A model of social stressor variables, social integration variables and demographic control variables was tested to assess their impact on alcohol use among the elderly. A secondary analysis of a survey on alcohol use among the elderly in Edmonton, Alberta, Canada, was conducted to test the major hypotheses of the study. Contingency table analysis, using gamma and partial gamma as correlation coefficients, was utilized in the data analysis. The first hypothesis, in regard to the positive relationship of social stressors with alcohol use, was confirmed. The best predictors of alcohol use among the social stressor variables were usual occupation, length of retirement, annual income, and subjective health status. The second hypothesis, that the social integration variables would be negatively related to alcohol use, received only moderate support. The results of the analysis indicated that six of the ten social integration variables were negatively related to alcohol use. Only three of these variables, retirement status, religious participation, and marital status, were statistically significant. Hypothesis three also was not confirmed. The introduction of the social integration variables did not substantially decrease the strength of the relationship between social stressors and alcohol use. Gender and age were also introduced as control variables for …