This paper examines the status of higher‐order need strength as a moderator of the relationship between job characteristics and employees' attitudes towards their jobs. It is argued that previous empirical work has been deficient in two important respects: (a) the measurement of the moderator has not been sufficiently independent of the measurement of the other variables in question, thus allowing response consistency to account for the results; and (b) analytical procedures have been employed which are severely restricted in their scope. The present investigation was designed to avoid both these deficiencies, and shows that higher‐order need strength moderates the relationship between job characteristics and job satisfaction. Comment is also offered on alternative statistical techniques for analysing data in which repeated measures of the moderator variable are used, or where the moderator is not linear in its effe