The attention now being paid to child welfare outcomes raises a number of important questions for researchers and practitioners. Among the most important questions are those that pertain to the issue of change and the statistical and methodological models used to understand how the delivery of services influences the path of change. In this article, I recommend that the study of change be divided into studies of individual-level and aggregate-level dynamics. I present a statistical framework for both types of analysis. I also stress the need to emphasize longitudinal research based on cohorts of children. In particular, I demonstrate how cross-sectional samples distort our impressions of both individual- and aggregate-level dynamics. A failure to recognize the bias in cross-sectional samples will hinder the study of outcomes in general.
No takes yet. Share an insight, caveat, or question.
Fred Wulczyn (1996) studied this question.