Review of theory and research on the relationship between job satisfaction and life satisfaction suggests the prediction that causal effects flow in only one direction: that is, that job satisfaction causes life satisfaction but that life satisfaction does not cause job satisfaction. Cross-lag correlation analysis and path analysis of longitudinal survey data so not support this prediction. Indeed, job satisfaction does not predict life satisfaction when the effects of other variables are controlled, nor does life satisfaction predict job satisfaction. In light of this finding, various models of work and non-work are discussed.
It has been eloquently argued that work strongly influences the quality of life even apart from the workplace (e.g., Davis and Chems, 1975;DHEW 1973). Recent reviews of the literature have concluded that influence of work life on non-work life is both pervasive and mutually reinforcing: i.e., good jobs lead to the good life and vice versa (Champoux, 1981 ;Kabanoff, 1980;Near et al., 1980; Staines, 1981). It is less dear, however, how and why work life and non-work life may influence one another; that is, a dynamic model describing how events in one domain (e.g., work) influence events in another (e.g., non-work) has not yet been developed (Cummings, 1982).
EVOLUTION OF A THEORETICAL FRAMEWORK
Attempts in the direction of building such a theoretical framework are still in their infancy. Yet a coherent pattern for a rudimentary schema seems to be developing. Campbell et al. (1976), working mostly independently from Andrews and Withey (1974 and 1976), nevertheless arrived at similar conclusions. Life satisfaction could be viewed as an overall measure of well-being. Satisfaction with various 'domains' of life (e.g., job, family, leisure) directly affected overall satisfaction with life. Living conditions and working conditions also played a role, but this was mediated by various domain satisfactions. Thus, people with boring and low-paying jobs might still be relatively satisfied with