This dissertation presents four papers that examine methods for improving the design, implementation, and analysis of the regression-discontinuity design (RDD). In RDD, units are assigned to treatment conditions solely on the basis of a cutoff score on a single continuous assignment variable. Units that score on one side of the cutoff are assigned to the treatment while units that score on the other side are assigned to the comparison. Treatment effects are estimated by measuring the displacement of the regression line at the cutoff point.

The first chapter reviews five studies that compare RD results with those from an experimental benchmark that share the same treatment and target population. Based on a qualitative synthesis of five within-study comparisons, I find that while the RDD often replicates experimental results, it does not do so in every case. However, design flaws in the within-study comparison may have confounded study results. The second chapter examines four methodological approaches for estimating treatment effects in a Multivariate Regression Discontinuity Design (MRDD). For a MRDD with two assignment variables, I show that the overall treatment effect at the cutoff frontier can be decomposed into a weighted average of two univariate regression-discontinuity effects and that the weights depend on the scaling of the assignment variables. Using a Monte Carlo simulation study, I find that the four approaches yield unbiased effect estimates when their required assumptions are met, and the contexts and conditions under which these assumptions are most likely to be met. The third and fourth chapters are applications of the RD design and methods. In Chapter 3, I use RDD to evaluate the effectiveness of five state pre-kindergarten programs on children’s receptive vocabulary, math, and print awareness skills, and in Chapter 4, I apply the “centering approach” discussed in the Chapter 2 to investigate school gaming practices under accountability reform. Using data from Pennsylvania, I provide empirical evidence that schools just below the state threshold engage in gaming practices and suggest mechanisms that these schools employ.