Designing Quantitative Research For Causal Claims

Facilitated by Dr. Paul Garton

Questions in educational research tend to have causal implications. What is the effect of some reading program on reading levels? Do private school vouchers cause public schools to change in quality? Direct cause-and-effect relationships are often a topic of interest, because we want to know how to enact a change or identify the (un)intended results of an existing program. Causal claims, however, require specific research designs and analyses. The gold standard of causal research is the experimental design, in which the researcher has full control over which participants receive treatment, but different methods can be used to approximate experimental designs without having control over treatment assignment. This seminar introduces the basic intuition behind what the scholarly community currently understands as causal designs, clarifies what it means for a study to be unable to make causal claims, and gives an overview of the most common approaches to causal analysis, difference-in-differences and regression discontinuities.