Open AccessDetecting Selection Effects in Community Implementations of Family-Based Substance Abuse Prevention Programs
Author(s) -
Laura Hill,
Scott Goates,
Robert Rosenman
Publication year - 2010
Publication title -
american journal of public health
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.284
H-Index - 264
eISSN - 1541-0048
pISSN - 0090-0036
DOI - 10.2105/ajph.2008.154112
Subject(s) - implementation , selection (genetic algorithm) , substance abuse prevention , substance abuse , sample (material) , substance use , computer science , program design language , population , medicine , risk analysis (engineering) , psychology , environmental health , machine learning , psychiatry , software engineering , chemistry , chromatography
To calculate valid estimates of the costs and benefits of substance abuse prevention programs, selection effects must be identified and corrected. A supplemental comparison sample is typically used for this purpose, but in community-based program implementations, such a sample is often not available. We present an evaluation design and analytic approach that can be used in program evaluations of real-world implementations to identify selection effects, which in turn can help inform recruitment strategies, pinpoint possible selection influences on measured program outcomes, and refine estimates of program costs and benefits. We illustrate our approach with data from a multisite implementation of a popular substance abuse prevention program. Our results indicate that the program's participants differed significantly from the population at large.