Proposal to Conduct Annual Workshops on Better Quasi-Experimental Design and Analysis

Project: Research project

Project Details


This proposal is to conduct three 2-week workshops on quasi-experimental methods, one per year over three years. The audience is 30 persons per workshop, mainly composed of faculty from Schools of Education. However, a few very advanced graduate students with substantive interests in educational research and with some capacity to spread the knowledge gained in the workshops will be recruited, as well as some local, state and federal officials employees of contract research firms. The workshop’s significance is much like that of its predecessor carried out at Northwestern University over the last 3 years. It can be expressed as a series of propositions: (i) valid causal research is needed in Education if the field is to become more evidence-based; (ii) randomized experiments are preferred for this task but are not universally applicable and sometimes go awry in their implementation; (iii) many quasi-experimental alternatives exist; but (iv) many – if not most – of them are of dubious validity; and (v) from both theoretical analysis and empirical research comparing their causal estimates to those of experiments on the same topic, some quasi-experiments produce closer approximations to experimental results than others; (vi) knowledge of these quasi-experimental practices deserve to be disseminated more widely; and (vii) in ways that practitioners can actually use. The main difference between the past and proposed workshops lies not in teaching students about the relevant theory and providing them with some memorable and convincing examples but teaching them in hands-on fashion about how to test the assumptions a particular design requires and then how to analyze the outcome data. Attendees of the proposed workshop should go away with (1) all of the knowledge presented in earlier workshops; (2) with knowledge of how to present this knowledge in terms of Rubin’s potential outcome notation; (3) with hands-on experience of how to calculate power for each design and how to analyze the resulting outcome data; and (4) with greater sensitivity to IV methods. Moreover, (5) attendees can spend more of their workshop afternoon and evening time specializing in just one or two quasi-experimental methods if their intellectual interest or work needs make that desirable for them. Compared to their predecessors, the proposed workshops offer a greater proximity to actual research practice, especially in analysis, and entail more time per student and thus a two-week format.
Effective start/end date9/1/148/31/18


  • Institute of Education Sciences (R305B140029)


Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.