Studies of California’s $20 fast-food minimum wage disagree about whether it reduced employment. Much of the disagreement arises from a substantial number of seemingly innocuous researcher decisions on comparison group construction. This column reports the results of 1,440 plausible synthetic-control specifications. More than 96% of the estimates show job losses, but the range is wide. Estimates that imply no effect sit in the tail of the distribution. By contrast, the estimates that arise from simple comparisons between California and the rest of the US imply an employment decline of around 3% and land near the distribution’s centre. The authors argue that results should be shown across a range of defensible choices.