The importance of the self-other distinction for understanding the relation between attributions and marital satisfaction is examined in two studies. In Study 1, causal attributions for naturally occurring behavior by the self and spouse were investigated. Study 2 examined both causal and responsibility attributions for hypothetical behaviors. In both studies, the attributions of spouses seeking therapy were investigated in relation to those of happily married persons in the community. The results showed that self-other attribution differences varied as a function of marital distress. Nondistressed spouses showed a positive attribution bias by making more benign attributions for partner behavior as opposed to self-behavior, whereas distressed spouses showed a negative attribution bias by making less benign attributions for partner behavior than for self-behavior. These findings suggest that self-attributions may, in part, determine the impact of attributions for spouse behavior on marital satisfaction. The clinical relevance of the results and their implications for research on actor-observer attribution differences are outlined.