Dehumanization is a broad conceptualization in which a human is treated as less than fully human (Haslam, 2006). Dehumanization has been described as the “inverse process of anthropomorphism” because it involves the denial of human traits (Waytz, Epley, & Cacioppo,2010, p. 58). Haslam's (2006) dual model of dehumanization defines two types. In animalistic dehumanization, targets are perceived similarly to animals, with decreased cognitive capacity, rationality, and civility. In mechanistic dehumanization, also called objectification, targets are perceived similarly to objects, with diminished agency, vitality, or warmth. Scholars have determined that these two types are often intertwined (Haslam & Loughnan, 2014), especially in regards to the treatment of women (Gervais, Bernard, Klein, & Allen, 2013).
Similarly, objectification theory posits that sociocultural forces, including media messages, influence people to diminish or deny the personhood of women and instead treat women as things (Fredrickson & Roberts, 1997; Fredrickson et al., 1998). Scholars have integrated objectification theory with models of dehumanization to help explain why women in particular are targeted (Gervais et al., 2013; Moradi,2013). When women are objectified, they are perceived as less competent, less moral, and less human (Heflick & Goldenberg, 2009; Heflick, Goldenberg, Cooper, & Puvia, 2011; Vaes, Paladino, & Puvia, 2011).
Several studies have examined the conditions under which people are more likely to dehumanize women, and found certain triggers or situations make women more likely to be objectified. When women are perceived to deviate from social norms, or behavior deemed appropriate in a particular context, including clothing choice, they are more objectified (Gurung & Chrouser, 2007; Infanger, Bosak, & Sczesny, 2012). The appropriateness of clothing worn also influences perceptions of credibility, likability, dominance, and competence (Gorham et al., 1999; Gurung & Chrouser, 2007; White, 1995). This also applies to digital embodiments, and research has shown that the clothing choice associated with avatars and virtual representations engages sex stereotyping and gender schema (Fox & Bailenson, 2009; Fox, Ralston, Cooper, & Jones, in press). Thus, the model in Figure 1 predicts that those in context appropriate clothing will be perceived to be dressed appropriately, and perceptions of appropriate dress will predict credibility.
Sexism and Objectification of Virtual Others
Synthesizing the predictions of CASA and objectification theory, the model in Figure 1 predicts that virtual women are likely to be objectified based on their appearance, particularly by sexist people. Previous studies have shown that virtual representations of women and femininity can trigger sexist reactions (Behm-Morawitz & Mastro, 2009; Fox & Bailenson, 2009; Lee, 2004, 2008) as well as objectification-related processes (Fox et al., 2013, in press). Given that CASA predicts that people use similar cognitive processes and biases to judge virtual representations and their levels of social potential, the same individual differences that predict person perception and objectification should influence how people process virtual representations and their messages.
Anthropomorphism is the perception of humanness of the other, whereas objectification is a process of dehumanizing the other (Loughnan et al., 2010; Vaes et al., 2011; Waytz et al., 2010). Human images are generally perceived as the most anthropomorphic, followed by animals, with objects generally perceived as the least anthropomorphic (Nowak & Rauh, 2005, 2008). If a perceiver objectifies a virtual woman, she will be dehumanized and seen as less anthropomorphic; in turn, finding the virtual woman less anthropomorphic would diminish perceptions of social potential and credibility. These predictions are illustrated in the model portrayed in Figure 1.
Individual traits and biases such as sexism lead people to interpret the same stimuli in unique ways and reach different conclusions about the credibility of the source and the value of associated messages (Hamilton & Nowak, 2010; Hamilton & Sherman, 1994; Nowak et al.,2009; Nowak & Rauh, 2008). Sexist individuals are overall more likely to objectify women regardless of other factors (Cikara, Eberhardt, & Fiske, 2011; Rudman & Mescher, 2012; Vaes et al., 2011), and women are generally less sexist than men (Swami et al., 2010; Swim, Aikin, Hall, & Hunter, 1995), explaining the predicted path from female sex to sexism. The model predicts that sexism will explain some of the variance in the perception of the image, with more sexist individuals being more likely to objectify and dehumanize the virtual representations of women than less sexist individuals. This means that sexist people will rate the women as less human and less credible, as shown in negative direct paths from sexism to both anthropomorphism and credibility, and this will influence the remaining judgments.
Anthropomorphism increases perceptions of social potential, and more anthropomorphic images are generally perceived to be more realistic (Bailenson et al., 2006; McGloin, Nowak, & Watt, 2014; Nowak et. al., 2009), explaining the predicted path from anthropomorphism to realism. Although there are different dimensions of realism, the one considered here is realism as typical, probable, or likely to exist offline (Busselle & Greenberg, 2000; Hall, 2003; Nowak et. al., 2009). This dimension of realism is also likely to influence perception of appropriate dress, as one would typically expect more realistic avatars and virtual images to be dressed appropriately for the interaction goals. Thus, images perceived to be more realistic are predicted to be rated as more appropriately dressed.