ERIC Number: ED475900
Record Type: RIE
Publication Date: 2001-Aug
Are Experiments the Only Option? A Look at Dropout Prevention Programs.
Agodini, Roberto; Dynarski, Mark
This study explores whether propensity score methods produce unbiased estimates of program impacts by comparing experimental and propensity score impacts of dropout prevention programs on four student outcomes (dropping out, absenteeism, educational aspirations, and self-esteem). It also estimates the standard error of propensity score impacts using a bootstrap approach. Data come from the federal evaluation of the School Dropout Demonstration Assistance Program (SDDAP) and the National Educational Longitudinal Study. Results find no consistent evidence that propensity score methods replicate experimental impacts of the SDDAP programs. This finding holds even when the data available for matching are extensive. Findings also suggest that propensity score methods can replicate standard errors of experimental impacts and can be computed using the standard analytic formula. Results suggest that evaluators should carefully consider the process by which programs target individuals and the process by which individuals enter programs. An appendix presents characteristics used to select comparison groups and impacts on other outcomes. (Contains 20 references.) (SM)
Descriptors: Academic Aspiration, Attendance Patterns, Dropout Prevention, Dropout Research, Evaluation Methods, Program Effectiveness, Program Evaluation, Secondary Education, Self Esteem
Mathematica Policy Research, Inc., P.O. Box 2393, Princeton, NJ 08543-2393 . Tel: 609-799-3535; Fax: 609-799-0005; Web site: http://www.mathematica-mpr.com.
Publication Type: Reports - Research
Education Level: N/A
Sponsor: Smith Richardson Foundation, Inc., Greensboro, NC.
Authoring Institution: Mathematica Policy Research, Princeton, NJ.
IES Cited: ED477483