Consumer Response to Composite Ratings of Nursing Home Quality, Marcelo Perraillon, Tamara Konetzka, Daifeng He, and Rachel Werner, in the American Journal of Health Economics, 2019, right here: https://www.journals.uchicago.edu/doi/abs/10.1162/ajhe_a_00115.
This article illustrates the bugaboo of regression discontinuity methods: elegant theoretically, they can sometimes yield nonsensical results in practice. This happens in Figure 1, which illustrates the trend lines on each side of a set of discontinuities for relatively small bandwidths. The anticipated slopes of these lines are clear from theory: all positive. In practice, they are all over the place. In this paper, description is saying something, if we stop to listen.