A three-wave panel study of auto plant closings focused on the mental health effects of unemployment on blue-collar workers. This paper explores how the impact of long-term unemployment varies across race and gender. We also examine whether other demographic factors can themselves modify the impacts of race and gender. Dependent variables include two measures of distress and two drinking measures. Results showed that the effect of long-term unemployment on distress and drinking was more severe among less educated workers, and responses of blacks were especially sensitive to level of education. In addition, men showed a greater association of long-term unemployment with depression (and to some extent anxiety) than did women. Marriage affected the responses of men but not of women, and of whites but not of blacks. Explanatory variables--the worker's experiences of financial hardship, other negative life events, and lack of a confidant--largely accounted for male-female differences. We conclude by discussing theoretical implications of these effects and address the limitations of the traditional term "vulnerability" in describing them.