Restaurant workers did fine (under both this NBER study and a prior Berkley one). It was a superset of low-wage employees including those restaurant workers that did poorly overall. This is one of the more interesting results since minimum wage studies in the past have favored studying restaurant workers (for reasons I don't know).