A persistent and sizeable literature argues that the reported job dissatisfaction of union members is spurious. It reflects either the sorting of workers across union status or the sorting of union recognition across jobs. We cast doubt on this argument presenting the first estimates that use panel data to hold constant both worker and job match fixed effects. The estimates demonstrate that covered union members report greater dissatisfaction even when accounting for sorting in both dimensions. Moreover, covered union members are less likely to quit holding job satisfaction constant and their quit behaviour is far less responsive to job satisfaction. The paradox of the discontented union member remains intact.