Attempts to predict coping, distress, and stress-related pathology wit
h person-related variables have had only limited success. With greater
attention to their controllability focus, and place in the causal seq
uence, it was believed that attributions might yet prove strong predic
tors. Two field studies, one large (N=679) and cross-sectional, the ot
her small (N=60) and longitudinal, tested this premise. Confirming hyp
otheses, the perceived controllability of an attribution proved essent
ial to prediction and attributions for coping failures proved superior
to attributions for stressful events as predictors. Results also reve
aled a causal sequence in which attributions affected distress both di
rectly and by influencing the choice of coping strategy. Implications
for future applications of attribution theory to stress research are d
iscussed.