JOB CHARACTERISTICS & ALCOHOLISM IN WHITE COLLAR WORKERS

到目前为止，关于工作与生活关系的研究成果

To date, the research findings on the relationship between work and drinking are quite equivocal, in part because of the highly inductive character of much of this research, the almost ad hoc nature of both the outcome (alcohol-related) variables and the job characteristics that are analyzed, the failure to control for nonwork-related factors such as marital stress or personality traits, and insufficient information about whether subjects' job experiences pre-date the development of alcohol problems which they are hypothesized to cause. Furthermore, most prior analyses have been based on blue collar samples. This study will address each of these issues by (1) testing several models that postulate a link between specific job characteristics and alcoholism, (2) systematically analyzing a variety of alcohol outcome measures (including alcohol abuse and dependence, quantity-frequency-variability measures, reasons for drinking, and attitudes toward drinking), and (3) analyzing the temporal relationship between onset of drinking problems and subjects' tenure with their present employer. The data source for the proposed analyses is a survey of 1870 white collar employees at Westinghouse Corporation conducted in 1988. This survey, originally designed to assess the prevalence rates of mood disorders and alcoholism in this population, collected detailed information about alcohol consumption, a variety of job characteristics, marital strain and several other nonwork-related factors that need to be controlled, etc. As such, this is the largest sample of white collar professionals for which comprehensive data on both alcohol use and job characteristics has been ascertained. The spouses of married employees completed a briefer questionnaire which included questions about alcohol-related behavior and attitudes toward alcohol. In addition to using standard multivariate statistical procedures (multiple and logistic regression and structural equation modeling) and survival analysis methods (proportional hazards models), two types of econometric analysis -- limited dependent variables analysis and switching regression models -- will be employed to try to improve the correspondence between specific conceptualizations of the work-alcohol relationship and the statistical models used to investigate this relationship. It is expected that the modeling approach developed for this study will be useful to those investigating similar questions among blue collar workers and that the results will help to inform the design of future studies (cross-sectional or longitudinal) of the relationship between work and alcoholism.