Evaluating alcoholism treatment programs: considerations and caveats.

There are many pitfalls in the design and implementation of treatment evaluations which can be avoided by careful scrutiny of previous efforts. To this end, the author reviews several key methodological and conceptual issues which affect evaluations of treatment outcome. The issue of sampling is discussed with primary attention given to biases resulting from common sampling errors. The measurement of treatment outcome is explored, underscoring the limitations and benefits of employing drinking behavior indices. Suggestions are made to improve follow-up response rates, and a strategy for analyzing follow-up losses is outlined. Finally, a number of frequently overlooked variables which may relate to treatment outcome are reviewed.