Methods for Estimating MRDD: Comparative Study of 3 Estimation Methods

Angebote / Angebote:

Regression discontinuity design (RDD) is originally a useful tool for evaluating a program or policy when a single variable is used to determine program eligibility. The RDD concept has been generalized to evaluate programs when eligibility is based on multiple variables. Multivariate regression discontinuity design (MRDD) may provide estimates of multiple parameters of interest (corresponding to the multiple possible treatment contrasts and for different subpopulations) so that the analyst is faced with a wider range of strategies for estimating treatment effects from a multiple rating score regression discontinuity. The choice among these different strategies has important implications for precision, bias, and generalizability. In this study, we compare the relative performance of three MRDD estimation approaches (frontier, centering and univariate) in Monte Carlo simulation study under different scenarios (different dimensional space of the assignment variables, presence of white noise and presence of correlation between score variables). Result attested that given correct model specification with two assignment variables.