Stereotyping, Homonegativity And Exposure To Gay-Related Media

The current study investigated two research questions and one hypothesis. Firstly, it investigated the relationship between stereotyping, homonegativity and exposure to gay-related media. Secondly, the study explored the moderating effect of four basic human values (clusters) onto the correlation between exposure to gay-related media and homonegativity, as well as the correlation between exposure to gay-related media and stereotyping of male homosexuals. The hypothesis predicted that exposure to gay-related media will negatively predict homonegativity even after controlling for all the four values, demographic variables and overall exposure to media, based on the findings of Bond and Compton (2015). Demographic variables were also included to control for them in the hierarchical regressions, moderations, and mediation, as done by the most recent studies in this field (e.g. Calzo & Ward, 2009; Bond& Compton, 2015; Sink & Mastro, 2017).

The predictive power of homonegativity and stereotyping

The results suggested that exposure to gay-related media is negatively related to homosexual prejudice even after controlling for the four values, demographic variables, and overall media exposure, supporting the H1. This is in line with Bond & Compton’s (2015) research that found the same results, only with motivation to seek out gay-inclusive TV as a moderating variable, which is thought to be a consequence of personal values (Schwartz, 1992). This thus again weakens arguments for reverse causation between the predictor and the outcome variables; it is less plausible that this association was found only because audiences who already strongly supporting gay equality seek out gay-inclusive media outlets. While the inclusion of values in the regression model also does not establish causality, it further solidifies the arguments that exposure to male homosexuals in media increases audiences’ acceptance of homosexuals, which is supported by wider array of most recent research (Sink & Mastro, 2017; Bond & Compton, 2015, Calzo & Ward, 2009).

Moreover, stereotyping was not a significant predictor of exposure to gay-related media, providing direct evidence in support of the claims of the “acceptance” strand, and suggests that content analyses used by the second “stereotyping” strand may not be the best predictors of media effects on people’s attitudes. However, personal values together were significant predictors of stereotyping, unlike gay-media exposure, which suggests that the tendency to stereotype male homosexuals can be explained by personal variances rather than any media cultivation effects. The current finding is one of the first pieces of evidence against claims about gay media exposure leading to greater homosexual stereotyping, however more investigation of the individual factors explaining the variance in homosexual stereotyping is needed in the future research. Moreover, the amount of gay male relations significantly predicted a decrease in stereotyping. This supports the “ancestor” of PCH, the Intergroup Contact Theory (Allport, 1954), which suggests that intergroup contact is one of the best reduce prejudice between majority and minority group members. Current findings then suggest that this theory can be extrapolated onto stereotyping reduction phenomenon, in addition to the empirically supported prejudice reduction phenomenon. This suggests that stereotyping and homosexual prejudice are somewhat related, even if not exactly the same (Armour, 1995). This is further supported by the following findings of the way homonegativity mediates the stereotyping-gay media exposure correlation.

Relating Stereotyping and Homonegativity

In the investigation of the first research question, results suggest that homonegativity significantly mediates the relationship between exposure to gay-related media and stereotyping. Specifically, exposure to gay media significantly decreases homonegativity, which in turn significantly decreases stereotyping. This supports the model, which suggests due to the conscious nature of prejudice (i.e. homonegativity here), if exposure to gay media decreases homonegativity, people are then more likely to suppress and overcome the unconsciously developed stereotypes about homosexuals, by realizing its overgeneralized nature (Armour, 1995). Ergo, these findings suggest homonegativity and stereotyping co-exist in a mutually inclusive relationship, as one decreases, the other seems to decrease too.

Even though the content analysis was not done, this mediation suggests that the media portray homosexuals in a diverse manner, which allows for not only decrease in prejudice but also in the reduction of the narrow stereotypical characterization of homosexuals (Raley & Lucas, 2006; Bond, 2014). This is also congruent with Sink and Mastro’s (2017) argument about the stereotyping paradox, where many scholars (e.g. Rothmann, 2013; Cartei & Reby, 2012; Levitan & Lloyd 2009) argue that although current media supports homosexual stereotyping, however, there is a lack of recent empirical evidence to support this claimed effect on people's attitudes, as the scholars have largely made these claims based on content analyses. This finding contradicts the McLaughlin’s et al. (2017) conclusions about concurrent existence of stereotyping and homonegativity. However, their finding was only marginally significant, which could have been due to its sample or situational biases of their study, which could explain why the current replication has not found any significance at all.

This mediation effect also supports the cultivation theory, which suggests that the more one is exposed to gay-related media, the more he/she is likely to perceive the real world in the way portrayed by the media messages (Morgan & Shanahan, 2010; Hefner et al., 2015). Framing then suggests that the positive frame of the gay-related media message further enhances acceptance of homosexuals (Pan & Kosicki, 1993). The mediating finding also supports the PCH (Schiappa et al., 2005), as the vicarious contact with diverse homosexual characters allows for normalization of the minority in people’s lives (Pan & Kosicki, 1993; Schiappa et al., 2005; Morgan & Shanahan, 2010), especially given that heterosexual audiences can now even see heterosexuals interacting with homosexuals in a positive way (Bond & Compton, 2015).

Values and their Moderating effect

Exploring the 2nd research question, none of the four values moderated the correlation between homonegativity and exposure to gay-related media or the correlation between stereotyping and exposure to gay-related media, while all the direct correlations between exposure to gay media and homonegativity remained significant. This suggests that exposure to gay media overrides individual differences in value systems, and decreases homonegativity. This finding thus favors the social psychology research, which suggests values are more malleable than previously thought, and can be significantly altered due to external social changes in culture or worldviews, all of which are at least partly shaped by the media (Rohan, 2000; Pakizeh, 2005; Schwartz, 2013). This also supports the cultivation theory suggesting exposure to media can foster or manipulate pre-existing beliefs, attitudes and values, thus suggesting future longitudinal experimental research could explore whether prolonged exposure to gay-related media can foster more tolerant and open-minded values. This, in turn, may cause or contribute to greater acceptance of homosexuals and reduce stereotyping (e.g. Greenberg et al., 1992; Gailliot et al., 2008; Maio et al., 2009; Smeekes, Verkyten and Poppe 2012).

Moreover, relating back to the concept of confirmation bias in the Internet age based on individual value systems (e.g. Bennett & Iyengar, 2008), the lack of such findings supports the scholars’ (e.g. Gentzkow & Shapiro, 2010; Holbert, Garrett, & Gleason, 2010) skeptical of the “new era of minimal effects” argument. These argue that since the development of online media, people are less likely to only selectively focus on the features of the media source (program, news article etc.) that are congruent with their attitudes and value systems. Conversely, people actually often allow exposure to a variety of even contradictory channels with information opposing their beliefs, which they can utilize in future decision-making (Donsbach, 2009; Knobloch-Westerwick & Kleinman, 2012). Therefore, the diversity of information available from the extensive media sources in the Internet age may be one of the reasons why personal values are less pertinent to decision making about minority groups (Maio, 2009; Schwartz et al., 2010).

Extending the Mediated Contact Field of Research

This study was unique and valuable to the field of research as it explored further whether and how people’s individual differences in personality, and thus their motivational antecedents (values), could account for the variance within and between the 2 correlations. Additionally, the simultaneous measurement of stereotyping and homonegativity further helped in the investigation of the two divergent strands in this field (e.g. Bassett et al., 2003; Newman, 2013). Moreover, it even highlighted a specific path through which stereotyping is lowered in response to gay-related media response.

Additionally, while specific genres of the media separately were explored to gain insight about their individual effects (Calzo & Ward; Gonta et al., 2017), none of the recent studies examined the overall consequence of engaging with gay-related media, which may be limiting as most young people engage with several outlets of the media every day (Bond, 2014). It is especially important to expand the measurement beyond television in this increasingly mediatized society, where new media outlets are launched frequently, and television is being less and less dominant and trusted (Prior, 2007; Stroud, 2011; McCombs, Holbert, Kiousis, & Wanta, 2011; Pew Research, 2013; Bond, 2014).

Furthermore, this study developed the aggregate measure with the specific aim not to ask about any specific programs as previously done in most of the research (McLaughlin et al., 2017; Sink & Masto, 2017; Bond & Compton, 2015). It is arguably too Western-centrically biased to assume that participants from the US may not be exposed to other non US media, given that a myriad of Asian or European shows and films and many other media outlets are available to view on the Internet (Raykov & Tobin, 2007; Moss, 2012; Kajinic, 2016; Davidson, 2018). Ergo, asking generally rather than specifically may be more suitable for the current high international media environment (Mihelj, 2011).

Limitations

Nonetheless, the limitations of the current research need to be addressed. Possibly the most prominent limitation is the use of convenience sampling, restraining the generalizability of the findings onto any wider population. Besides, given that majority of participants were female also suggests the findings may be more skewed towards more tolerant, as females are known to be inherently more tolerant towards male homosexuals (Herek, 2007; Bond & Compton, 2015; Ayoub & Garretson, 2017). Additionally, because of the study’s cross-sectional design, the results are based on a limited snapshot of youth’s gay-related media habits and attitudes towards homosexuality. Attitudes are known to be often inconsistent due to situational and social pressures (Davidson & Jaccard 1979), especially due to the uncontrolled web-based design of the survey. To gain a more accurate measure of people’s attitudes, future research should utilize a longitudinal design, monitoring even respondent’s exposure to media more closely, as the current measure of asking participants to remember and estimate their consumption can be inaccurate (Hoffman & Young, 2014).

Moreover, explicit attitude measures such as Likert scales suffer from high social desirability bias, when participants report attitudes, which they deem as wanted by the current society (Poldrack, 2006; Amodio, 2009; Ito & Bartholow, 2009; Deniz & Citak, 2010; Amodio, 2014), even when explicitly told their answers are anonymous. This means that their explicit attitudes are often unreliable as they more or less shaped such social norms, tangible rewards and the beliefs of individuals about themselves (McClelland, 1985; Steffens, 2004; Breen & Karpinski, 2013). This calls for an advancement in the current field of media effects on homosexual attitudes to investigate whether the well-praised decrease in homosexual prejudice associated with exposure to gay media is still prominent even in the implicit attitudes, seen as close to one's core attitude compared to explicit attitudes (Greenwald, McGhee & Schwartz 1998).

Furthermore, the lack of moderation by values may be caused by the value measure, as the current study had to use a shortened version of the scale given the limited space, and need not to cognitively overload the participants. This meant each of 2 of the 4 examined values were measured using only a single item, which is often criticized as an inadequate amount, due to their low reliability, sometimes considered as a fatal error, and requests at least three-item scales (Wanous, Reichers & Hudy, 1997; Gardner et al., 1998; Loo, 2002). Ergo, a replication of the study could be made with more detailed Schwartz’s Value scale (1992), to be able to make more precise conclusions about the moderating role of personal values in this context.

Therefore, given these limitations, definite conclusions should be made from neither the non-significant nor the significant findings shown above. Future research should use a more longitudinal experimental research from which causal inferences can be made, and due to the longitudinal nature, provide more ecologically valid results. Along with larger, more representative samples and more extensive scales, this approach should give more definite answers whether the findings from this research are reliable and valid even in the overall population aged 18-25 years old.