The purpose of this study was to evaluate treatment specificity and potential mediators of two online therapies for depressive symptoms. We conducted a randomized controlled trial in which 263 participants were randomized to online cognitive behavioral therapy (CBT), online problem-solving therapy (PST) or a waiting list control group. Both treatments were more effective than the control group in reducing dysfunctional attitudes, worry, negative problem orientation and enhancing feelings of control. No differences between the treatments were found on each of the potential mediators. Furthermore, results suggest that dysfunctional attitudes, worrying, a negative problem orientation and perceived control all played a mediating role in CBT as well as in PST. Our findings suggest that regardless of the theoretical background to the therapy, the psychological processes necessary for symptom reduction seem to be comparable.