Several experiments have demonstrated a "camera perspective bias" in evaluations of videotaped confessions: videotapes with the camera focused on the suspect lead to judgments of greater voluntariness than alternative presentation formats. The present research investigated potential mediators of this bias. Using eye tracking to measure visual attention, Experiment 1 replicated the bias and revealed that changes in camera perspective are accompanied by corresponding changes in duration of fixation on the suspect and interrogator. A path analysis indicated that visual attention partially mediated the bias, with at least one additional factor independently contributing to it. A proposed second factor was changes in available visual content that naturally coincide with alterations in camera perspective. Experiment 2 directly manipulated observers' focus and thus more conclusively established visual attention as one mediator of the camera perspective bias. Together the two experiments provide plausible evidence that differences in visual content may also mediate the bias. (Contains 5 figures and 2 tables.)