Decades of research have shown that the correlation between job satisfaction and job performance is modest in magnitude, yet lay people are thought to believe strongly that satisfied or 'happy' employees are more productive at work. This paper first documents the strength and pervasiveness of belief in several versions of the happy-productive worker hypothesis (Study 1), then proposes and explores potential substantive explanations for these beliefs (Study 2). It is possible that lay people base their beliefs on genuinely stronger relationships that occur at a different level of analysis than usually studied by researchers, and/or that exist between satisfaction-like and performance-like variables other than the constructs typically investigated by scholars. Study 2 provides data relevant to several of these possibilities. The most compelling findings were at the within-person level of analysis. The average within-person correlation between momentary task satisfaction and concurrent perceived task performance was 0.57. Individuals feel more satisfied than usual when they believe they are performing better than usual for them. If lay persons mistakenly generalize from their own within-person experiences of satisfaction-performance covariation to the between-persons level, this relationship may be the basis for the strong lay belief that satisfied workers perform better. Copyright (C) 2003 John Wiley Sons, Ltd.