Objectives This study tested and compared three sequential interpretations of Beck's cognitive model of the development of depression (1996). The causal mediational interpretation identifies dysfunctional attitudes as most distal to depressive symptoms, followed by cognitive distortions, the cognitive triad, and negative automatic thoughts, with each construct successively more proximal to depressive symptoms. By contrast, the symptom model reverses the causal chain with negative automatic thoughts as the most proximal consequence and dysfunctional attitudes as the most distal consequence of depression. The bidirectional model merges both interpretations into one model. Previous studies on sequential interpretations of Beck's model have not included cognitive distortions or the cognitive triad and did not test the bidirectional model finding contradictory empirical evidence for the sequential order. Method In the 3‐wave longitudinal study, 308 German university students without clinically significant depressive symptoms (245 female, average age: 23.69 years) completed self‐report questionnaires measuring their dysfunctional attitudes, cognitive distortions, cognitive triad, negative automatic thoughts, and depressive symptoms. Results The bidirectional model with partial mediation fit the data best and cognitive distortions mediated the relationship between dysfunctional attitudes and negative automatic thoughts and vice versa. Conclusions The findings have important consequences for the prevention of depression. Prevention programs may want to focus on cognitive distortions, the only construct in Beck's model that influences every other construct in the model.