| Session Title: Regression Discontinuity Design: Lessons Learned From a Real World Application |
| Demonstration Session 474 to be held in International Room on Thursday, November 8, 5:15 PM to 6:00 PM |
| Sponsored by the Quantitative Methods: Theory and Design TIG |
| Presenter(s): |
| Elizabeth Autio, Northwest Regional Educational Laboratory, autioe@nwrel.org |
| Abstract: In the many instances when purely experimental designs are not feasible, evaluators are increasingly encouraged to adopt quasi-experimental methods, such as the regression discontinuity design. While the methodology of regression discontinuity is described in several texts, this presentation connects research to practice by describing its application in the "real world" environment of evaluation. We explore lessons learned when applying the design to student assessment data from a large-scale education evaluation. Issues addressed will include: time and cost investment; the challenges of appropriate model specification, pseudo-effects, and ceiling and floor effects; wrestling with the possible violation of assumptions; and the overarching question of client utility. The presentation will invite a larger audience to share in what we have learned as well as to offer their own insights to the discussion. |