Abstract Friedlander's (1965) results as well as the conclusion that his data represented support for the two-factor theory of job satisfaction ( Herzberg, 1959; 1966) were critically examined. 442 male and female white-collar workers employed by the same company and living in a company town in Canada were asked to indicate their satisfaction and judge the importance of 21 environmental and job characteristics. Satisfaction with 16 of the characteristics was measured by 5-point Likert scales. Cumulative point adjective check lists were used to measure satisfaction with the remaining five characteristics. The importance of all characteristics was assessed on a 5-point Likert scale. Results indicated that for all environmental or job characteristics a V-shaped function between satisfaction and importance was obtained providing that both satisfaction and importance were assessed by Likert scales. No V-shaped relationship was found when satisfaction was measured by cumulative point adjective check lists. It was concluded that Friedlander's results depend on the measurement scale used and thus are basically artifactual. His results therefore cannot be interpreted as giving support for the two-factor theory of job satisfaction.