I don’t need your attention: ostracism can narrow the cone of gaze

Interestingly, it has been found that awareness that one is being ostracized by a computer program does not make the experience any less distressing (Zadro, Williams, & Richardson, 2004). However, it could moderate other outcomes of ostracism, such as its effects on motivation and behavior. For instance, a recent study found that ostracism decreased participants’ prosocial behavior only when they were led to believe they were interacting with other human beings, rather than with a computer (Kothgassner et al., 2017). To ensure our results would not be influenced by participants’ suspicions, we chose to exclude participants who indicated awareness of the deception in Cyberball. This decision was made before collecting the data.

We also analyzed eye contact impression strength, combining data from both response windows (R1 and R2). The impression strengths were set in an ascending order (1 = not looking at me, strong impression, 2 = not looking at me, intermediate impression, 3 = not looking at me, weak impression, 4 = looking at me, weak impression, 5 = looking at me, intermediate impression, 6 = looking at me, strong impression). A Kruskal–Wallis test revealed that there were differences between the groups in the mean eye contact impression strength scores (χ²(2) = 6.36, p = 0.042). Mann–Whitney U tests showed that excluded participants reported weaker eye contact impression strengths than participants in the inclusion group (mean ranks were 17.98 and 27.02, respectively, U = 142.50, p = 0.019), and the control group (mean ranks were 18.95 and 26.87, respectively, U = 164.00, p = 0.043). There were no differences between the inclusion and control groups (mean ranks were 23.14 and 22.87, respectively, U = 250.00, p = 0.946). These results expand on our main finding by showing that ostracized participants (compared to participants in inclusion and control groups) not only considered a narrower range of gaze directions as being directed at them, but also reported weaker impressions of being looked at.

In the small-scale experiment, we randomly allocated participants (N = 42) to be included or excluded in Cyberball (inclusionary status, between subjects factor), ostensibly played with other participants either present in the laboratory, or located elsewhere (experiment setting, between subjects factor). After the manipulation, participants completed basic need, mood and pain questionnaires, as well as the Need to Belong Scale (Leary, Kelly, Cottrell, & Schreindorfer, 2013), Solitude-seeking Scale (Ren et al., 2016), and State Hostility Scale (Anderson, Deuser, & DeNeve, 1995). We did not find evidence that ostracized participants in the two different settings responded differently on these scales. Two-way between-subject ANOVAs found an expected effect of inclusionary status on basic needs, mood, and pain (all ps < 0.001), but no effect of experiment setting (all ps > 0.1), or an interaction between the two factors (all ps > 0.1). On the other scales, which were more crucial for our research question, we found no main effects (all ps > 0.1) or interactions (all ps > 0.4). We chose not to proceed further with this research, as the scales did not appear to be very sensitive to our manipulations, and a very large sample size would likely be needed to detect interactions, and this was outside the scope of the current study.