A critique of Herzberg's incident classification system and a suggested revision

Abstract

It is argued that Herzberg's two-factor theory of job satisfaction may not be an artifact of the critical incident method as such, but rather of the system used to classify the incidents. It is shown that Herzberg's classification system confuses two levels of analysis, events (what happened) and agents (who made it happen). A new classification system was developed which classified separately by event and agent. Classification of critical incidents from four employee samples revealed that the same classes of events led to both job satisfaction and dissatisfaction (good and bad days on the job); but that different agents were judged to be responsible for these events. “Motivator” and Task related events were seen as predominantly responsible for both good and bad days; the self was typically given credit for good day events while others were primarily blamed for bad day events. The implications of the results concerning events for Herzberg's theory and method are discussed. The data concerning agents are discussed with reference to the problem of subject defensiveness.