Toward a general model of stress control: An investigation of the relationship between life stressors and substance use strains in an unemployed population. A test of the generalizability of Spector's Job Stress Control Model (P. E. Spector)

The purpose of this research was to explore how Spector's Job Stress Control Model (1998) might be generalized to a population of unemployed adults. The criterion variable studied was the behavioral strain substance use. In general, support was found for Spector's (1998) model to be generalized beyond the stressors associated with the workplace to life stressors associated with being unemployed; and to the use, abuse and dependency of drugs and alcohol. A total of 529 surveys were completed by people from two distinct samples. The first sample, composed of a group of people that currently participate in a panel study for the Florida Mental Health Institute, were sent a mail survey. A second group of people consisted of a convenience sample of individuals who were visiting a local unemployment office. Results supported the moderating effects of Negative Affect on Perceived Stressor's influence on Emotional Response, where people who have high negative affect are more likely to have an exaggerated emotional response to the perceived stressors. There was also some support for the mediating influence of Emotional Reaction on the relationship between Perceived Stressors and the Strain of Substance Use. There was good support for the hypothesis that the relationship between stressors and substance use depends on the length of time a person is unemployed. Moreover, these relationships appear to be curvilinear in nature with cubic regression curves providing the best fit in many associations. Perceived Control was also found to be strongly associated with the time a person is unemployed, and again, the best regression curve was cubic. No conclusions could be reached in an analysis of gender effects because of a lack of statistical power. Race effects showed that people who self-identified as Black were significantly more likely to use Control Strategies including contacts with religious leaders to help cope with their unemployment situation. Further research is encouraged that will address some fundamental issues about Spector's Stress Control Model when it is used in this non-work environment. In particular, it was noted in a post-hoc analysis of the data that there may need to be another mediating variable added between Emotional Response and Strains possibly called: Constraints. This is because people participating in the research with the lowest family incomes had financial burdens such that their ability to obtain certain illegal drugs was apparently curtailed and instead abused alcohol, the opposite of the people with the highest family incomes. Among those people with the lowest family income it would be difficult to purchase certain illegal drugs, even if they had the Emotional Response that motivated them to engage in that behavior. Finally, a suggestion was made to consider the presence of the many curvilinear relationships found as evidence of a complex process that may have contributed to the nearly 100 years of confusion in the literature. The literature has shown that while some research has found that, for example, alcohol consumption goes up during periods of unemployment, other research has shown it to go down and yet others have found no relation. The solution may be in the curvilinear nature of the relationship and further research is suggested to help explain these results.