Avoiding minorities: Social invisibility

Three experiments examined how self-consciousness has an impact on the visual exploration of a social field. The main hypothesis was that merely a photograph of people can trigger a dynamic process of social visual interaction such that minority images are avoided when people are in a state of self-reflective consciousness. In all three experiments, pairs of pictures—one with characters of social minorities and one with characters of social majorities—were shown to the participants. By means of eye-tracking technology, the results of Experiment 1 (n=20) confirmed the hypothesis that in the reflective consciousness condition, people look more at the majority than minority characters. The results of Experiment 2 (n= 89) confirmed the hypothesis that reflective consciousness also induces avoiding reciprocal visual interaction with minorities. Finally, by manipulating the visual interaction (direct vs. nondirect) with the photos of minority and majority characters, the results of Experiment 3 (n=56) confirmed the hypothesis that direct visual interaction with minority characters is perceived as being longer and more aversive. The overall conclusion is that selfreflective consciousness leads people to avoid visual interaction with social minorities, consigning them to social invisibility. Copyright © 2012 John Wiley & Sons, Ltd. Research into visual behaviour has a long history in psychology (e.g. Buswell, 1935; Cline, 1967; Gibson, 1963; Yarbus, 1967). However, recent advances in new technologies—such as eye tracking—have opened new possibilities for analyzing complex cognitive processes involved in visual behaviour. During the exploration of a visual scene, the eyes move about three or four times a second via saccadic eye movements. This movement process in gathering visual information is necessary to reorient the fovea through the scene. Eye fixations through a scene serve ongoing perceptual and cognitive activities (Henderson, 2003). Eye tracking provides a millisecond (non-conscious) record of these eye fixations (Duchowski, 2007), allowing us to test complex hypotheses concerning the relationships between cognitive processes and visual behaviour. Many factors that drive the gaze have been studied (Calvo & Lang, 2004; Henderson, 2007; Henderson & Ferreira, 2004; Land & Hayhoe, 2001; Van Gompel, Fischer, Murray, & Hill, 2007). In a review of these studies, Henderson (2003) argued that the two main sources of information used to select specific scene areas for further scrutiny are stimulus-based information generated from the scene (i.e. colour, intensity, contrast, orientation, motion) and top-down, memory-based information generated from visual and cognitive systems (e.g. familiarity, characteristics of the perceiver in interaction with stimulus characteristics, expectations, emotional meaning; Calvo & Lang, 2004; Henderson, 2007; Henderson & Ferreira, 2004; Land & Hayhoe, 2001; Van Gompel et al., 2007). Both stimulus-based information and cognitive knowledge structures used in a top-down manner affect how fixations are deployed in a scene (Parkhurst, Law, & Niebur, 2002). At a second level of analysis, several studies have considered how looking at images with social content works. Indeed, the presence of specific people in the picture influences the exploration of the whole scene (Birmingham, Bischof, & Kingstone, 2009; Keysar, Barr, Balin, & Brauner, 2000; Zwickel & Vo, 2010). For instance, some studies have shown that when there is the presence of a person, the participants’ gaze is directed sooner, more frequently, and for a longer amount of time to the person in the scene (Castelhano, Wieth, & Henderson, 2007). Furthermore, Birmingham, Bischof, and Kingstone (2007) and Birmingham et al. (2009) found that fixations stay predominately on the region of the head. The use of images with social content has become particularly relevant in the study of autism. A differential characteristic of autism is an impaired social interaction and communication, including less frequent eye contact and a worsened ability to differentiate between expressions and perceived emotion. By using eye tracking, some studies have shown that individuals with autism avoid looking at people in a scene and look less in the eyes of the people than non-autistic individuals (Boraston & Blakemore, 2007; Dawson, Meltzoff, Osterling, Rinaldi, & Brown, 1998; Klin, Jones, Schultz, Volkmar, & Cohen, 2002; Riby & Hancock, 2009; Sasson, 2006). Most of these studies have used static images, in which people are just photographed. *Correspondence to: Stefano Passini, Department of Education, University of Bologna, Via Filippo Re 6, Bologna 40126, Italy. E-mail: [email protected] European Journal of Social Psychology, Eur. J. Soc. Psychol. 42, 864–874 (2012) Published online 15 October 2012 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/ejsp.1889 Copyright © 2012 John Wiley & Sons, Ltd. Received 27 October 2011, Accepted 20 April 2012 Interestingly, despite being no more than photographs, it was found that visual performance is a function of the (virtual) social interaction between the participant (autistic vs. non-autistic) and the person shown in the picture. This is not a breakthrough for scholars of virtual reality. For instance, an experiment of Slater, Pertaub, Barker, and Clark (2006) involved people who were confident public speakers and people who were phobic. Half of each group spoke within a virtual environment depicting an empty seminar room and the other half within the same room but populated by a neutrally behaving virtual audience of five people. The people with phobia showed a significant increase in signs of anxiety when speaking to the virtual audience (even if they knew perfectly well that the observers were avatars) compared with the empty room, whereas the confident people did not show any difference in the two conditions. Thus, autistic individuals avoid contact with people’s eyes (even in a simple photograph), and phobics felt anxiety at having to speak to an audience even if they knew it was onlymade up of avatars. These studies suggested the main hypothesis— underlying the three experiments that will be presented in the succeeding texts—that a virtual social interaction (with the character of the photo) is induced when individuals look at pictures of people although they know they are only photographs. In psychological terms, this means that observers behave as if their gaze is not only used to seeing the others but also to communicate a sentiment of the perceiver to the observed people. In particular, we supposed that the way in which participants view the person they see in the image can be considered as an index of the (virtual) social interaction with that person. To our knowledge, no previous study on visual behaviour has considered this virtual social interaction hypothesis with images of people, even if they are only