Applying Regression Discontinuity Designs to Measure Treatment Effects