The researchers found that adding an untreated comparison to the basic regression discontinuity design can improve its weaknesses compared to a randomized clinical trial. By using pretest observations, they created a comparison function that made the assumptions more transparent and produced more precise causal estimates. The results showed that the pretest-supplemented design produced causal effects similar to an RCT, not just at the cutoff but also away from it. This approach can help improve the accuracy of causal estimates in studies using regression discontinuity design.