A comparative study of two measures of adaptive behavior for mentally retarded adults

• Main Author: Mongrain, Susann Louise
• Language: English
• Published: 2010
• Online Access: http://hdl.handle.net/2429/19423

Since 1959, adaptive behavior has been defined by the American Association on Mental Deficiency (AAMD) as one of the two dimensions to be used in diagnostic classification in mental retardation. This use of adaptive behavior has received support in the literature in this field, but it has not been adopted in clinical practice. This may be the result of problems that have been associated with the conceptualization and measurement of adaptive behavior by early scales. In recent years two new measures of adaptive behavior have been published: the AAMD Adaptive Behavior Scale (ABS) and the Vocational Rehabilitation and Research Institute's (VRRI) Adaptive Functioning Index (AFI). The ABS consists of two parts. Part I measures the extent to which an individual is able to cope with the personal and social demands of his/her environment. Part II is designed to measure maladaptive behaviors. The AFI consists of three parts. The Social Education Test (SET) is an individualized test of practical academic and related skills. The Vocational Check List (VCL) is designed to assess basic work habits and skills. The Residential Check List is used for assessing self-help skills related to independent living. Very little information concerning the utility of these scales is available. Therefore, the purposes of this study were to provide estimates of the reliabilities of these scales; to examine their ability to discriminate among individuals at all levels of functioning; to assess their validity by examining their mutual relationship as well as the relationship of each scale to measured intelligence; and finally, to assess the utility of these scales for programming purposes. A sample of fifty-one mentally retarded adults were tested on the Revised Stanford Binet Intelligence Scale (RSB) and on the SET. In addition, each subject was rated by two staff members familiar with his/her daily behavior on the ABS Parts I and II, the VCL, and the RCL. Three months after this data was collected, thirty subjects were retested on the SET. The results of this study indicated that the internal consistencies of the ABS and the AFI scales were sufficiently high and comparable to those of standardized tests such as the RSB. The stability of SET scores also appeared to be adequate. The interrater reliabilities of the ABS Parts I and II, the VCL, and the RCL were assessed using generalizability analyses. These analyses revealed great variation in the extent to which rater pairs agree in their judgments. Some inconsistencies appeared to be the result of a tendency for different raters to assign different scores to the same persons. This bias was attributed to ambiguities in the scales as well as to differences in the opportunities for observation raters have and/or use. Both the ABS and the AFI contain items that were not useful in discriminating among members of this sample. Some items were ambiguous; others were too easy or too difficult. In general, the ABS appeared to be more useful for discriminating among more severely handicapped individuals, whereas, the AFI seemed appropriate for individuals functioning at higher levels, The validity of these scales was assessed using a multitrait-multimethod approach. Evidence was provided for the convergent validity of sixteen of the nineteen traits that were examined. However, support for discriminant validity emerged for only seven of these traits. The practical validity of the five domains of the ABS Part II that were examined was supported. Many aspects of adaptive behavior were found to be significantly related to RSB mental age. The SET showed the most striking relationship to mental age. Finally, the ABS appeared to be more useful for programming in institutional environments, whereas the AFI seemed more useful for more independent settings. The AFI was judged to have certain desirable psychometric characteristics, such as equal numbers of items in the subtests of a given test, that make it more desirable for programming. However, the usefulness of both the ABS and AFI was considered to be limited by their low interrater reliabilities and hence their limited validity. === Education, Faculty of === Educational and Counselling Psychology, and Special Education (ECPS), Department of === Graduate