Published evidence, based largely on bivariate methods of analysis, generally suggests that pay, race, occupational prestige, supervisory status, and work autonomy are associated with job satisfaction and that sex is not. Regression analysis of a representative sample of U.S. workers, which partialled out the effects of a number of other variables, affirms that supervisory status makes an independent contribution to job satisfaction and that sex is unrelated, but suggests that the zero-order effects of pay, race, occupational prestige, and work autonomy are spurious.