who had been previously 
classified at different levels of adaptive behavior 
according to clinical judgement. 
In a study of 531 retarded institutionalized 
adults, Greenwood and Perry (Note 3) showed that all of the 
Part One scores and some of the Part Two domain scores 
significantly discriminated among those persons who had 
been placed into five homogenous administrative units, that 
16 
is, medical, educational, vocational, preplacement and 
release units in a residential treatment unit for mentally 
retarded persons. 
Foster and Foster (Note 4) reported a study based 
on 41 retarded children and adolescents. The results 
indicated that three domain scores from Part One, and Part 
Two total scores showed a significant change from pre-test 
over a two year period when intense operant treatment 
regimes were used. 
Another study investigated a group of 260 retarded 
subjects who were divided into groups considered psychia­
 trically and non-psychiatrically impaired. Six of the do­
 mains in Part Two of the Adaptive Behavior Scale were found 
to be significantly discriminating between impairment 
groups, although the groups had the same IQ and general 
level of functioning (Foster and Nihira, 1969). These six 
domains were: Untrustworthiness, Psychological Dis­
 turbances, Self-Abusiveness, Rebelliousness, Antisocial 
Behavior, and Violence and Destructiveness. 
Christian and Malone (1973) studied the relation­
 ship between WISe and Stanford-Binet IQ scores, Wide Range 
Achievement Test scores, and Adaptive Behavior Scale scores 
of 129 children and adolescents in a special education 
program. Significant correlations were obtained between 
Wide Range Achievement Tests and IQ scores, and between 
Adaptive Behavior Scale scores and IQ. 
Another study found significant relationships 
17 
between the class placement level of EMR (Educable 
Mentally Retarded) pupils and their respective domain 
scores on the AAMD Adaptive Behavior Scale. This tended 
to be even more significant since a population sample of 
2600 children was used (Lambert et al., 1975). 
Inter-rater reliabilities are reported in the man­
 ual of the AAMD Adaptive Behavior Scale (AAMD, 1974). This 
new revision was administered to 133 persons from three 
different settings. Each individual was independently 
rated by two ward personnel, who were from the "day" and 
"evening" shifts respectively. The mean rater reliability 
found on Part One of the scale was .86; the mean relia­
 bility found on Part Two of the scale was .57. These mean 
scores were based on the reliability coefficients of each 
individual domain score. In addition, these reliability 
results were based on inter-rater agreement, which is con­
 sidered to be one of the important factors to be weighed 
when interpreting behavior ratings (Leland et al., Note 5). 
Inter-rater reliabilities were obtained in con­
 junction with the factor analytic studies on which the 
scale is based. In the adult study 48 subjects were rated 
independently by two different judges with the resultant 
reliabilities ranging from .89 to .35 on the domain scores; 
the median reliability was .72. Between continuous vari­
 ables, the Pearson product-moment and Phi coefficients were 
used, and between dichotomous variables, the Biserial cor­
 relation coefficients were used (Nihira, 1969). Another 
18 
factorial study on children and adolescents reported inter­
 rater reliabilities performed on 48 subjects who were rated 
by two independent judges. The range of the reliability 
coefficients was from .89 to .35 on the domain scores, 
while the reported median was .72. The statistical tests 
used in obtaining the reliability coefficients were the 
same as reported in the previous study (Nihira, 1969). 
There has been some disagreement in regards to whether or 
not only trained raters should be used to establish the 
criteria in that trained raters would maximize objectivity 
so that more reliable results would be obtained (Leland, 
et al., Note 1). 
Congdon (1973) described modifications in the 
Adaptive Behavior Scale that produced scales limited to 
defining profound mental retardation and maladaptive behav­
 ior. This was achieved by dropping many of the high compe­
 tency items found in Part One of the scale. 
The Adaptive Behavior Scale has also been cited to 
be suited for further research in language acquisition 
(Perozzi, 1972). Language acquisition theories were found 
to be related to three aspects of adaptive behavior, namely 
maturation, learning and social adjustment. 
Regarding mentally retarded educational program­
 ming, Bogen and Aanes (1975) found that the Adaptive Behav­
 ior Scale proved useful in the development of behavorial 
norms for mentally retarded population grouping. The norms 
were then utilized objectively in determining individual 
19 
and group program needs. 
Before the 1974 revision of the Adaptive Behavior 
Scale it contained faults in both its content and scoring 
procedures according to Bhattacharya (1973). In his short 
report, the author pointed out that although the scale was 
based on the results of factor analysis, some domains over­
 lapped and some occupied two extreme points on a behavior 
scale. He indicated that the scale could be improved if 
changes were made. First, the number of points could be 
made uniform and more significant by making the distance 
between two points approximately equal. Second, in order 
to make the scale more comprehensive, new dimensions should 
be added. Third, since the negative personality traits are 
unsystematically placed in the scale, reorganization of the 
personality items was needed. Finally, his fourth point 
was that a scale profile containing information about the 
relative standing of each retardate on each variable was 
needed to make the scale more comprehensive. 
Chapter 3 
METHODS AND PROCEDURES 
Since only one study on inter-rater reliability was 
completed on the 197~ revision of the AAMD Adaptive Behav­
 ior Scale, the need for further research into inter-rater 
reliability with this instrument was apparent. The intent 
of this study was to investigate the claims of the AAMD on 
its revised Adaptive Behavior Scale, namely that this 
scale can be used accurately by untrained raters as well as 
by trained raters. Its claims, while based on a large 
sample of retardates in three different institutional 
settings utilized only one pair of raters from each insti ­
 tution. 
Although this present study used retardates and 
raters from only one institutional setting, it had more 
than one pair of raters complete the Adaptive Behavior 
Scale. In addition, it was decided that this study would 
differ from the AAMD study by having both trained and 
untrained raters instead of just using untrained raters. 
Since the AAMD manual reported that the reliability coeffi ­
 cients on Part Two of the Adaptive Behavior Scale were con­
 siderably lower than those on Part One, only Part Two 
scores were compared in this present study. In spite of 
utilizing only Part Two scores, the raters in this present 
20
 
21 
study completed both Part One and Part Two of the AAMD 
Adaptive Behavior Scale. Information regarding the popu­
 lation sample of retardates and raters chosen are explained 
in this chapter. In addition, the design of the study, 
collection of data and data analysis applied are also 
detailed. 
POPULATION AND SAMPLING 
It was decided to select three different resi­
 dential units from the institutionalized retardates of the 
Kansas Neurological Institute, Topeka, Kansas to be rated 
with the AAMD Adaptive Behavior Scale by child care workers 
who work with these retardates on their living units. The 
child care workers selected were those individuals who work 
the "morning" shift (6:00 a.m. - 2:00 p.m.) and those indi­
 vlduals who work the "afternoon" shift (2:00 p.m. - 10:00 
p.m.). While one of the units chosen had three workers on 
each shift, the other two units had only two workers on 
each shift. In order to have three pairs of raters from 
each unit, the child care worker supervisor from the "morn­
 ing" and "afternoon" shifts from these two units were 
selected to participate as raters in the two units that had 
only two pairs of raters. Since these two units were in 
the same administrative section, both units had the same 
"morning" and "afternoon" supervisor. Therefore, these two 
supervisors participated as raters in two of the units 
selected. Consequently, although these two units had three 
22 
pairs of raters, the number of raters from these two units 
totaled only 10 individuals. Added to the six raters in 
the remaining unit, the total of all raters from all three 
units was 16 individuals. Therefore, each of the three 
units chosen had three pairs of raters. 
In order to have trained raters as well as 
untrained raters, one pair of raters from each unit was 
selected to be trained; consequently, they received special 
instruction in the use of the AAMD Adaptive Behavior Scale 
which qualified them to participate as trained raters. 
Therefore, each unit consisted of one pair of trained 
raters and two pairs of untrained raters. Since the two 
child care worker supervisors served as raters in two of 
the units, they were selected to participate, after train­
 ing, as trained raters in two of the units. The trainees 
from the remaining unit were selected on the basis of their 
expressed interest in this study because training required 
additional effort and time. 
Although it was intended to have at lea.st two 
retardates in each retarded IQ classification from all 
three units, only one unit met the criterion of having two 
mild, moderate, severe and profound mentally retarded resi­
 dents. The second living unit had two retardates in each 
classification with the exception of having only one mildly 
retarded resident. The third living unit had retardates in 
each classification with the exception of having only one 
profoundly retarded resident. Therefore, one unit was 
23 
composed of eight retarded subjects, and the other two 
units were composed of seven retarded subjects each. The 
retarded sUbjects from all three living units totaled 22 
individuals. 
The population sample of retardates was not ran­
 domly sampled in that only three living units were found to 
have retardates in all four IQ classifications. The 
Stanford-Binet or Wechsler IQ scores for each retardate was 
obtained from institutional records. However, four of the 
IQ classifications were based on tests other than the 
Stanford-Binet and Wechsler IQ scales. Two IQ's were based 
on the Merrill Palmer Scale of Mental Tests, one IQ was 
based on the Cattell Infant Intelligence Scale, and another 
was based on the Interim Hayes-Binet. 
The age range and sex of the retardates in the unit 
composed of eight subjects were ages 14 through 21, with 
four males and four females. The age range and sex of the 
retardates in the unit composed of seven subjects with only 
one profoundly retarded subject were ages 17 through 22, 
with six females and one male. The age range and sex of 
the retardates in the unit composed of seven subjects with 
only one mildly retarded subject were ages eight through 
16, with seven males. 
In order to facilitate description, the living unit 
with four females and four males was labeled unit A. The 
living unit with six females and one male was labeled unit 
B. The remaining unit with seven males was labeled unit C. 
24 
MATERIALS AND INSTRUMENTATION 
The only materials and instruments used in this 
study were the AAMD Adaptive Behavior Scale, 1974 Revision. 
Specific instructions for completing the scale are pres­
 ented in the scale booklet. These instructions are repro­
 duced in Appendix A. Although the instructions are con­
 tained in the scale booklet, the administrator should be 
made aware of three important considerations. First, where 
certain items are not applicable to the individual being 
rated (for example, the item referring to money changing 
ability may be inappropriate where the individual has no 
opportunity to handle money) the rater can ignore the item. 
Second, where items deal with situations that are against 
regulations, such as using the telephone, the administrator 
must complete the rating, and indicate if the person could 
perform the task if it were allowed. This is done to 
insure that no one will be penalized for conditions beyond 
their control. Third, where items describe maladaptive 
behaviors not usually seen in very young children, such 
items as those referring to aggressive behavior may not be 
appropriate; however, the rater should try to record the 
behavior as accurately as possible (AAMD, 1974). 
The first part of the Adaptive Behavior Scale was 
designed for the measurement of an individual's skills and 
abilities in the following domains of behavior: independ­
 ent functioning, physical development, economic activities, 
25 
language development, number and time concept, occupation 
(domestic), occupation (general), self-direction, responsi­
 bility and socialization. The second part of the scale was 
designed for the measurement of the following domains: 
violent and destructive behavior, antisocial behavior, 
rebellious behavior, untrustworthy behavior, withdrawal, 
stereotyped behavior and odd mannerisms, inappropriate 
interpersonal manners, unacceptable vocal habits, peculiar 
or eccentric habits, sexually aberrant behavior, self­
 abusive behavior, hyperactive tendencies, and psychological 
disturbances (Nihira and Shellhaas, 1970). 
In Part One of the scale, scores are summed for 
each item statement, and the addition of these scores 
yields either a subdomain score or directly yields the 
domain score. The subdomain items are those statements 
that are included in the sphere of a particular domain, 
for example Domain II, Physical Development contains items 
dealing with two aspects of this domain, namely A) Sensory 
Development and B) Motor Development. Six of the domains 
in this part of the scale contain subdomains, while the 
remaining four domains do not have subdomains, and hence 
are directly scored from the sum of the item scores. Like­
 wise, in Part Two of the scale domain scores are the direct 
result of summing the item scores, since this part of the 
scale does not have any subdomain categories. To reit­
 erate, when the item scores are summed, the rater obtains 
the subdomain score depending on the presence or absence of 
26 
subdomains for that particular domain. The sum of the sub­
 domain scores produces the domain score. When the subdo­
 main and domain scores are computed, they are placed onto 
the Data Summary Sheet located in the back of the scale 
booklet (AAMD, 1974). A sample Data Summary Sheet for both 
Part One and Part Two scores was reproduced, and can be 
found in Appendix B. 
Although test profiling was not performed as part 
of this study, this test does provide a method for obtain­
 ing individual age related profiling. The test profile is 
achieved by entering the raw scores into the designated 
space at the bottom of the Profile Summary Sheet. Two such 
sheets are provided in the back of each scale booklet; one 
sheet pertains to Part One domain scores, and the other 
pertains to Part Two domain scores. The domain raw scores 
are then converted into percentile ranks by using one of 
the 22 age related tables. The age range of these tables 
is from three through 69. The obtained domain percentile 
score is then placed onto the Profile Summary Sheet. After 
this is done with each domain score, the adaptive behavior 
profile is obtained (AAMD, 1974). Reproductions of the 
Profile Summary Sheets for both Part One and Part Two of 
the scale are included in Appendix C. 
DESIGN 
After each of the three living units were selected, 
each of the 16 child care workers was interviewed by the 
27 
researcher in order to establish an amicable relationship 
that would enhance their interest and cooperation in their 
participation in this study. Following these interviews, 
and after the selection of those raters who were to par­
 ticipate in the study as trained raters, the researcher 
initiated the training procedure. The training sessions 
began with a session devoted to didadtic preparation. This 
included a discussion concerning the administration 
instructions contained in the AAMD Adaptive Behavior Scale 
manual as well as the instructions contained in the scale 
booklet itself. In addition, scoring procedures such as 
mathematical computation and counting were also discussed. 
After this session, the two pairs of trainees were asked to 
complete three ratings on retardates with whom they were 
familiar, and who were not subjects in the main study. The 
second and final session was devoted to pointing out the 
raters item score dissimilarities. In addition, their 
misconceptions concerning administration technique were 
discussed and clarified. Finally, an attempt was made to 
resolve their differing views of the retardate on the 
particular item being questioned. 
After the training session was completed, both the 
trained and untrained raters on each of the three living 
units chosen were asked to independently rate the assigned 
retardates from their respective living units with the AAMD 
Adaptive Behavior Scale, 1974 Revision. Due to the length 
of time required to complete each rating (45 to 60 minutes) 
28 
they were not required to complete more than one rating 
per day. This was done to insure that rater fatigue would 
not influence the test results. It was reported in the 
previous section that there were three considerations each 
administrator should be made aware of. These consider­
 ations were reproduced on an additional instruction sheet, 
and each rater received a copy prior to performing his 
first rating. This additional instruction sheet was 
reproduced, and can be found in Appendix D. 
DATA COLLEC TION 
Prior to each rating the child care worker was 
requested to follow the instructions contained in the scale 
booklet. The raters were not asked to complete scale pro­
 files from the raw scores they obtained. They were 
instructed that if they were unsure of how to score an 
item they should re-read the instructions in the scale 
booklet, or consult the additional instruction sheet pro­
 vided. In order to insure that each rater independently 
rated each child from his respective unit, each rater was 
given the scale booklet just prior to the time he was 
scheduled to go off duty for the day, and at that time he 
was informed of the retarded person's name whom he was 
to rate. After the child care worker completed each 
rating, he was instructed to deliver the booklet to the 
section secretary where it was collected either at 8:00 
a.m. or 3:00 p.m. daily until all ratings on the chosen 
29 
retardates were collected. After the completed scales 
were retrieved from the section secretary, each rating was 
checked for accuracy in addition and subtraction by this 
researcher before the scores were placed on the Data Summary 
Sheets as being true raw scores. 
DATA ANAI.JYSIS 
The data in this study were analyzed for the purpose 
of determining the degree of reliability between raters who 
rated the same person, or the inter-rater level of agree­
 ment. The data were grouped according to untrained/trained, 
untrained and trained rater pairs. 
Comparisons of the obtained domain scores between 
each pair of trained raters from each living unit were com­
 puted with the Pearson product-moment (r) and the corre­
 lation coefficients were obtained with the formula: 
Nl:XY - (L:X) (L:Y)
 rx•y ­
 / [NL:X 2 - (L:X)2J [NL:y 2 - (L:y)2J 
Comparisons of the obtained domain scores between 
each pair of untrained and each pair of untrained/trained 
raters were computed with the Kendall Coefficient of 
Concordance (W). and correlation coefficients were 
obtained with the formula: 
\J = s 
1/12 k2 (r~3 - N) 
30 
Both of these formulas were used because there was 
more than one variable, the variables were score data, and 
because there was one score for each rater. 
All computations were performed on a Rockwell hand 
computer, Model 3lR. 
The .05 level was selected as a significant 
relationship. The significance levels for rand W were 
obtained from significance tables for r (degrees of 
freedom = N - 2), and from sienificance tables for w. 
However, since the table for W was limited to N's of seven, 
the significance level for unit A, which had an N of eight, 
was obtained from the chi s~uare table (degrees of freedom 
= N - 1) after the data were converted with the following 
formula: 
x2 = k(N - l)W 
Tables were developed that showed domain corre­
 lations and levels of significance obtained by the raters 
according to the respective units where they completed 
their ratings. 
Chapter 1+ 
ANALYSIS OF DATA 
The data resulting from the scores obtained by the 
child care workers who completed the AAMD Adaptive Behavior 
Scale, Part Two were compared to ascertain the level of 
inter-rater agreement. The problem was to determine if a 
significant relationship existed between untrained raters, 
trained raters and between untrained/trained raters. 
RESPONSE ANALYSIS 
In Chapter Three it was indicated that three sepa­
 rate units of retardates would be rated with the AAMD 
Adaptive Behavior Scale by three pairs of child care workers 
from each respective living unit. Although both Part One 
and Part Two of the scale were completed, only those scores 
from Part Two were analyzed in this study. As indicated in 
Chapter Three, the child care workers were asked to rate 
each retardate from their respective unit with the Adaptive 
Behavior Scale within the time they went off duty and 
reported for work on the following day. However, there 
were varying degrees of cooperation in meeting these re­
 quested time limits. The raters from units A and B were 
most cooperative regarding the time limitations with the 
exception of the trained raters from these two units. In 
31 
32 
comparison, unit C raters were usually tardy in returning 
their booklets at the specified times. In addition, the 
raters from unit C failed to comply with the request to 
leave their completed booklets with the section secretary, 
while the raters from lmits A and B generally complied with 
this reql1est. Furthermore, more adding errors were made by 
the raters from unit C than either units A or B. Therefore, 
as a group, the raters from unit C tended to be less cooper­
 ative and more careless than the raters from the other two 
units. 
STATISTICAL ANALYSIS 
The Kendall Coefficient of Concordance, W, was the 
statistical method used to compute the level of inter-rater 
agreement between the untrained raters, and between the 
untrained/trained raters. Due to the inapplicability of 
using the Kendall Coefficient between the trained raters, 
the Pearson product-moment was used. 
In order to test the null hypotheses that there were 
no significant relationships between the untrained/trained, 
l.mtrained and trained raters, the significance level for W 
and r was obtained from tabled sources. The hypotheses were 
tested for each of the 14 domain scores from Part Two of the 
AAMD Adaptive Behavior Scale. In order to show the strength 
of relationships, the .05 level was chosen as being signif­
 icant. Statistical analyses were computed for all three 
units. The Kendall Coefficient of Concordance, W, and the 
33 
Pearson product-moment, r, coefficients as well as the 
respective significance levels for unit A appear in Table 
1, page 34. 
In unit A the relationships between the untrained/ 
trained raters did not achieve significance at the .05 
level in the following domains: VII, Inappropriate Inter­
 personal Manners; VIII, Unacceptable Vocal Habits; IX, Un­
 acceptable or Eccentric Habits; XI, Hyperactive Tenden­
 cies; and XII, Sexually Aberrant Behavior. Therefore, the 
null hypothesis that there are no significant relation­
 ships between the untrained/trained raters was accepted 
for these domains. The remaining domains showed signifi­
 cance to at least the .05 level. These domains were: I, 
Violent and Destructive Behavior; II, Antisocial Behavior; 
III, Rebellious Behavior; IV, Untrustworthy Behavior; V, 
Withdrawal; VI, Stereotyped Behavior and Odd Mannerisms; 
X, Self-Abusive Behavior; XIII, Psychological Disturbances; 
and XIV, Use of Medications. Therefore, the null hypoth­
 esis that there are no significant relationships between 
the untrained/trained raters was rejected for these do­
 mains. The relationships between the untrained raters 
from this unit that failed to achieve significance at the 
.05 level were: VII, Inappropriate Interpersonal Manners; 
VIII, Unacceptable Vocal Habits; IX, Unacceptable or Ec­
 centric Habits; X, Self-Abusive Behavior; XI, Hyperactive 
Tendencies; and XII, Sexually Aberrant Behavior. There­
 fore, the null hypothesis that there are no significant 
34­
 Table 1 
Unit A Rater Score Comparisons 
Domain Untrained~Trained 
Wa (x) 
I. 
II. 
III. 
IV. 
V. 
VI. 
VII. 
VIII. 
IX. 
X. 
XI. 
XII. 
XIII. 
XIV.
 
Note.
 
.574- (24.11)**
 
.868 (36.4-1)**
 
.810 (42.81)**
 
.663 (27.85)**
 
.609 (25.58)**
 
.462 (19.36)**
 
.24-3 (10.16)
 
.314- (13.15)
 
.318 (13.36)
 
.4-30 (18.06)*
 
.214- ( 8.99)
 
.322 (13.4-8)
 
.473 (19.82)**
 
.380 (15.92)*
 
Since the n for unit A was 
conversions for ~ was indicated. 
a~ =	 chi square 14.07, E(.05

 chi square 18.4-8, ~<.01
 
b~ =	 .707, E (.05

 .834-, ~ <.01
 
*E <.05 
**:Q <..01 
Untrained 
Wa (x2) 
Trainedb 
r 
-
 .738 (20.66)** .828* 
.861 (24-.11)** .968** 
.891 (24-.95)** .805* 
.664­ (18.95)** .830* 
.528 (14-.78)* .94-4-** 
.638 (17.86)* .224­
 .225 ( 6.30) .775* 
.34-2 ( 9.58) .271 
.301 ( 8.4-3) .733* 
.4-68 (13.10) .4-88 
.296 ( 8.23) .04-8 
.4-19 (11.70) -.362 
.628 (17.58)* .333 
.512 (14-.34-)* -.04-5 
eight, chi square (x2) 
35 
relationships between the untrained raters was accepted 
for these domains. The remaining domains showed signifi­
 cant relationships to at least the .05 level. These do­
 mains were: I, Violent and Destructive Behavior; II, 
Antisocial Behavior; III, Rebellious Behavior; IV, Un­
 trustworthy Behavior; V, Withdrawal; VI, Stereotyped Be­
 havior and Odd Mannerisms; XIII, Psychological Disturbances; 
and XIV, Use of Medications. Therefore, the null hypoth­
 esis that there are no significant relationships between 
the untrained raters was rejected for these domains. The 
relationships between the trained raters from this unit 
that failed to achieve significance at the .05 level were: 
VI, Stereotyped Behavior and Odd Mannerisms; VIII, Unac­
 ceptable Vocal Habits; X, Self-Abusive Behavior; XI, Hyper­
 active Tendencies; XII, Sexually Aberrant Behavior; XIII, 
Psychological Disturbances; and XIV, Use of Medications. 
Therefore, the null hypothesis that there are no signifi­
 cant relationships between trained raters was accepted for 
these domains. The remaining domains showed significant 
relationships to at least the .05 level. These domains 
were: I, Violent and Destructive Behavior; II, Antisocial 
Behavior; III, Rebellious Behavior; IV, Untrustworthy Be­
 havior; V, Withdrawal; VII, Inappropriate Interpersonal 
Manners; and IX, Unacceptable or Eccentric Habits. There­
 fore, the null hypothesis that there are no significant 
relationships between the trained raters was rejected for 
these domains. Those domains that achieved significance in 
36 
all three rater types in unit A were: I, Violent and De­
 structive Behavior; II, Antisocial Behavior; III, Rebel­
 lious Behavior; IV, Untrustworthy Behavior; and V, With­
 drawal. The domains that achieved significance between the 
trained raters, but not between untrained/trained raters 
were: VII, Inappropriate Interpersonal Manners; and IX, 
Unacceptable or Eccentric Habits. The domains that achieved 
significance between both the untrained raters and un­
 trained/trained raters, but not between the trained raters 
were: VI, Stereotyped Behavior and Odd Mannerisms; XIII, 
Psychological Disturbances; and XIV, Use of Medications. 
The only domain that achieved significance in the untrained/ 
trained raters, but not in either the untrained or trained 
raters was domain X, Self-Abusive Behavior. 
The relationships between the raters domain scores 
for unit B is presented in Table 2, page 37. For this 
group the domain score relationships between the untrained/ 
trained raters that failed to achieve significance at the 
.05 level were: VII, Inappropriate Interpersonal Manners; 
VIII, Unacceptable Vocal Habits; IX, Unacceptable or Eccen­
 tric Habits; X, Self-Abusive Behavior; XI, Hyperactive Ten­
 dencies; and XII, Sexually Aberrant Behavior. Therefore, 
the null hypothesis that there are no significant relation­
 ships between the untrained/trained raters was accepted for 
these domains. The remaining domains showed significant 
relationships to at least the .05 level. These domains 
were: I, Violent and Destructive Behavior; II, Antisocial 
37 
Table 2
 
Unit B Rater Score Comparisons
 
Domain Untrained/Trained Untrained Trained 
a CW Wb r 
I. .649* .665** .750 
II. .864** .883** .820* 
III. .548** .571* .935** 
IV. .700** .784** .688 
V. .533** .524* .186 
VI. .410* .516* .342 
VII. .149 .169 .000 
VIII. .329 .271 .558 
IX. .212 .215 .415 
X. .315 .263 .966** 
XI. .191 .203 .766* 
XII. .238 .343 .167 
XIII. .581** .546* .911** 
XIV. .434** .492* .471
 
Note. The n for unit B was seven retarded subjects.
 
aW = .333, B <.05 
.419, p. <: .01 
bW = .484, p. <.05
 
.591, B <.01
 
cr = .754, p. <.05 
.874, p. <.01
 
*Po <.05
 
**p. <.01
 
38 
Behavior; III, Rebellious Behavior; IV, Untrustworthy Be­
 havior; V, Withdrawal; VI, Stereotyped Behavior and Odd 
Mannerisms; XIII, Psychological Disturbances; and XIV, Use 
of Medications. Therefore, the null hypothesis that there 
are no significant relationships between the untrained/ 
trained raters was rejected for these domains. The domain 
score relationships between the untrained raters that 
failed to achieve significance at the .05 level were: VII, 
Inappropriate Interpersonal Manners; VIII, Unacceptable 
Vocal Habits; IX, Unacceptable or Eccentric Habits; X, 
Self-Abusive Behavior; XI, Hyperactive Tendencies; and XII, 
Sexually Aberrant Behavior. Therefore, the null hypothesis 
that there are no significant relationships between the un­
 trained raters was accepted for these domains. The remain­
 ing domains showed significant relationships to at least 
the .05 level. These domains were: I, Violent and De­
 structive Behavior; II, Antisocial Behavior, III, Rebel­
 lious Behavior; IV, Untrustworthy Behavior; V, Withdrawal; 
VI, Stereotyped Behavior and Odd Mannerisms; XIII, Psycho­
 logical Disturbances; and XIV, Use of Medications. There­
 fore, the null hypothesis that there are no significant 
relationships between the untrained raters was rejected for 
these domains. The domain score relationships between the 
trained raters that failed to achieve significance at the 
005 level were: I, Violent and Destructive Behavior; IV, 
Untrustworthy Behavior; V, Withdrawal; VI, Stereotyped Be­
 havior and Odd Mannerisms; VII, Inappropriate Interpersonal 
39 
Manners; VIII, Unacceptable Vocal Habits; IX, Unacceptable 
or Eccentric Habits; XII, Sexually Aberrant Behavior; and 
XIV, Use of Medications. Therefore, the null hypothesis 
that there are no significant relationships between the 
trained raters was accepted for these domains. The remain­
 ing domains showed significant relationships to at least 
the .05 level. These domains were: II, Antisocial Behav­
 ior; III, Rebellious Behavior; X, Self-Abusive Behavior; 
XI, Hyperactive Tendencies; and XIII, Psychological Distur­
 bances. Therefore, the null hypothesis that there are no 
significant relationships between the trained raters was 
rejected for these domains. Those domains that achieved 
significance in all three rater combinations in unit B 
were: II, Antisocial Behavior; III, Rebellious Behavior; 
and XIII, Psychological Disturbances. The domains that 
achieved significance between the trained raters, but not 
between the untrained/trained raters were : X, Self-Abusive 
Behavior; and XI, Hyperactive Tendencies. Those domains 
that achieved significance between the untrained and un­
 trained/trained raters, but not between the trained raters 
were: I, Violent and Destructive Behavior; IV, Untrust­
 worthy Behavior; V, Withdrawal; VI, Stereotyped Behavior 
and Odd Mannerisms; and XIV, Use of Medications. Those 
domains that were significant between the trained raters, 
but not between either the untrained or untrained/trained 
raters were: X, Self-Abusive Behavior; and XI, Hyperactive 
Tendencies. 
40 
The relationships between the raters domain scores 
for unit C is presented in Table 3, page 41. For this unit 
the domain score relationships between the untrained/trained 
raters that failed to achieve significance at the .05 level 
were: III, Rebellious Behavior; X, Self-Abusive Behavior; 
and XIV, Use of Medications. Therefore, the null hypoth­
 esis that there are no significant relationships between 
the untrained/trained raters was accepted for these do­
 mairls. The remaining domains showed significant relation­
 ships to at least the .05 level. These domains were: I, 
Violent and Destructive Behavior; II, Antisocial Behavior; 
IV, Untrustworthy Behavior; V, Withdrawal; VI, Stereotyped 
Behavior and Odd MaIlnerisms; VII, Inappropriate Interper­
 sonal MaIlners; VIII, Unacceptable Vocal Habits; IX, Unac­
 ceptable or Eccentric Habits; XI, Hyperactive Tendencies; 
XII, Sexually Aberrant Behavior; and XIII, Psychological 
Disturbances. Therefore, the null hypothesis that there 
are no significant relationships between the untrained/ 
trained raters was rejected for these domains. The domain 
score relationships between the untrained raters that 
failed to achieve significance to the .05 level were: III, 
Rebellious Behavior; IV, Untrustworthy Behavior; VI, 
Stereotyped Behavi.or B.nd Odd Marmerj 8ms; VII, Inappropriate 
Interpersonal Manners; IX, Unacceptable or Eccentric Habits; 
X, Self-Abusive Behavior; XII, Sexually Aberrant Behavior; 
and XIV, Use of Medications. Therefore, the null hypoth­
 esis that there are no sie;nificant relationships between 
Table 3 
Unit C rater Score Comparisons 
4-1 
Domain Untrained/Trained 
Wa 
Untratned 
W 
Trained 
r 
C 
-
 seven 
.785** 
.629** 
.4-36 
.4-63 
.609** 
.4-35 
.4-25 
.4-85* 
.4-27 
.285 
.593** 
.378 
.552* 
.201 
retarded 
.84-8* 
.665 
.223 
.728 
-.225 
.64-5 
.4-93 
-.111 
.671 
.311 
.565 
.74-3 
..516 
.679 
subjects. 
I. 
II. 
III. 
IV. 
V. 
VI. 
VII. 
VIII. 
IX. 
X. 
XI. 
XII. 
XIII. 
XIV.
 
Note.
 
.699** 
.591** 
.286 
.4-52** 
.503** 
.370* 
.4-13* 
.396* 
.527** 
.230 
.4-50** 
.4-66** 
.500** 
.304­
 The n for unit C was 
aw = .333, R(.05
 .4-19, :Q(.01 
bw = .4-84-, :Q <.05 
.591, R (.01 
cr 
-
 = .754-, R (.05
 .4-19, R (.01 
*R (.05 
**R <.01 
42 
the untrained raters was accepted for these domains. The 
remaining domains showed signifjcant relationships to at 
least the .05 level. These domains were: I, Violent and 
Destruc tive Behavi.or; II, Antisocial Behavi or; V, Wi th­
 drawal; VIII, Unacceptable Vocal Habits; XI, Hyperactive 
Tendencies; and XIII, Psychological Disturbances. There­
 fore, the null hypothesis that there are no significant 
relationships bet\"~f'r the l.mtrained raters was rejected for 
these domains. The domain score relationships between the 
trained raters that failed to achieve sienificance at the 
.05 level were: II, Antisociel Behavior; III, Rebellious 
Behavior; IV, Untrustworthy Behavior; V, Withdrawal; VI, 
Stereotyped Behavior and Odd Mannerisms; VII, Inappropriate 
Interpersonal Manners; VIII, Unacceptable Vocal Habits; IX, 
Unacceptable or Eccentric Habits; X, Self-Abusive Behavior; 
XI, Hyperactive Tendencies; XII, Sexually Aberrant Behavior; 
XIII, Psychological Disturbances; and XIV, Use of Medi­
 cations. Therefore, the null hypothesis that there are no 
significant relationships between the trained raters was 
accepted for these domains. Domain I, Violent and Destruc­
 tive Behavior, was the only domain that achieved signifi­
 cance at the 005 level. Therefore, the null hypothesis 
that there are no significant relationships between the 
trained raters was rejected for this domain. The domains 
that achieved significance between the untrained/trained 
raters, but not between the untrained or trained raters 
were: IV, Untrustworthy Behavior; VI, Stereotyped Behavior 
43 
and Odd Mannerisms; VII, Inappropriate Interpersonal 
Manners; IX, Unacceptable or Eccentric Habits; and XII, 
Sex1lally Aberrant Behavior. Ther.e were no domains between 
either the trained or untrained raters that were indep­
 endently significant. 
When considering all three units, the trained raters 
achieved significance less times than the untrained raters. 
However, when considering the combined lUltrained/trained 
raters, they achieved significance more frequently than did 
either the trained or untrained raters individually. 
Another relevant aspect when considering all three units 
together was that the trained raters had more correlation 
coefficients above + .50, which was interpreted as a fairly 
high level of inter-rater agreement, more often than did 
the untrained or the untrained/trained raters. 
The domains that more often achieved significance 
irrespective of untrained/trained, untrained or trained 
rater grouping were: I, Violent and Destructive Behavior; 
II, Antisocial Behavior; III, Rebellious Behavior; IV, tJn­
 trustworthy Behavior; V, Withdrawal; VI, Stereotyped Be­
 havior and Odd Mannerisms; and XIII, Psychological Distur­
 bances. The domains that achieved significance less often 
irrespective of the type of rater grouping were: VII, 
Inappropriate Interpersonal Manners; VIII, Unacceptable 
Vocal Habits; IX, Unacceptable or Eccentric Habits; X, 
Self-Abusive Behavior; XI, Hyperactive Tendencies; XII, 
Sexually Aberrant Behavior; and XIV, tJse of Medications. 
44 
Those domains that had the hiehest rate of achiev­
 ing signifi.cant relationships irrespective of the type of 
rater grouping were: I, Violent and Destr1J.ctive Behavior; 
and II, Antisocial Behavior. The domain that had the 
lowest rate of achieving a significant relationship irre­
 spective of the type of rater grouping was domain XII, 
Sexually Aberrant Behavior, which was followed closely by 
domains VII, Inappropriate Interpersonal Manners; VIII, 
Unacceptable Vocal Habits; IX, Unacceptable or Eccentric 
Habits; X, Self-Abusive Behavior; and XI, Hyperactive 
Tendencies. Domain II, Antisocial Behavior, had the high­
 est overall correlation average when all three units were 
considered together, while domain XII, Sexu.ally Aberrant 
Behavior, had the lowest overall correlation average when 
all three units were considered together. 
Chapter 5 
SUMMARY, CONCLUSIONS AND RECOMMENDATIONS 
A study was devised to test the level of inter­
 rater agreement when mental retardates were rated by child 
care workers with the AAMD Adaptive BehaNior Scale. Based 
upon the results of this study it was found that untrained 
raters achieved significant relationships in their domain 
scores more often than did trained raters. From these 
results it could be concluded that training had a negative 
effect; however, this result could also be related to the 
training technique, or other factors. 
SUMMARY 
After the revision of the AAMD Adaptive Behavior 
Scale in 1974, the AAMD cited a reliability study on the 
revised scale. This study was based on 133 institution­
 alized retardates who were rated by two ward personnel. 
The AAMD claimed that its scale can be effec­
 tively used by individuals with little or no training. 
This claim, according to the AAMD manual, is based on the 
relatively simple scoring and administration procedures. 
Since the AAMD manual does not indicate that the raters 
used in the 1974 study were trained raters, it was 
assumed that they were not trained. The range of the 
45 
46 
reliability coefficients in the 1974 study was from + .37 
to + .77 on Part Two of the scale, and the mean relia­
 bility computed for Part Two was + .57. 
The AAMD specified, in 1961, that a diagnosis of 
mental retardation must include deficiencies within the 
dimensions of adaptive behavior as well as measured intel­
 ligence. Therefore, it became incumbent on the AAMD to 
develop a means for measuring adaptive behavior. In 1969, 
the AAMD developed an instrument proported to measure the 
adaptive behavior dimension, which was revised in 1974. 
This study was proposed in order to empirically 
test the claims that the Adaptive Behavior Scale can be 
administered effectively by untrained raters. In 
addition, it was decided that a reliability study based on 
only one pair of rater scores was insufficient. There­
 fore, this study undertook the task of using three pairs 
of raters of which one pair was trained in administration 
and scoring procedures, while the remaining two pair were 
not trained. Three different living units with retardates 
in each IQ classification of mental retardation were 
chosen from a state institution. Each unit was rated by 
three pairs of raters. The three units were composed of a 
total of 16 raters and 22 retardates. 
The Kendall Coefficient, W, was used to test the 
null hypotheses that there were no significant relation­
 ships between the scores of the untrained/trained and 
untrained raters in each of the 14 domains of Part Two of 
47
 
the AAMD Adaptive Behavior Scale. The Pearson product­
 moment, r, was used to compute the score relationships 
between the trained raters. The results varied from one 
unit to another. The trained raters achieved less domain 
significance than either the untrained/trained raters, or 
the untrained raters. However, it should be pointed out 
that the trained raters accum.ulated more domain corre­
 lation coefficients above + .50 than did the rater combi­
 nations of untrained/trained and untrained raters when all 
three groupings of units were considered together. 
Those domains that achieved significance across 
all rater types with only two exceptions were domains I, 
Violent and Destructive Behavior; and II, Antisocial 
Behavior. Both exceptions were the trained raters from 
units B and C; however, their respective correlation coef­
 ficients, in both instances, were well above + .50, which 
were considered rather high levels of inter-rater 
agreement even though significance was not attained. When 
considering the trained raters in all three units, no 
domain showed consistency in obtaining significance. How­
 ever, when considering the grouping of untrained/trained 
and untrained rater relationships in all three units, four 
domains were found to be consistently significant. These 
domains were: I, Violent and Destructive Behavior; II, 
Antisocial Behavior; V, Withdrawal; and XIII, Psycho­
 logical Disturbances. Therefore, when considering these 
two rater types in all three units together, the null 
48 
hypotheses that there were no significant relationships 
between the untrained/trained and untrained raters was 
rejected for these four domains. It follows that the null 
hypotheses were accepted for the remaining 10 domains when 
considering all three units and rater types together. 
Furthermore, the null hypothesis is also accepted for the 
trained raters when considering all three units together. 
However, when considering each rater type with 
his respective unit acceptance and rejection of the null 
hypotheses that there were no significant relationships 
between the rater types varies from unit to unit. Spe­
 cific information relating to each of the three rater 
types can be found in Chapter Four, but briefly unit A 
achieved more significant relationships in the different 
domains than did unit B or C, and unit B achieved more 
significant relationships in the different domains than 
did unit C. 
CONCLUSIONS 
The problem encountered in this study was to 
determine if there is a significant relationship between 
the scores of the untrained/trained, untrained and trained 
raters. Referring to the results discussed in Chapter 
Four, the data indicated that trained raters achieved sig­
 nificant relationships less often, and on less domains 
than did the untrained raters. Therefore, this suggested 
that training, for one reason or another, produced a 
49 
negative effect. Specifically, the particular method of 
training raters in this study may not have been the most 
effective method for training raters, and may be highly 
related to the negative findings between the trained 
raters in this study. In Chapter Four it was pointed out 
that rater attitudes were reflected by the degree of coop­
 eration in complying with the researcher's instructions. 
The least cooperative raters were found to be the trained 
raters from units A and B, and all rater types in unit C. 
Therefore, this negative attitude could also be a factor 
that led to the negative effect produced by the trained 
raters. Finally, the factor of the amount of time the 
trained raters normally spent with the retardates in com­
 parison to the amount of time the untrained raters spent 
with the retardates could also affect the inconsistencies 
between the scores of the trained and untrained raters. 
This factor was mentioned because the two child care 
worker supervisors, who functioned as trained raters in 
units A and B, do not spend as much time in direct super­
 vision of the retardates as do the untrained raters. How­
 ever, since the trained raters from unit C were regular 
child care workers and not supervisors, and since they 
achieved significance less often than the trained raters 
from either units A or B, this factor was not considered 
very significant. 
Since each of the three units had varying degrees 
of success, what variables may have influenced this 
50 
variability? Since unit C was less cooperative in follow­
 ing the researcher's instructions, and since as a group, 
unit C achieved significance on the domains less often 
than did either units A or B, the attitude or motivation 
couJ.d be a related factor to the result that unit C 
achieved significance on the domains less often than did 
the other two units. 
Unit A achieved significance on more domains than 
did unit B and unit B achi.eved significance on more 
domains than di.d unit C. Since the age ranges of the 
retardates was highest in unit A and lowest in unit C, 
the factor of age range could also be related to the 
results of this study. 
Unit A, in comparison with units Band C, was 
equally divided by sex having four females and four males. 
Therefore, equally divided sex groups could be a variable 
of some importance since unit A had more significant rela­
 tionships than the other two units. It should also be 
noted that unit A was composed of eight retardates, one 
more than either units B or C. 
As indicated in Chapter One, another reason for 
pursuing this study was to see if the data from this study 
supported or failed to support the results of the relia­
 bility study cited in the AAMD manual. Based upon the 
reliabilities of the raters in unit A of this study, the 
reliability coefficients reported in the AAMD manual of the 
Adaptive Behavior Scale are supported for the following 
51 
domains: II, Antisocial Behavior; III, Rebellious Behav­
 ior; V, Withdrawal; and XIII, Psychological Disturbances. 
In unit B the following domains are likewise supported: 
I, Violent and Destructive Behavior; II, Antisocial Behav­
 ior; IV, Untrustworthy Behavior; V, Withdrawal; and XIII, 
Psychological Disturbances. In unit C this study sup­
 ported domains I, Violent and Destructive Behavior; V, 
Withdrawal; VIII, Unacceptable Vocal Habits; and XIII, 
Psychological Disturbances. Those domains not mentioned 
received lower reliability coefficients than were obtained 
in the study cited in the AAMD Adaptive Behavior Scale 
manual. 
Based on this data it was concluded that the 
results from Part Two of the AAMD Adaptive Behavior Scale 
in this study did not reflect or reproduce the results 
reported by the AAMD's reliability study since more than 
half of the 14 domains showed inconsistent inter-rater 
agreement. It was further concluded that the domain 
scores of the trained raters showed less significant rela­
 tionships when compared with the untrained raters. There 
was also less significant relationships when the trained 
raters scores were compared with the grouping of the 
untrained/trained raters scores. Finally, there were no 
significant relationships in any domain between the 
trained raters when all three units were considered toget­
 her. 
52 
RECOMME~mATIONS 
Contrary to the claims made by the AAMD, the data 
obtained in this study did not produce high levels of 
inter-rater agreement on Part Two of the Adaptive Behavior 
Scale as reported in the AAMD Adaptive Behavior Scale 
manllal. However, when viewing this scale from the per­
 spective of individual domains, domains I, Violent and 
Destructive Behavior; II, Antisocial Behavior; IV, 
Untrustworthy Behavior; V, Withdrawal; and XIII, Psycho­
 logical Disturbances, showed rather high inter-scorer 
reliabilities according to the data obtained in this pres­
 ent study. 
Based upon an examination of those domains with 
low inter-rater reliabilities, it appeared that some of 
those domains had subjective content. For example, domain 
XII, Sexually Aberrant Behavior, is an area that could 
mean different things to different persons depending on 
the individual's cultural mores. The same argument 
applied for domains VI, Stereotyped Behavior and Odd 
Mannerisms; VII, Inappropriate Interpersonal Manners; 
VIII, Unacceptable Vocal Habits; and IX, Unacceptable or 
Eccentric Habits. Although domains III, Rebellious Behav­
 ior; and X, Self-Abusive Behavior, appeared to have more 
objective content, they tended not to be as objective as 
those domains with higher reliabilities. Therefore, it 
is recommended that more research be implemented regarding 
53 
the improvement of the objectivity of these domain items. 
In addition to the faults in scoring and content of 
the Adaptive Behavior Scale reported at the end of Chapter 
Two, the dichotomization of scoring into "occasionally" 
signifying that the behavior occurs once in a while or now 
and then, and "freQuently" signifying that the behavior 
occurs Quite often or habitually appeared to be more of a 
subjective than an objective scoring system. In addition, 
the definitions for these terms appeared to be somewhat 
vague, for the booklet instructions do not specify the num­
 ber of times a behavior should occur before it is labeled 
"occasionally" or "freQuently". Therefore, a change in 
this scoring system to a more objective and less vague 
one is recommended. 
Although this study used two more rater pairs than 
did the study cited in the AAMD manual, the latter study 
used 133 retardates, while this study used only 22 retar­
 dates. Therefore, it is recommended that further research 
into the reliability of this scale consider increasing the 
number of retardates to at least 133 to correct what must be 
considered a weakness of this study. In addition, since 
age and sex may influence test results, it is recommended 
that further research into the reliability of this scale 
consider controlling for sex and age variables. Further­
 more, since attitude and motivation of the raters was con­
 sidered a possible factor related to the negative resul.ts 
in this study, it is recommended that future research in 
54 
this area consider these factors, and attempt to obtain 
raters who are both interested and motivated to participate 
in the study. 
Finally, it is recommended that if another study is 
undertaken using this particular study's method and proce­
 dure that a more intense training procedure be utilized to 
train raters in the scoring and administration procedures. 
This is recommended due to the negative effect of the 
trained raters in this study. 
Reference Notes 
1.	 Leland, H., Nihira, K., Foster, R., & Shellhaas, M. 
Proceedings of the First Congress of the International 
Association for the Scientific Study of Mental 
Deficiency, Surrey, England 1968. 
2.	 Nihira, K. Ten Dimensions of Maladaptive Behavior in 
Mentally Retarded Early Adolescents. Proceedings of 
the annual convention of the American Psychological 
Association, 7(Pt. 2), 1972. 
3.	 Greenwood, D., & Perry, R. Use of the Adaptive Behavior 
Checklist as a Means of Determining Unit Placement in a 
Facilitx for the Retarded. A Paper presented at the 
meeting of the Rocky Mountain Psychological Association, 
Denver, May 1968. 
4.	 Foster, R., & Foster, C. The Measurement of Change in 
Adaptive Behavior. A Paper presented at the Region V 
meeting of the American Association on Mental 
Deficiency, Denver, October 1967. 
5.	 Leland, H., Nihira, K., Foster, R., & Shellhaas, M. 
Conference on Measurements of Adaptive Behavior: II, 
Parsons State Hospital and Training Center, Parsons, 
Kansas 1966. 
55 
References 
American Association on Mental Deficiency. AAMD Adaptive 
Behavior Scal~ Manual (Rev. ed.). Washington, D. C.: 
Author, 1974. 
American Association on Mental Deficiency. A Manual on 
Terminology and Classification in Mental Retardation 
(Rev. ed.). Washington, D. C.: Author, 1973. 
Bhattacharya, S. Adaptive Behavior Scale Refinement. 
Mental Retardation, 1973, 11(1), 27. 
Bogen, D., & Aanes, D. The Adaptive Behavior Scale as a 
Tool in Comprehensive Mental Retardation. Mental 
Retardation, 1975, 13(1), 38-41. 
Christian, W. P., & Malone, D. R. Relationships among 
Three Measures Used in Screening Mentally Retarded for 
Placement in Special Education. Psychological Reports, 
1973,33(2),415-418. 
Congdon, D. M. The Adaptive Behavior Scale Modified for 
the Profoundly Retarded. Mental Retardation, 1973, 
11(1),20-21. 
Foster, R., & Nihira, K. Adaptive Behavior as a Measure of 
Psychiatric Impairment. American Journal of Mental 
Deficiency, 1969, 74(3),401-404. 
56 
57 
Guarnaccia, V. J. Factor Structure and Correlates of 
Adaptive Behavior in Noninstitutionalized Retarded 
Adults. American Journal of Mental Deficiency, 1976, 
80(5),543-547. 
Lambert, M., Windmiller, M., Cole, L., & Richard A. 
Standardization of a Public School Version of the AAMD 
Adaptive Behavior Scale. Mental Retardation, 1975, 
13(2),3-7. 
Leland, H., Shellhaas, M., Nihira, K., & Foster, R. 
Adaptive Behavior: A New Dimension in the Classifi­
 cation of the Mentally Retarded. Mental Retardation, 
1967,4(3),359-378. 
Nihira, K. Factorial Dimensions of Adaptive Behavior in 
Adult Retardates. American Journal of Mental Deficiency, 
1969,73(6),868-878. 
Nihira, K. Factorial Dimensions of Adaptive Behavior in 
Mentally Retarded Children and Adolescents. American 
Journal of Mental Deficiency, 1969, 74(1), 130-141. 
Nihira, K., Foster, R., & Spencer, L. Measurement of 
Adaptive Behavior: A Descriptive System for Mental 
Retardates. American Journal of Orthopsychiatry, 1968, 
38, 622-634. 
Nihira, K. & Shellhaas, M. A Study of Adaptive Behavior: 
Its Rationale, Method and Implications in Rehabilitation 
Programs. Mental Retardation, 1970, 8, 11-16. 
Perozzi, J. A. Language Acquisition as Adaptive Behavior. 
Mental Retardation, 1972, 10(2),32-34. 
58 
Tomiyasu, Y., Matsuda, S., & Murakamai, E. Y. Structure 
of the Adaptive Behavior of the Mentally Retarded: 
I. A Factor-Analytic Study. Japanese Journal of 
Special Education, 1974, 12(1), 10-23. (Abstract) 
6~
 
areos .I0 lAB-q8g: 8A1=t-dBPV 
awvv 8qt-BU1+81dUloo .IoJ sU01=t-0n.It-suI 
V XIaN~(IdV 
60
 
AAMD 
ADAPTIVE BEHAVIOR SCALE 
For Children and Adults 
1914 Revision 
N.mc (I..t) (first) 
Speci ..1 
ldentifiuuon 
D.le 
(mol (day) (y..tl 
S..:~ Date "f B"t/I 
(mol (day) {~c:~;.)-
 Nil1IO of penon flllin~ oul Scale _ 
Source of info,m"'liun .md relollion~hip 10 perW>n be,nK eV.J.lu.tC'd (weh 015 "Juhn Doc . P..rtnl," .,f "Self· 
Pl\y..'i.....I _ 
Addition.llnform.tion: _ 
Thi' Sui, ,on~i'l' of il nUHlbcr of s141emenls which deKrihc 50rne of the WolY "cop" ..,t in dilferent __ih""'h""l,). 
ThaI:: Me 5Cller..1 WioYS of .dmini\lcring the S"de; thete, .and "h:uiled \4.:uring :'hln,iccium,• ..appC'Jr In the 
KComP4nvina Manuill. 
InsulICtions (0' lhe Mcond pMt of lhe Sui. immedi"'lcly precede [he Wl".und h.. lf of Ittl.. bO<Jllcl. 
INSTRUCTIONS FOR PART ONE 
There arc two Idnds of llem, in the first poilU of the Su.I•. The I'lht require, IhJI 'l'uU ~clt'ct only ONE uf lhe 
lI.e,al poHibl. ,elpon"l. For .umple: 
IZI E.Uft. ift '.blic ""e1. onl~ ~) 
Ord...n COf1lvlrlP m",.I, In r",-.fuar",nn 
Orden IImple mIP..h ,,\LIP holmbu,.en 
or hoi do.,) 
O'k'" \Oh d,,"It., Ilt ~. to~n,••n 
Of une.." °0
 Due, nat arcH, .t public eIll.nl pi ..,., 
NOli,. lhal lh••tilemont. aro a,,"nged In urder of diffl,;YllY: 1,2,1,0. Ci"le Ih. UI1' Il"ement ",hi.h best 
.Krlbtl th. mai'difli,ufr t,U" tho person , ..n l,nually man.,~\}. tn thn, 1l)t04mple. Ihll incSi\llduoIl LUBn, <Jbwr'lt\:d .;.a.n 
or_ limpl. mull Iik. hambYrs." ur hot dose (2), bYI ..nnu! urder. ,omplet. dlllll" (3). rhor.fore. (2) i, "'dod 
In lilt ...mpl••bovo. In "orins. 2 II enlered in 'ho ,j"le 10 th. ,11lh1. 
C lllOll, la74, la76 Am."can Anacillion on M.nll' O.lici.noy 
61
 
The IeC:Dftd type of item nils you to chKk ALL ttMemenu whl(h appIv .... ,h. pnWfl. Fot ...mp&lo 
141 T...... -.. 
(Check ALL 'Lil..m~n'~ _"lIch .pply) 
8·numbe1' 
5.,,110"" food without ctwwlnl chKlted .a. 
aww, food .....uh mouth oven ..:L 
Drop, food on tlibl. ~ tilAU' 
Uwt n.(Jllin Incorrectly ~ not Ii' .11 32 
1.lk, .....Ith mouth j ull o 
T<&keto 'GOd off Olh.r~· pl••" 
E.l~ 100 fUI IX 100 slow 
Pwys In food ..... Ith Ilnile,S 
_.ltlwobooe __ 
Qrop, not .lJp1v.• I . bt-l' ..use ... IX 
~t' ... complt',.ly o.prndt-nl un
 
ocht"n (If cht'<:lr.ed. _onl.r '·0'· In
 
ltw (1((1. lQ c... ,llj,hl )
 
In 11\, ...",ple obove. lI\e ,""nd o"d fourth Iteml .,e ,h.r;k.~ to indi<ote thot the pc"on """WI food ";111
 mow'" open" lind "ut.et n~kin inc;orrectly." In s.coriNI, the number of Ilcm\ checkc:d. 2, I~ ,>ubtr.u;led trom tJ,.it1d 
thl Item ItOr., 6, is entered in the circle to the riaht. Mo~t items. do nOl, however, require thii \UbtroiLlll>n; ,"sle~. 
the number ChKkcd c.ln be £iirectly entered.lS the iGore. The \uu:mcnt "None of the ibovc," whu;h.i~ included fO' 
odmlnillroll.e purpowl only. i. not to be ",unted in Korinl here. 
Some Items moy d... with beh.viors lI..t .,. ,le.,ly o..inS! 1"'01 r.~ul.tion •• (e.~ .• use ot the ",Iephonel, 01 
behivion thit ill not poS\ible for i per~n to perform becau~ the upponunity duC\ not u,in, (e.I., tatln, In 
reluur,ants II not po~.jb'e fur KJmeone who i. bedridden). In these inSlan,,·~., you muu nil! ~",mplete your ral,n;. 
GI•• th' porton "edit tor 11\, item if you feel obtolutely «ruin thot he or ,he <4n .nd would perform the beh ••iar 
wlll\out oddltionol trolninl h.d h. or .he th. opporlunilY to do to, W, ite "AR" fur "AK.,n,t IIrK"I.t;o ..·." 01 
"HNO" for IlHas No Opportunity" n' .. t to the r,atin. mide in the.. '.-tet. Thl'" nol,Ulon, will not ,jrfed the 
, ..ntuM "orlnl ot thot item, but will ,onllibute io 11\, underoundinl .n~ interp« lion of the per",n's .~.pti.e 
blho.lor oncl environment. 
",,,, obM'" 11\, tollowln. pnerol rul.. In ,ompl,tln. the 5,01" 
1. In IlIml w111'h lpo,lfy "with h,'p" or "with u.l,ton"" for completion 01 l.I.k, thU' mean with dirKr 
pII'lfluI.,,;,uM•. 
1	 GI.. the porton "edit for on Item ..en if h' or she needs verbo' promptlnK or remindIn. ro ,ampl.te the us. 
unl... th' Item definitely SUte, "wlthour promprln," or "wi/hour remimJ"." 
Thl. SQI, II propared for IOneraJ u... Therefore. tome of tile Item. m.y nOl be .ppropriate tor your .pe<ifi, 
..tdlll. but pi.... do try to ,omplete 011 of them. 
62
 
INS TRUCTIONS FOR PAR T 1>\'0 
Part Two conlam. only one type of ir..m. the; lollowjn~ IS an e.ample 
\1\ 0 ........." .., ..... P'OIWrt,.
 
f.l.,P\. 1..,J.r\_ Of" c:h~"" .. 0",,('1 dOlhlnl 
S,ud .. (Mon pro~rt.,. 
rt!'.Ir~ up g""n ,n"loIllnt'~. buuIr.\, 
Of' 01n.r pcI)'lor""lOn} 
Otht"l t \Pf'fltv 
_ None' ., tht .bolo.. 
Select those 0' Iht" ~t.lt~m~nt§, which cUt' true of Ihe mdl\llduoJl bt'IOIo:
 
eVAluated, dod clrclt' (lJ If the beohd .... lor uccuri oceJslonally, or ti) II It 0< (Or)
 
frequently Check "Non~ of the Abo"e" when:· etpproprl~llt· In "(lJrln~. tul.&!
 
elleh column on tht> bottom (l otal) line. dod !pnler th~ ,"Unl 01 th ....... fotdol'io In tht>
 
circle to the 'IKohl When ·'Nont> of th.' J.hovt'" h (h~cked. (-nft'r 0 In tlw
 
Circle to the rlKht In tht' dbo"e e)lamplt:, the fir'll ~tdhmH:'nl ai trut' U{C.l\IUI1<.1II..,..
 
and tht> Id~t two :Jl..ltemt'nh are true frequently, lheretore, 4 )(Qrt> of ) h.J')
 
been entered 
"Oc(ch)onally" s,lgnlfte~ lhdllh~ bt:'havlor (Jf(urs, once In d ~... hdp 01 no\-'Io dnd
 
then, and I"frequently' ,)Igmhe\ that the betld'w'lor O'llUn qUltt!' oltt'n. or
 
habItually.
 
• U,e the .pace lor "Oth"r" ."hen 
Th. per !ton has rpldted b.-hd\"or problem't In dlhllflon 10 lhosf' urllpct 
The penon ha' heh•• ,or problem, lh..t arc nul lO.Cfl·d b. an. (J( \he 
e••mpld Ii.led 
The b,h••,or Iitled und,r "Other" mu51 b" ••pe"',, (>,ampl. oj lhe
 
Iwh••ior proble'" tlated In Ihe Ilem
 
. t 
Some of III@ hem' ,n P"rt Two dt:'",.b~ bf,hd\llf)r't which ntlf'U nof Ol,l
 
COn\ldE'ft!'U mi1lad.lph\'t!' tQr ver~ 'YounM, l.htldren (tor l'k,IlIIpll', pu,hlflM mht·r .. j
 
Th. que'l,on 0/ whelh"r a 1I,••n b,-ha.,or It ad.pll ... or maladapl .." d"!"'lld.
 
on 'h. w.~ that pdrtlcular bphavltJr IS Vll'Wl"ll hy prop I..• In our ,Ollt"h'
 
Nonelhelell, ,n complel,oll th" Scal.. you ar" ",kl!d !O ,,-,"rei. P.""o',
 
behhlor .. Ilcural.ly a' ~/)."hll!, '/l"OIlOIl, f"I' tho mo,m'nr, ,u", 111"'''0,,1
 
bl'''I, lhlfn, wh'," yuu lalor iOlt'rpl'l>tlh~ ,mpact ui th. r"purleLl bOI1J.,O", ,uu
 
'hould lakl! IOto ~un"deratiun ,o"otal altllutJe.
 
11 
t:9 
a: XlaN~ddV 
'1
 ;i 
'1 
1 
:1 
I
 1 
~ 
.~ 
64
 
. _-----_.__.__ . _._-_. ~ .. ----------------_..... 
IOenHf"""lIon 
All" 
Sox 
O.lt of AdlT\ifll~U.llon 
DATA SuMMARY SHEET· AAMO ADAPTIVE BEHAVIOR SCALE
 
PART ONE
 
A. E.ring 
II. rj".lilir Uj•.
 
C CIUfllm.u.
 
D. Ap~f.n ...... 6f~"l:~A66 " 
E. e.f. of ClvthlllQ 
F. Df.ui'¥J & UndftJuln/l 
G. Tf.~/ 
·: .. ·.L., ., '-1 
H. G.MfM Ind.p.ndllnt I . ',dlonin/l 
... -_...-+ C.-J 
I. INOfPfNDfNTfUNCTlONINCi. 
A. s.nlDry D'~/opm.lJt.. 
II. Motof D.II.'opm,nr 66 
II. PH'fSICAL DEV;'LOPMfNT 
A. Mo,...y H.ndlm~ -I1f,", BLJdg.ring . 
tj Shopping SA III, 6./\·0__I
 .... L. \, -.JIII. ff<QNQM/C ACT.lJt../.l.X. 
-10-1 .. 
A. EApr,sua" . 
B. Comp,.htJnllon. l~'~66C. $ocI.,1 LrJrJgUJg. D.~.'opm*m. 
.., .. CJIV. LANGUAGE DEvELOPMENr 
V. NUM8EIJ5 AND r,,~lf 
·CJ 
A. C/,.",ng
 
IJ. Kirc"'" DUll.'
 
C. Den.' Dom*IlIC Ai.:rllli".~, 666 ,-,
 VI, POMEstlC ACTIVITY . .. L.-J~--~·CJVII. VOCArlONAb ACTlVlr'f 
A. 111/11.rl:,ll, .• , "."." 
- l~66
 B. h".""4n<" , , . , , .
 
C, L,i,~" rim'. , , , ,
 
".... ..Cj
 VIII. SELF·DIRECrION \0,. 
IX.~~ =--====---.C]
 -----.. C]X. SOCIAL IZA TION 
65
 
DATA SUMMARY SHEET 
'ART TWO
 .'. ~.' ,: 
..~. 
I. VIOLENT AND DESTRUCnVE BEHAVIOR 0 
/I. ANTISOCIAL BEHAVIOA /I
 . 0 
III. REBELLIOUS BEHA IIIOR 1/1 
IV. UNTRUSTWORT/I¥ BEHAVIOR IVB 
II. WITHDRAWAL V0 
VI. STEREOTYPED BEHAVIOR AND 000 MANNtlIIS",S VI0 
V/I. INAPPROPRIATE INTERPERSONAL MANNERS V/I0 
VI/I. UNACCEPTABLE VOCAL HABITS VIII0 
IX., UNACCEPTAIJLE OR ECCENTRIC HABITS /)(0 
X. SELF·ABUSIVE BEHAVIOR ~0 
XI. HYPERACTIVE TENDENCIES ;<'10 
X/I. SEXUIHL Y ABERRANT BEHAVIOR )(110 
XI/I. PSYCHOLOGICA L DISTURBANCES )(//10 
XI V. USE OF MEDICA TIONS O)(/V 
99
 
s+eeQS A~BlliillUS e[1Jo~d 
::) X:IITNaddV 
r--­
 II>
 :;: 
, 
~ I 0' " 
I I 
~~. 3NO !lI'dd n'd~S 1I0'"'dHU 311I!d'd0'd OMtV 
..... 
P··fIUV 
1'-'-f-.----(O~I_ 
10 
_. 
----
 --/IIZ.I_
 ZO 
.. 
(O~'- '£0 
._-_.--... 
. -1--
 --1--. (O!L .a 
I---'---
 -._.. _. (O!,_ 
90 
,....--
 _.-"'-
 --. ---_lO~L 
i 
10. 
11I~'- ~~_..---
 _... 
....-I--
 LO 
i (O~L 1---,'--, ....---t-. 
I----I 
1 10 
_._-(O~t --.-----10 
i 1---'-;--------_.. ---._--_. 
i ~ [g I < a 
~ ~ .. 
z ,.. 
l:' ';J ;' 
I: " 
. 0 
~ ..
 ! 
:I :I " 
.. 'C 
I 'l! " Ill· ...
 I ~ If ~ .. a il ~ I o· f .. II 
~ ! I ~ i I 
n' 01 K' a 
~ 
go ;' :r
 ... I g. I ~ II 
~ 
... 
l !i' :!. !!. "
 ...
 i 
!l ~. 
.. c 
~ ~. . ... :I
 1I ... .. 
• 
.. 
it .. 
l! l! e. 
.I 
XI III'" 11/\ .... ... /\1 III II I --
 -.. 
_...._'---
 l_ AlI'dWWnS 3'I~Olld ... -----. .-----~-_.__ ._--_.-.~-_.-
 ,-----. -'.----­
 "5 
---------------------I8'<f 
-------------UO~llD!j!aUApt 
1 I 
i I
 i I
 I I . I 
1----+ I 
I .­-t 
f J' I \." 
":. c 
~ 
I 0 
. 
~ 
J> 
c 
_. 
0 
0 
[
 ~ -
 ~ 
~ 0 
.\1, IIiX 11X IX 
, ,I lilA 11.\ 'A A L ..A~ III l~I .. _.~ _ 
OMI IHVd ~lV:lS Hnl"VH3~ 3f\lIdVOV OWWV 
AHVW\'H1S ]ll~O\ld 
-----.---~-----.----_....----------uOI,,:)IJ'l".PI 
89
 
69
 
+aa1.1S U01+ ~n.:I:+SUI p mo 1+1"P"PV 
([ XI([N8.({dV 
70 
Additional Instruction Sheet 
In addition to following the instructions in the 
scale booklet, the rater should be aware of the following 
scoring considerations: 
First, where certai.n items are not applicable to 
the individual being rated (for example, the item referring 
to money changing abi.lity may be inappropriate where the 
individual has no opportunity to handle money) the rater 
should ignore the item. Second, where items deal with 
situations that are agai.nst regulations, such as using the 
telephone, the rater should complete the rating by deter­
 mining whether or not the indjvidual could. perform the 
task if it were allowed. Third, where items describe mal­
 adaptive behaviors not usualJy seen in very young children, 
such as items referring to agressive behavior, the rater 
should try to record the behavior as accurately as 
possible. 
717 Lindenwood Ave. 
Topeka, Kansas 66606 
October 12, 1977 
Dr. Albert Berkowitz, Executive Director,
 
American Association on Mental Deficiency,
 
5101 Wisconsin Ave. N.W.,
 
Washington, D. C. 20016
 
Dear Dr. Berkowitz:
 
I am a graduate student from Emporia State University, 
Emporia, Kansas. This correspondence is directed for 
permission to reproduce parts of the AAMD Adaptive Behavior 
Scale, 1974 Revision, and to include these reproductions 
in the Appendix section of my masters thesis. 
The subject of my thesis is a reliability study of the 
revised scale. Specifically, score comparisons were made 
between child care workers from the Kansas Neurological 
Institute, Topeka, Kansas who rated a selected retardate 
sample. When this study is fully accepted by the graduate 
department, it will be placed in the university library. 
As I indicated during our phone conversation, I will 
need a letter from you granting me permission to reproduce 
the following parts of the scale: A) Instructions for both 
Part I and Part II.B) Data Summary Sheets for both Part I and 
Part II.C) Profile Summary Sheets for both Part I and Part II. 
Thanking you and hoping to hear from you, I am, 
jkgk.?·tj~ 
West~. Tatman 
e4AMD
 
FOUNDED 1816 
AMERICAN ASSOCIATION ON MENTAL DEFICIENCY 
5101 Wisconsin Avenue, N.W., Washington, D. C 20016 
202/686-5400 
October 26, 1977 
Mr. Westley E. Tatum 
717 Lindenwood Ave. 
Topeka, Kansas 66606 
Dear Mr. Tatum: 
In response to your request for permission to reprint material from the AAMD 
Adaptive Behavior Scale, I have reviewed your request and permission is hereby 
granted. It is understood that full acknowledgment to the AAMD will appear upon 
distribution of the paper. 
I wish you contipued.,
 , 
success. 
J 
I Si~e~e,tY yours, 
/'f/ 1\/
 / ~/LtLClt '/ ~ /,( /?,J l.U /~ 
filbert J. Berkowffz, Ed.D.;
 £xecutive Direct6~ 
AJB:gel 
i