Meta-analyses have been conducted within the last twenty years in four areas in the field of perceived social reality. Their results will be summarized here. It has to be noted from the start, though, that any meta-analysis summarizes only the existing literature at the time. The findings reported here should not be taken to reflect current knowledge or thinking in the area, but instead, only the state of knowledge as of the respective dates of publication of the meta-analyses.
Meta-Analysis Of The False Consensus Effect
The earliest meta-analysis in the area of perceptions of social reality was undertaken on the corpus of literature on the false consensus effect. The false consensus effect refers to our tendency to perceive our own behaviors, attitudes, and opinions as relatively more common than those of people who behave differently or have alternative opinions or beliefs. As such, it is a relative phenomenon, not an absolute one. We do not necessarily believe our own opinions or behaviors are in the majority but, rather, we tend to overestimate how common our opinions and behaviors are. Many explanations for this phenomenon have been offered, ranging from selective exposure to people who are similar to ourselves to motivation to believe that our behaviors, choices, and attitudes are normal and good (see, e.g., Marks & Miller 1987).
In the typical false consensus study, participants are asked to provide information about their own beliefs or behavior or the extent to which they exhibit certain characteristics, and they are also asked parallel questions about some other group or the population at large. Mullen et al. (1985) published the first meta-analysis of existing false consensus studies. For inclusion in this meta-analysis, studies had to meet three criteria: (1) they had to report some statistical test of false consensus, (2) the outcome measure of interest had to be free-response, self-generated percentage estimates of the prevalence of a particular characteristic, opinion, or behavior in some population, and (3) the statistical tests of the false consensus effect had to be derived from original data gathered to assess the phenomenon. These criteria yielded 23 studies containing 115 tests. The results indicated statistically significant support for the existence of a false consensus effect that is, on average, moderate in magnitude. Furthermore, they found that the effect was larger when participants were asked to respond to fewer questions about the behavior and beliefs of others, or when participants were first asked to provide reports of the behavior or beliefs of others prior to reporting their own, compared to when the questions were asked in the opposite order.
Robbins and Krueger (2005) published a subsequent meta-analysis of studies of the false consensus effect focused on whether such false consensus or “projection” effects are larger for perceptions of ingroups relative to outgroups. Examining 19 studies yielding 48 hypothesis tests based on over 5,000 participants, Robbins and Krueger report that, indeed, people tend to believe members of their ingroups are more similar to themselves in a variety of ways than are members of outgroups. However, ingroup false consensus effects tended to be larger in perceptions of laboratory-contrived ingroups than real ingroups existing in the participant’s world outside of the laboratory context.
Meta-Analysis Of Cultivation Effects
Perceptions of false consensus are driven at least in part by selective exposure to people who are similar to ourselves. Through our daily interactions with like others, we come to see our world as populated by people who share many of our own characteristics, attitudes, and values. The advent of television, however, gave people a new lens through which to view the world. Of course, people differ in how much time they spend watching television. According to cultivation theory, cumulative long-term exposure to television cultivates a worldview that is consistent with television portrayals. In other words, light and heavy television viewers will come, over time, to perceive the world in a very different way. Namely, the perceptions of heavier viewers will tend to become more consistent with the “reality” of the television world – as one that is relatively violent, hostile, and unfriendly – more so than will those of lighter viewers.
Shanahan and Morgan (1999) published the first meta-analysis of cultivation studies, examining the relationship between television-viewing frequency and perceptions of the world on such dimensions as the prevalence of violence, likelihood of being a victim of crime, and so forth. In 58 independent tests of the cultivation hypothesis, they report an average effect of around 0.10 in correlation terms, an effect that they admit is small but statistically significant, thus supporting cultivation theory predictions. In general, heavy television viewers report perceptions that are more consistent with the television world than light viewers. However, Shanahan and Morgan also found evidence that the effect sizes were heterogeneous across studies. They found no evidence that the magnitude of the cultivation effect varies across “perceptual domains” – the effect size was around 0.10 regardless of whether participants were asked about crime, sex roles, the “meanness” of the world, or a variety of other domains. They also report no evidence that the magnitude of the effect differs between men and women. However, they did find weak evidence that political liberals seem more susceptible to cultivation than conservatives, and white participants may show greater cultivation effects than other ethnic groups. On the whole, given the number of moderators they looked for and the relatively small differences between studies varying on potential moderator variables examined, their conclusion was that important moderators of the cultivation effect remain “at large” (Shanahan & Morgan 1999, 132).
Meta-Analysis Of The Spiral Of Silence
The remaining two meta-analyses described here are less concerned with studies on perceptions of reality but instead on the potential effects of perceptions of reality. According to spiral of silence theory (Noelle-Neumann 1974; 1993), our perceptions of the beliefs of others influence our willingness to express our own beliefs publicly. When we perceive our own opinion to be popular or gaining in strength with the public, we are more likely to express that opinion publicly than when we think our opinion is in the minority or becoming less popular. Because our own behavior serves as input into others’ perceptions of reality, this theory describes one mechanism by which our own perceptions of reality can affect the perceptions of others’ reality.
Glynn et al. (1997) conducted a meta-analysis of research on the spiral of silence by focusing on studies that met four selection criteria: (1) the study had to assess participants’ perceptions of opinion distribution, and compare that perception to the participants’ own opinion, (2) the investigator could not have experimentally manipulated perceptions of the climate of opinion, (3) the study had to measure perceptions of the climate of opinion rather than which opinion in the distribution was in reality more or less common, (4) the study had to include a measure of how willing the participant was to express the opinion publicly. A total of 17 studies met these criteria, some of which were published, some unpublished, allowing Glynn et al. to compute 123 separate effect sizes.
This meta-analysis revealed a weak but positive and statistically significant relationship between perceived support for our opinion and willingness to speak our opinion publicly. Tests of heterogeneity revealed evidence of significant variation in effect size across studies. Glynn et al. considered several potential moderators of effect size, including whether the study was published, whether respondents were asked about willingness to express an opinion or just about engaging in a conversation on the topic, whether the audience for the opinion expression was presumed to agree or disagree with the respondent, whether the audience for the opinion expression was a stranger to the respondent or a member of the respondent’s community, and whether the opinion expression was directed to a media representative (e.g., a newspaper reporter) or someone else. However, none of these potential moderators of the effect size were statistically significant, leaving unanswered the question of what accounts for between-study variation in the size of the relationship between perceptions of the opinion distribution and willingness to express one’s opinion.
Meta-Analysis Of The Third-Person Effect
People often fail to recognize or are unwilling to admit that the messages they receive through the mass media can affect their own thoughts, feelings, and behavior in negative ways. Yet there is reason to believe that, by contrast, people are quite aware of and comfortable acknowledging the detrimental effects that mass media can have on others or on society as a whole. This tendency to perceive mass media as having a bigger effect on others than on oneself has been dubbed the “third-person effect” (Davison 1983). In the typical third-person effect study, participants are asked to indicate how probable it is that others (e.g., friends, members of the participant’s community, or people in general) will be affected by certain mediated messages (e.g., pornography, advertising, etc.). Participants also indicate how likely they are themselves to be affected by such messages. Like the false consensus effect, explanations for the third-person effect are numerous, such as biased attribution or self-esteem enhancement or maintenance.
Paul et al. (2000) conducted a meta-analysis to determine whether a third-person effect exists and what determines its size, restricting analysis to studies that (1) explicitly included a measure of perceived effects of certain media messages on others and the self, and (2) reported results in a format that could be converted to a common metric. This selection process resulted in 32 studies representing 121 separate effect sizes. Paul et al. found support for the third-person effect hypothesis in the corpus of literature, with an average effect size that was moderate in strength. People do indeed perceive media content as having a greater effect on others than on themselves. However, the size of the effect was heterogeneous across studies, suggesting the presence of one or more moderators.
An examination of sampling method, respondent (student or not), and message content (e.g., pornography, violence, political ads, etc.) revealed them to be significant moderators. Third-person effects were larger when sample collection was nonrandom rather than random. But this difference may have been the result of the fact that effects based on nonrandom samples are typically calculated from college student participants, whereas random samples usually include members of the general population. Indeed, third-person effects were generally larger in studies that used college students than in studies that relied on samples from the general population. Importantly, the meta-analysis revealed no statistically significant evidence that the third-person effect is larger when participants are asked about the effects of socially desirable media messages (e.g., wearing seatbelts, behaving nicely to others) compared to undesirable messages (e.g., pornography, television violence, political scandals), a distinction that has itself motivated substantial third-person effects research.
The Paul et al. meta-analysis tells the research community that there is a systematic tendency for people to overestimate the effects of media on others compared to themselves. But this meta-analysis does not provide information about the potential effects of such perceptions. The third-person effect is considered important because, presumably, differential perceptions of the effect of media themselves have certain effects. For instance, studies suggest that such perceptions are related to a willingness to censor objectionable media messages, presumably out of a desire to protect a naïve public from negative effects. To date, no meta-analysis of the literature on the correlates of third person perceptions has yet been conducted, and one is certainly warranted, given the large number of studies on the topic that currently exist.
- Davison, W. P. (1983). The third-person effect in communication. Public Opinion Quarterly, 47, 1–15. Glass, G. V., McGraw, B., & Smith, M. L. (1981). Meta-analysis in social research. Beverly Hills, CA: Sage.
- Glynn, C. J., Hayes, A. F., & Shanahan, J. (1997). Perceived support for one’s opinions and willingness to speak out: A meta-analysis of survey studies on the “spiral of silence.” Public Opinion Quarterly, 61, 452–463.
- Griffin, D., & Ross, L. (1991). Subjective construal, social inference, and human misunderstanding. In M. P. Zanna (ed.), Advances in experimental social psychology, vol. 24. New York: Academic Press, pp. 319–360.
- Hunter, J. E., Schmidt, F. L., & Jackson, G. B. (1982). Meta-analysis: Cumulating research findings across studies. Beverly Hills, CA: Sage.
- Krueger, J. (1998). On the perception of social consensus. In M. P. Zanna (ed.), Advances in experimental social psychology, vol. 30. New York: Academic Press, pp. 163–240.
- Marks, G., & Miller, N. (1987). Ten years of research on the false consensus effect: An empirical and theoretical review. Psychological Bulletin, 102, 72–90.
- Mullen, B., Atkins, J. L., Champion, D. S., Edwards, C., Hardy, D., Story, J. E., & Vanderlok, M. (1985). The false consensus effect: A meta-analysis of 115 hypothesis tests. Journal of Experimental Social Psychology, 21, 262–283.
- Noelle-Neumann, E. (1974). The spiral of silence: A theory of public opinion. Journal of Communication, 24, 43–51.
- Noelle-Neumann, E. (1993). The spiral of silence: Public opinion, our social skin. Chicago: University of Chicago Press.
- Paul, B., Salwen, M. B., & Dupagne, M. (2000). The third-person effect: A meta-analysis of the perceptual hypothesis. Mass Communication and Society, 3, 57–85.
- Robbins, J. M., & Krueger, J. I. (2005). Social projection to ingroups and outgroups: A review and meta-analysis. Personality and Social Psychology Review, 9, 32–47.
- Rosenthal, R. (1991). Meta-analytic procedures for social research. Newbury Park, CA: Sage.
- Shanahan, J., & Morgan, M. (1999). Television and its viewers: Cultivation theory and research. Cambridge: Cambridge University Press.