This article examines the hypothesis that cognitive therapy (CT) is more effective than interpersonal therapy (IPT) for treatment of depressed patients with an elevated level of avoidant personality, whereas the reverse holds for depressed patients with elevated level of obsessive personality. This hypothesis was derived in part from the preliminary results of previous unpublished pilot work, which examined the course of dynamic and cognitive therapies for avoidant and obsessive-compulsive personality disorders. With the "completer" data set available from the Treatment for Depression Collaborative Research Program (I. Elkin et al., 1989), the expected significant interactions between treatment (CT vs. IPT) and avoidance and between treatment and obsessiveness were found. A significant interaction was also found between marital status and treatment, indicating that married patients did better after CT, whereas single and noncohabiting patients improved more after IPT. Similar patterns of results were found using the Hamilton Rating Scale for Depression and the Beck Depression Inventory. A matching factor formula of patients to CT vs. IPT is presented.