Abstract
We continue our investigation of the relationship between preference judgments and orienting behavior. We reported (VSS ′01) an increasing, cascade-like gaze bias towards choice in preference tasks (cascade effect), when stimuli were either faces, or abstract, unfamiliar shapes. We proposed a model in which orienting behavior (active gaze) interacts with cognitive assessment of stimuli in a positive feedback loop leading to the decision.
If our model is correct, biasing observers' gaze should influence preference. To assess the validity of this claim, we showed human faces, whose baseline attractiveness was matched, on a computer screen, while subjects were instructed to actively follow the display with their eyes. Faces were presented side by side alternatively, so that only one was present on the screen at any given time. We biased the fixation duration by presenting one face for a longer time (900 vs. 300 ms). This sequence was repeated 2, 6 or 12 times before subjects were allowed to respond which face appeared more attractive. The results show a clear bias towards the longer presented face when the number of repetitions was larger than 2 (59.0 % for 6 repetitions, 59.2% for 12 repetitions, both p<0.01), consistent with our model.
To exclude the possibility that the preference bias was due to mere exposure, we performed a control experiment in which the presentation sequence was identical, but both images were presented in the middle of the screen, at the point of fixation. Retinotopic exposure sequence was identical, yet this experiment did not reveal any preference for the longer exposed face. Moreover, to show that the effect was not due to general perceptual fluency, we repeated the original experiment asking subjects to report which face was rounder, instead of more attractive. No bias towards the longer shown face was found.
We conclude that orienting behavior, in the form of active gaze shift, is critical in deciding preference, directly influencing it.