The present study explored how eye contact at different levels of visual awareness influences gaze-induced joint attention. We adopted a spatial-cueing paradigm, in which an averted gaze was used as an uninformative central cue for a joint-attention task; prior to the onset of the averted-gaze cue, either supraliminal (Experiment 1) or subliminal (Experiment 2) eye contact might be presented. The results revealed a larger subsequent gaze-cueing effect following the supraliminal eye contact, compared to a no-contact condition. Nevertheless, the gaze-cueing effect was reduced by the subliminal eye contact, in comparison to the no-contact condition. These findings suggest that the facilitation effect of eye contact on coordinating social attention depends on visual awareness, and that the subliminal eye contact might have a distinct impact on subsequent social attention processes, different from that of supraliminal eye contact. This study highlights the need for further investigation on the role of eye contact in implicit social cognition.

Past research examining implicit self-evaluation often manipulated self-processing as task-irrelevant but presented self-related stimuli supraliminally. Even when tested with more indirect methods, such as the masked priming paradigm, participants' responses may still be subject to conscious interference. Our study primed participants with either their own or someone else's face, and adopted a new paradigm to actualize strict face-suppression to examine participants' subliminal self-evaluation. In addition, we investigated how self-esteem modulates one's implicit self-evaluation and validated the role of awareness in creating the discrepancy on past findings between measures of implicit self-evaluation and explicit self-esteem. Participants' own face or others' faces were subliminally presented with a Continuous Flash Suppression (CFS) paradigm in Experiment 1, but supraliminally presented in Experiment 2, followed by a valence judgment task of personality adjectives. Participants also completed the Rosenberg Self-Esteem Scale in each experiment. Results from Experiment 1 showed a typical bias of self-positivity among participants with higher self-esteem, but only a marginal self-positivity bias and a significant other-positivity bias among those with lower self-esteem. However, self-esteem had no modulating effect in Experiment 2: All participants showed the self-positivity bias. Our results provide direct evidence that self-evaluation manifests in different ways as a function of awareness between individuals with different self-views: People high and low in self-esteem may demonstrate different automatic reactions in the subliminal evaluations of the self and others; but the involvement of consciousness with supraliminally presented stimuli may reduce this dissociation.