One essential feature common to all social and behavioral phenomena is variability across units of analysis at different levels, such as individual and family characteristics and contextual (such as neighborhood and clinic) characteristics. The development of statistical methods so as to better understand and accommodate such variability has been a major methodological achievement of modern social and behavioral sciences. Individuals differ greatly not only in background attributes but also in how they respond to a treatment, intervention, or stimulation. We call the second type of variability """"""""heterogeneous treatment effects."""""""" The proposed research assembles an interdisciplinary team, encompassing such diverse fields as statistics, sociology, psychiatry, economics, and public health, to develop methodological tools that can be used to better understand and investigate the consequences of both types of heterogeneity, with a special emphasis on heterogeneous treatment effects. Specifically, the proposed research has four aims: (1) It will illustrate, through empirical studies, the pervasiveness of heterogeneous treatment effects in social and behavioral sciences. Examples will be drawn from a variety of fields, using both experimental and observational data. (2) It will demonstrate, through micro-level simulations, that heterogeneity in treatment effects can give rise to composition biases in estimated treatment effects when selecting criteria for receiving treatment change. (3) It will develop a set of diagnostic and analytical tools that will help researchers and practitioners to detect and utilize heterogeneous treatment effects so as to better match interventions to the individual and/or social/clinic setting. Special attention will be paid to heterogeneous treatment effects in social, behavioral, and medical sciences. (4) To demonstrate the usefulness of the tools, it will apply the diagnostic and analytical tools to substantive research in three areas: (a) the effects of a job training program on the productivity of low-skilled workers, (b) the protective effect of a spouse on health among the elderly, and (c) the effectiveness of psychiatric treatments on mental health among veterans. 1 The proposed research will provide new knowledge and practical solutions on how to deal with heterogeneous treatment effects in social and behavioral sciences. This work has general implications for almost all research areas dealing with human subjects. In addition to scientific articles, a book resulting from the project will be produced that will be accessible to a wide variety of social, behavioral and medical scientists.
Showing the most recent 10 out of 15 publications
