I wanted to flag this terrific paper by Gregory Warrington on the various quantitative measures of partisan asymmetry that have been proposed in recent years. Warrington calculates each metric over more than 1000 historical elections as well as over a range of simulated outcomes. He then examines how the metrics are correlated with one another, which plans they identify as the most skewed, and under what circumstances the metrics disagree. His conclusion, which mirrors my and Eric McGhee’s judgment, is that all of the common measures are reliable in competitive electoral environments. In uncompetitive settings, though, partisan bias and the mean-median difference should not be used, while the efficiency gap and the declination still may be.