Using state data sets and meta-analysis of low-powered studies to evaluate a school-based dropout prevention program for students with disabilities |
| |
Institution: | Westat, 1600 Research Boulevard Rockville, MD, 20850, USA |
| |
Abstract: | This study explores the use of state data sets and meta-analysis of low-powered studies to evaluate a school-based dropout prevention program for students with disabilities. The program was implemented in several states. A randomized controlled trial was infeasible because schools were not chosen at random; furthermore, pretest data were minimal. The use of extant state data allowed these obstacles to be overcome by providing valid pre- and post-intervention outcomes as well as a large selection of schools and variables to create reasonable matches for the treatment schools. Results from four states were synthesized meta-analytically to evaluate whether the program had a significant impact on any of seven proximal and distal outcome variables. No such impacts were demonstrated. More importantly, this paper demonstrates and explains the methodological steps and choices involved in a quasi-experimental evaluation approach that may be applied to cases for which large amounts of extant data are available. |
| |
Keywords: | Meta-analysis Dropout prevention Impact evaluation Difference-in-difference Low-power studies |
本文献已被 ScienceDirect 等数据库收录! |