This research explores the role of perspective taking in self-serving biases. Assisted by a confederate, 80 subjects performed an impression-formation task and were given either success or failure bogus feedback. One week later, half of the subjects watched their performance on videotape and provided causal attributions ('observers'). The other half simply gave causal attributions (actors). Thus, the experiment employed a modified version of the actor/observer paradigm with one group of subjects taking the perspective of observers (observers) and one group of subjects keeping their original perspective (actors). The aim of this study was to test whether the change of perspective would increase dispositional causal attributions both in success and failure conditions. Results showed that subjects gave greater causal weight to internal factors (ability, effort) and less causal weight to external factors (task characteristics, collaboration with the partner) in the success than in the failure condition. Moreover, in a direct comparison task, subjects attributed a greater percentage of responsibility to themselves than to their partner in the success than in the failure condition. However, the type of perspective produced no significant effects, but showed an attenuation of self-serving biases for observers as compared to actors. A motivational explanation of the results is proposed.