A reliability study of measures assessing the impact of deinstitutionalization

Abstract

This study examined the reliability of several scales and indices used to measure outcome variables (independence, integration, productivity, and satisfaction) among people with developmental disabilities. A stratified random sample of 112 people was interviewed twice in a two-week period and included equal numbers of verbal and nonverbal consumers, of parent versus other caregivers, and consumers with diagnosed level of retardation being dichotomized into high and low. In addition, half of the interviews were test-retest and half were interrater. After stratifying on these four variables, the sample was chosen randomly within subgroups from the total database of 3,700 individuals who receive services through the Developmental Disabilities Services Division of the Oklahoma Department of Human Services. Correlation and proportion agreement analyses were performed on the pre- and post-tests and comparisons made on each scale for each stratification to examine variations in reliability. Acceptable correlations and matched agreements of at least 0.70 for all measures were found, with the Adaptive Development Scale having particularly strong correlations. In addition, responses from people with developmental disabilities on items of the Consumer Satisfaction scale were acceptably reliable.