When evaluating the effect of a program upon a target population, the ideal study design is a randomized experiment; however, such a study is often impractical or impossible to perform. Matching methods replicate the conditions of a randomized experiment by selecting treatment and control groups from observational data that differ only in their participation in the program of interest. This presentation will review matching methods and their application to program evaluation in a post-secondary setting. A case study will be presented in which matching methods are used to evaluate the effectiveness of an advising outreach program at Humber College.

Jeffrey Daniel (Humber College)