Behavioral data were collected for 713 residents of a public institution in 1978 and 1980; the same data were collected for 174 residents of community living arrangements (CLAs) in a neighboring urban area in 1979 and 1981. The research question was whether the two groups displayed different rates or patterns of developmental growth. Matching clients by initial adaptive and maladaptive behavior scores, gender, and age, 104 matched pairs were identified. The institu tional sample showed no change over 2 years, while the CLA sample showed significant developmental progress in reduction of maladaptive behavior. Prob lems with matched comparison designs are discussed. It is concluded that, despite its flaws, this design is the best presently feasible to address the question of client development in the two settings.Studies of deinstitutionalization are designed to answer the question of whether clients are "better off in small, community-based settings after leaving large, segregated institutions. This question is receiving in tense scrutiny as part of a longitudinal study of courtordered deinstitutionalization in Pennsylvania (the Pennhurst Longitudinal Study). The meaning of "better off must be seen to include developmental growth (in creased adaptive behavior and reduction of maladap tive behavior), changes in health, safety, family satisfaction, client satisfaction, amount of appropriate services delivered, and community reaction. The longitudinal study, to which this report is related, ad dresses each of these elements. This report compares the developmental growth of noninstitutionalized clients and comparable institutionalized clients. IIn their review of research on community residen tial alternatives, Heal, Sigelman, and Switzky (1978) concluded that most studies were one-shot, more longitudinal work was needed, and more powerful ex perimental and analytical designs were essential. The empirical information available to compare the quali ty and utility of institutional and community options was perceived to be poor. Heal et al. did, however, note that the most recent studies tended to support the notion of low cost and high effectiveness in community-based, culturally normative residential settings.