William P Eveland Jr. The Persuasion Handbook: Developments in Theory and Practice. Editor: James Price Dillard & Michael Pfau. Sage Publications. 2002.

Often the role of the mass media in Persuasion is perceived to be limited to the impact of overtly persuasive communications such as newspaper editorials, television advertisements, and public service announcements (PSAs). However, other media content may have a more subtle, but more Pervasive, impact on the beliefs of the public, regardless of whether or not this was the intention of the message producer. That is, public beliefs are often shaped by subtle but repetitive messages contained in news and entertainment media content that are not overtly Persuasive. These beliefs may eventually translate into opinions and even socially relevant behaviors. It is this typically unintentional Persuasive impact of news and entertainment media on public beliefs—specifically Perceptions of social reality—that is the focus of this chapter.

Introduction to the Study of Social Reality Perceptions

A common focus of social science research is the public’s Perceptions of their neighbors, community, state, nation, or world. The public’s Perceptions of the climate of public opinion on policy issues and elections and the public’s beliefs about the rate of crime, social norms, stereotypes, the impact of the mass media on public attitudes and behaviors, and many other social reality Perceptions have been studied in fields such as public opinion, sociology, psychology, and communication.

Social reality Perceptions are best defined as “individuals’ conceptions of the world” (Hawkins & Pingree, 1982, p. 224). Because there are so many Perceptions that individuals have about the world, so many potential causes of these Perceptions, and so many levels of analysis at which to examine them, research on social reality Perceptions has spanned a number of related fields (Shrum & O’Guinn, 1993). Often this research has taken complementary directions despite a relative lack of cross-citations. For example, one common finding of the research on social reality Perceptions is simply put as follows: They are often wrong—very wrong—and this can have important implications for both individuals and society as a whole. Unfortunately, social reality Perception research has tended to be somewhat compartmentalized because different fields both use different terminology to describe public misperceptions of social reality and offer different explanations for these misperceptions.

Generally speaking, the notion that the public misperceives the opinions and behaviors of others may be labeled “pluralistic ignorance,” although some scholars using this term define it more narrowly as specific types of misperceptions. Research on pluralistic ignorance as defined here has been conducted in psychology, sociology, and public opinion under the rubric of the false consensus effect (Ross, Greene, & House, 1977), the false idiosyncrasy effect (Sherman, Presson, & Chassin, 1984), social projection or the “looking glass Perception” (Fields & Schuman, 1976), and disowning projection (Fields & Schuman, 1976).

Research conducted in communication has also examined Perceptions of media content and effects, including the third-Person effect (Davison, 1983), the hostile media phenomenon (Perloff, 1989; Vallone, Ross, & Lepper, 1985), and the Persuasive press inference (Gunther, 1998). Noelle-Neumann’s (1993) spiral of silence model is a broader and more well-defined theory that incorporates research on Perceptions of public opinion as well as their causes and implications. Another more fully specified theory originating in the field of communication, cultivation theory (Gerbner, Gross, Morgan, & Signorielli, 1994), is concerned with misperceptions of social reality in the form of beliefs about the characteristics of individuals and society. These communication-related theories also deal with pluralistic ignorance in one way or another, but for the sake of clarity I do not use the term pluralistic ignorance to describe theories in communication.

These theories—or at least hypotheses— about Perceptions of social reality generally focus on misperceptions of different facts such as public opinion, media effects, and social indicators. Persuasion research can benefit from better understanding how Perceptions of social reality are shaped by a number of factors, including the use of entertainment and news media, and from a review of the potential influence of social reality Perceptions on attitudes and behaviors. The organization of this chapter is as follows: (a) a discussion of the concept of pluralistic ignorance, which may be defined as misperceptions of social reality, (b) models of social reality Perceptions in communication, (c) integration of Perspectives on misperceptions of social reality, (d) integration with Persuasion theories, and (e) applications to communication campaigns.

Pluralistic Ignorance: Contexts and Causes

Before discussing theories specifically focusing on the role of mass media in the development of Perceptions of social reality, it is useful to review evidence pertaining to the general accuracy of public Perceptions of social reality. While theories focusing on mass media provide evidence of inaccurate Perceptions of social reality, the greatest number of studies demonstrating inaccuracy come from the fields of public opinion, psychology, and sociology.

Research on Perceptions of public opinion has consistently found that people are inaccurate in their judgments of the climate of opinion (for a rare opposing view, see Nisbett & Kunda, 1985). This inaccuracy has been demonstrated at both the aggregate and individual levels. Among other labels, researchers have termed this general inaccuracy about the distribution of opinions pluralistic ignorance.

Pluralistic ignorance is a term that was introduced by Floyd Allport (e.g., Katz & Allport, 1931) to refer to a situation in which members of the majority believe that they are actually in the minority. Research on this topic was continued by one of Allport’s students, Richard Schank, in a study of a small community in New York (Schank, 1932). Since the early days of pluralistic ignorance research, the original definition has been contested, becoming for some authors more restrictive and for other authors less so. Breed and Ktsanes (1961) argued that simple inaccuracy in estimates of public opinion is evidence of pluralistic ignorance. Similarly, Merton (1968) expanded the definition to include “the unfounded assumption that one’s own attitudes and expectations are unshared and the unfounded assumption that they are uniformly shared” (p. 431). Despite the advantages of these more inclusive definitions, some theorists have continued to argue for a more narrow conceptualization. For instance, Miller and McFarland (1987) argued that the term should apply only to “those illusions of uniqueness that emerge in the face of evidence of behavioral similarity” (p. 298, note 1). However, I concur with O’Gorman that pluralistic ignorance is “an erroneous cognitive belief shared by two or more people regarding the ideas, sentiments, and actions of other individuals” (O’Gorman, 1975, p. 314, note 3) or, more briefly, “false social knowledge of other people” (O’Gorman, 1988, p. 145).

This expanded definition of pluralistic ignorance subsumes several other concepts that have been studied in the area of Perceptions of others’ opinions. For example, the “false consensus effect” has been defined as occurring when people “see their own behavioral choices and judgments as relatively common and appropriate to existing circumstances while viewing alternative responses as uncommon, deviant, or inappropriate” (Ross et al., 1977, p. 280, emphasis added). Mullen (1983) argued that the false consensus effect should be distinguished from the overestimation of consensus in general. He noted that the false consensus effect is demonstrated when a Person sees his or her position on an issue as more common than someone who holds an alternative position would see it. Thus, for the false consensus effect, the misperception is relative to others and would be measured as a correlation between own opinion and Perception of the opinion of others. Alternatively, the overestimation of consensus is when people holding a given opinion Perceive more support for this opinion than there is in reality. This misperception is “absolute” in the sense that it is objectively false. Overestimate of consensus is tested as the aggregate difference (paired samples t test) between own opinions and Perceived opinions, with a significant difference in the correct direction considered supportive evidence.

Still other terms have been used to describe similar misperceptions. Fields and Schuman (1976) noted that people tend to presume that others hold the same opinions as they do themselves, which they termed “looking glass Perceptions” or social projection. Researchers have tested for the presence of looking glass Perceptions in several different ways. One of these tests is analogous to the test for a false consensus effect: a correlation between Perceptions of others’ opinions and Perceptions of one’s own opinion. Thus, from this Perspective, false consensus effects and looking glass Perceptions are equivalent (Eveland, McLeod, & Signorielli, 1995), although some still distinguish them (e.g., Glynn, Ostman & McDonald, 1995) and use different methods to test for them.

Finally, Fields and Schuman (1976) described what they termed “disowning projection” as providing the socially desirable response when asked for one’s own opinion but then projecting one’s true opinion onto the majority. In this way, one can validate one’s unpopular opinion by saying that it is held by most others while avoiding possible scorn from the interviewer for expressing a deviant opinion. While Fields and Schuman concluded that there was no evidence for this effect in their data, Glynn et al. (1989) suggested that it could account for some of her findings. In fact, this hypothesis could explain situations where the majority believes it is in the minority and the minority believes it is in the majority on sensitive social issues. Note, again, that this is specifically what Schank (1932) found and called pluralistic ignorance.

Because I have followed others and broadened the term pluralistic ignorance to include all errors in Perception of public opinion, I follow Sherman et al. (1984) and specifically label the majority perceiving itself as a minority the “false idiosyncrasy” effect (although others [Korte, 1972] have called this the “silent majority” effect). Also in regard to terminology, I use the terms false consensus effect, projection, and looking glass Perceptions interchangeably but distinguish them from the concept of overestimation of consensus, as Mullen (1983) suggested. Despite the fact that many of the theories in communication may also be subsumed under the label of pluralistic ignorance as defined here, for clarity I refer to them separately by name in this chapter.

Now that definitions have been clarified, it is possible to review some of the empirical findings in the pluralistic ignorance literature. After the work of Allport and his students Katz and Schank, pluralistic ignorance research seemed to enter a long Period of hibernation (O’Gorman, 1986). The next explicit study appearing in the literature was in 1961, when Breed and Ktsanes reported their findings on attitudes and Perceptions of segregation in Louisiana. They found that people Perceived more segregationist attitudes than actually existed but that, at least in one of two tests, the more educated respondents were less likely to hold incorrect Perceptions. They noted that pluralistic ignorance is nearly always in a conservative direction because “the error will tend to favor the older existing beliefs in the system rather than the direction of change” (Breed & Ktsanes, 1961, p. 383).

Apparent contradictory evidence for this “conservative bias” explanation was reported more than a decade later in the next study of pluralistic ignorance, when Korte (1972) revealed that students at Vassar consistently overestimated how radical the student body was in two different surveys over the course of 2 years. He labeled this phenomenon a “radical bias.” Glynn (1989) found both a conservative and a liberal bias and thus coined the term “ideological bias.”

Most other research appearing under the label of pluralistic ignorance during the mid-1970s focused on racial attitudes. O’Gorman (1975, 1979, 1976) and Fields and Schuman (1976) found consistent evidence of pluralistic ignorance in terms of false idiosyncrasy errors, overestimates of consensus, and projection effects. To date, research appearing under the label of pluralistic ignorance has expanded beyond its original home in social psychology and public opinion, appearing in areas as diverse as criminology (e.g., Kauffman, 1981; McGarrell & Sandys, 1996; Toch & Klofas, 1984), education (e.g., Long & Willower, 1980), sex research (e.g., Cohen & Shotland, 1996), and gerontology (e.g., O’Gorman, 1980) while retaining its presence in the broad fields of public opinion (e.g., O’Gorman, 1988; Shamir, 1993, 1998) and social psychology (e.g., Miller & McFarland, 1987; Miller & Prentice, 1994; Schroeder & Prentice, 1998).

It was not until the mid-1970s that social psychologists coined the term false consensus effect (Ross et al., 1977). Since the introduction of this term, the phenomenon has been studied extensively. Throughout the 1980s, researchers replicated and extended the findings of Ross et al. (1977) and attempted to provide theoretical explanations for the effect (Marks & Miller, 1987). In their meta-analysis, Mullen and his associates (1985) found 115 tests of the false consensus effect in the literature at that time and determined that the effect was quite robust. Research on the false consensus effect and overestimation of consensus has remained relatively separate from research appearing under the label of pluralistic ignorance, despite their obvious similarities. In part, this is the result of differing definitions of what constitutes an important mispPerception and different methods used to test hypotheses.

Generally speaking, pluralistic ignorance researchers have two distinct issues to address (Eveland et al., 1995). First, why are Perceptions of public opinion inaccurate? Second, what are the implications of this inaccuracy? Neither of these questions has been a substantial focus of the lines of research described in this section. This is likely in part because, as Glynn et al. (1995) noted, the research has stemmed from evidence of an empirical generality and not a complete theory, thus leading research in this area to focus on “product” over “process.” Nonetheless, recently there has been some progress made toward developing explanations for pluralistic ignorance and its potential social relevance.

Prentice and Miller (1993) argued that there are at least four possible causes of pluralistic ignorance. Impression management (sometimes termed the “self-serving bias”) is a motivational explanation suggesting that people hope to see themselves as unique or common, depending on the context. For instance, when one is successful, one is motivated (for self-impression reasons) to consider oneself as unique or “better than the rest.” On the other hand, when one fails, one is motivated to see oneself as typical; as the saying goes, misery loves company. In terms of Perceptions of the opinions of others, it has been argued that people rarely want to see themselves as part of the minority.

The second explanation for pluralistic ignorance is cognitive errors. This explanation, which has its roots in the work on accessibility biases and selective attention (e.g., Tversky & Kahneman, 1974), suggests that people derive their estimates of public opinion based on the biased pieces of information that are accessible in their memory. Thus, (a) because people tend to associate with others who are similar to themselves, (b) because consistent information is often more likely to be attended to and recalled than is inconsistent information, and (c) because one’s own opinions are more accessible than others’ opinions, cognitive errors based on this biased accessibility are believed to be the source of misperceptions of public opinion.

The differential interpretation explanation suggests that individuals publicly misrepresent their opinions, expressing more support for what they believe to be the normative position on an issue than is accurate. This is a conscious decision. However, when they use the public statements of others to assess public opinion, they do not take into account the possibility that others are also being disingenuous. This inability to interpret others’ public statements for what they are—essentially exercises in politically correctness—despite doing the same oneself leads to a state of pluralistic ignorance.

Finally, the differential encoding explanation is similar to the differential interpretation explanation except it does not assume that individuals intentionally misrepresent their own opinions publicly. For example, if a given individual is opposed to the community norm, then the individual will assume that his or her behaviors and statements will make this clear to others. However, this is not always the case. For example, Prentice and Miller (1993) noted, “Students may fail to recognize how pro-norm their public behavior actually is, mistakenly believing that their private discomfort with alcohol practices is clear from their words and deeds” (p. 253). By the same token, these same individuals will assume that others’ behaviors and statements are indicative of their respective beliefs. Once again, this is not necessarily true. Thus, when one bases his or her judgment of the opinions of others on their statements and behaviors, the judgment could be in error.

The most common of these four explanations in the social psychological literature are the impression management explanation (Gilovich, 1990; Hollander, 1994; Sherman et al., 1984) and the cognitive errors explanation (e.g., Marks & Miller, 1987; Mullen, 1983; Mullen et al., 1985). The impression management explanation and the cognitive error explanation both have been proposed as causes of biased attribution generally. However, as with the attribution literature, it is nearly impossible to distinguish these two explanations for pluralistic ignorance empirically because neither explanation is developed enough to provide for a critical test (Tetlock & Levi, 1982). Despite this problem, some researchers (e.g., Mullen, 1983) have made flawed assumptions (e.g., that impression management is necessarily conscious and intentional) and then concluded that the evidence indicates that this explanation is implausible. Perhaps the best effort to sort out the possible explanations in the pluralistic ignorance literature comes from Sherman et al. (1984). They found evidence in support of both the cognitive error explanation (effects of information on Perceptions of others’ opinions, controlling for own success or failure) and the impression management explanation (greater false consensus effects when the self fails than when the self succeeds, controlling for information effects) in the same experiment. This suggests that both processes may occur simultaneously.

Models of Social Reality Misperceptions in Communication

In this section of the chapter, I review theories dealing specifically with communication and social reality Perceptions. I describe the context of each theory, relevant evidence for and against the theory, and potential causal mechanisms that could account for the effects.

Research in communication pertaining to misperceptions of social reality can be classified into two general types. The first type of misperception studied is a misperception of media content or impact. Beliefs about what the content of media is (e.g., media bias), or beliefs about the effects that this content may have on individuals or society, are studied by researchers in the areas of third-Person Perceptions and the hostile media phenomenon. The second type of misperception is of the state of public opinion, or of social reality more generally, with a focus on some form of communication as a key contributing factor to the misperception. Research of this type is conducted under the rubric of the spiral of silence, the Persuasive press inference, and cultivation. Thus, these two types of research can be distinguished given that the first examines Perceptions of mass media content or impact as a dependent measure, while the second considers mass and/or interpersonal communication as an independent variable accounting for misperceptions of social reality.

Misperceptions of Media Impact and Content

Third-Person Effect. The third-Person effect hypothesis, as formulated by Davison (1983), has both a Perceptual and a behavioral component. The first component of the hypothesis (the “third-Person Perception”) predicts a biased Perception such that, in the aggregate, people will perceive greater media impact from Persuasive messages on others than on themselves. The second (behavioral) component builds on the Perceptual component, predicting that these biased Perceptions will result in action designed to mitigate the Perceived negative impact on others. Perloff’s (1996) review of the third-Person effect literature concluded that the Perceptual component was strongly supported but that the behavioral component had not received similar support. Recent evidence, however, suggests that there is at least some support for the behavioral component as well (e.g., Gunther, 1995; McLeod, Eveland, & Nathanson, 1997; Rojas, Shah, & Faber, 1996).

To demonstrate that the Perceptual component of the third-Person effect is actually a misperception, it is necessary to determine what an accurate Perception would be. This can be done in two ways. Most third-Person Perception research compares the aggregate Perceptions of media effects on self to aggregate Perceptions of media effects on others via paired sample t tests. While a significant difference provides evidence for a third-Person Perception, it is not necessarily evidence of an individual-level misperception given that self-reports of media effects on oneself, unlike statements of one’s own opinion on an issue, are not necessarily accurate. Therefore, an aggregate measure of Perceived effects on self is not necessarily the appropriate baseline to determine whether aggregate Perceptions of media impact on others are inaccurate in the absolute sense.

Some researchers, however, have actually been able to provide estimates of the real effect of the media content on the sample, which is an appropriate baseline for a determination of accuracy of Perceptions of others. In one of the first empirical studies of the third-Person Perception, Cohen, Mutz, Price, and Gunther (1988) found that people tended to underestimate the effects of libelous newspaper articles on themselves and generally overestimated the effects on others. Further research by Gunther (1991) demonstrated an overestimation of effects on others as compared to actual opinion change from elite and tabloid newspapers, while Perceptions of effects on self were not significantly different from the actual effects. Similar conclusions about the accuracy of Perceptions of media impact on others judged against some objective measure of effects have been provided by more recent studies as well (e.g., Gunther & Thorson, 1992; Perloff, Neuendorf, Giles, Change, & Jeffres, 1992; Price, Tewksbury, & Huang, 1998).

A brief examination of the attribution literature makes it clear why researchers have tapped attribution theory as a possible explanation for third-Person Perceptions, just as it has been tapped by pluralistic ignorance researchers. Most of the literature on biases in attribution could be applied to third-Person Perceptions; the trick, apparently, is choosing which type of bias, if any, should be seen as the basis of third-Person Perceptions. Researchers have not yet come to a firm conclusion on this point.

Rucinski and Salmon (1990) argued that the fundamental attribution error (FAE) runs counter to third-Person Perceptions because it concerns the attribution of dispositional factors to others and the attribution of situational (in this case media) factors to self. However, the exception based on effectance motivation makes this internal versus external distinction more appropriate. Effectance motivation suggests that people make attributions to maintain a belief that they have control over the world around them, as evidenced in terms such as “unrealistic optimism” (Weinstein, 1980), “just world beliefs” (Lerner, 1980), and the “illusion of control” (Langer, 1975). Therefore, although normally actors attribute the responses of self to situational factors (the FAE), the effectance motivation essentially reverses the FAE. In this case, actors would not want to see themselves as easily swayed by the media messages (and hence unable to control the world outside), so instead they infer a dispositional ability to resist Persuasive media messages. But they maintain the belief that others will respond to this situational factor (the media message). This is consistent with research in social psychology that has found that sometimes actors do avoid situational attributions of their own actions (e.g., Miller & Norman, 1975).

Gunther (1991) interpreted the FAE link somewhat differently. He argued that the actor sees himself or herself as responding to situational cues (e.g., the intent to Persuade, the possible harmful effects of the media message) in order to avoid negative effects. However, the actor perceives others to be unresponsive to these situational cues. When it comes to possible positive media influence (e.g., PSAs), Gunther stated that effectance motivation takes over and the actor Perceives himself or herself as being likely to be influenced by a positive message due to an internal disposition.

Gunther and Mundy (1993) relied more heavily on the notion of effectance motivation, specifically the notion of “unrealistic optimism.” They argued that the tendency to believe that bad things (e.g., negative media effects) will not happen to the self, but can happen to others, is a suitable explanation for third-Person Perceptions. They see this unrealistic optimism as a means of ego enhancement. A similar motivational explanation of biased optimism and ego defensiveness has been promoted by Perloff (1989) and Hoorens and Ruiter (1996), although Perloff (1996) also acknowledged that an information processing (or “cognitive error”) account for the third-Person Perception may also be appropriate.

Eveland, Nathanson, Detenber, and McLeod (1999) provided yet another explanation for third-Person Perceptions that draws on attribution theory. Their explanation is derived from the more general underlying notion of attribution theory that suggests that individuals have an innate need to provide explanations for things they observe in the world around them and that these explanations are likely to be rather naive. They suggested that individuals hold a schema for media effects that is consistent with the so-called “magic bullet” theory such that Perceptions of the media exposure of others are directly linked to Perceptions of media effects on those others. However, the evidence (McLeod, Detenber, & Eveland, 2001) suggests that the process of estimation of media effects on oneself is more complex than that on others and takes into account different factors.

The key implication of third-Person Perceptions is dealt with in the behavioral component of the third-Person effect hypothesis. Davison (1983) initially noted that the social relevance of the third-Person Perception was that individuals would take some action based on the belief that others were influenced by some media message. Initially, there was little evidence for this component of the hypothesis. However, the mid- to late 1990s produced a wealth of data suggesting that, if nothing else, Perceptions of media impact in the form of Perceived effects on self (e.g., Gunther & Hwa, 1996; McLeod et al., 1999; Rojas et al., 1996), Perceived effects on others (e.g., McLeod et al., 1999; Rojas et al., 1996; Salwen, 1998), or specifically the third-Person Perception (e.g., Gunther, 1995; Gunther & Hwa, 1996; McLeod et al., 1997; Rojas et al., 1996; Salwen, 1998; Shah, Faber, & Youn, 1999) can lead to support for censorship of the corresponding media content. Thus, it is still unclear exactly which component of Perceptions of media impact is the central force behind support for censorship. If third-Person Perceptions (in the aggregate a misperception) are the cause, then we can be confident that support for censorship is at least in part based on a misperception. However, given that it is not completely clear whether Perceived effects on self or perceived effects on others are accurate or not, we cannot say that evidence of a correlation between support for censorship and Perceived effects on self or perceived effects on others is definitively the result of a misperception.

Hostile Media Phenomenon. Political observers have noted that claims of media bias by politicians as well as academics have come from both the left and right sides of the political fence. Indeed, often both liberals and conservatives will claim to observe bias against their position in the same media coverage. The hostile media phenomenon describes the situation in which partisans on both sides misperceive what is probably relatively neutral news content as biased against their own position.

The seminal study of the hostile media phenomenon was undertaken by psychologists at Stanford University during the early 1980s (Vallone et al., 1985). They conducted a pilot study just before the 1980 presidential election and found that a majority (66%) of the respondents did not perceive media bias in the election. However, among those who Perceived bias, it was nearly always in a hostile direction. Nearly 90% of those Perceiving bias perceived it to be taking place against their favored candidate. This trend was slightly stronger for Reagan supporters (96%) than for Carter (83%) or Anderson (88%) supporters. Gunther’s (1992) survey findings also indicated that Republicans believed that both television news and newspaper coverage of Democrats was too positive, whereas Democrats thought that both television news and newspaper coverage of Republicans was too positive. A more recent summary of poll findings from the 1996 presidential campaign again demonstrated that both Republicans and Democrats Perceived bias in the news but that the bias they perceived was in different directions (Dautrich & Dineen, 1996).

Vallone and his colleagues (1985) also found that news coverage of the Beirut massacre in 1982 was perceived differently by pro-Israeli and pro-Arab partisans. Pro-Israeli partisans believed that the coverage was biased across seven different items. Pro-Arab partisans, on the other hand, believed that the direction of bias was in favor of Israel (and against the Arabs). These means were significantly different in each case. In addition, on four of these tests, a “neutral” group was significantly different from both of the partisan groups in Perceptions of media bias (falling between them). Later research has, in part, replicated these findings in similar contexts (Giner-Sorolla & Chaiken, 1994; Perloff, 1989).

When partisans on different sides of an issue have strongly divergent Perceptions of bias in the exact same news coverage, at least one of the groups must be inaccurate. Thus, the hostile media phenomenon indicates that Perceptions of the slant of news may be inaccurate. It may be assumed that the neutral groups—those without an ax to grind—have the most accurate Perceptions. However, even this is difficult to ascertain because neutral groups, as well as partisan groups, have preconceived notions that may contribute to misperceptions of the nature of media coverage.

Vallone et al. (1985) described two possible mechanisms that might produce the hostile media phenomenon. Both explanations assume that, in reality, the media content is relatively unbiased. The first explanation is based on the notion of biased assimilation (see also Kressel, 1987). It suggests that, based on past experience and observation, partisans believe that the bulk of information and factual evidence supports their own position. When presented with news content that is effectively neutral (showing equal parts favorable and unfavorable information), this appears to be inconsistent with what partisans believe to be the broader population of information available on the topic. Thus, partisans will perceive bias because the content of the media is not consistent with their Perceptions of the actual broader information environment.

The second explanation suggests that partisans on different sides of an issue may actually perceive different things when viewing, for instance, the same television newscast. Each will selectively recall hostile information better than information that is consistent with their own viewpoint. Thus, when making decisions about the fairness of a media report, partisans from different camps will effectively be making their assessments based on the different parts of the report they are best able to recall. Each will conclude that the content was biased against their position because each is better able to remember the information from the report that was hostile to their opinion.

There was evidence that each of these processes played some role in the findings of Vallone et al. (1985). Partisans did in fact have different Perceptions of the content, suggesting support for the second explanation just provided. However, even after controlling for these Perceptual differences, partisans still differed in their beliefs about media bias. Through a subtractive logic (i.e., if the second explanation did not account for the difference completely, then the first explanation must also be valid), the authors concluded that both mechanisms played a role in the hostile media phenomenon.

Misperceptions Produced by Media Content

Spiral of Silence. The spiral of silence, introduced to American researchers during the early 1970s, is a cross-level model of opinion dynamics that has received considerable research attention, if not support, in the United States. My discussion of the spiral of silence hypothesis draws heavily from the English-language writings of Noelle-Neumann (1974, 1977, 1979, 1985, 1989, 1991, 1993, 1995).

The spiral of silence hypothesis consists of several different stages and includes several important variables. First, the theory assumes that (a) most people fear social isolation and believe that to express minority opinions on “moral” issues is to risk social isolation; (b) people have a “quasi-statistical sense” that allows them to determine, if not the actual distribution of public opinion, then at least the relative trends in public opinion (sometimes called “future trend”); and (c) these Perceptions come from evidence in the environment, most importantly mass and interpersonal communication.

Thus, the public opinion process, according to the spiral of silence, works like this. To avoid the negative feelings of social isolation, people scan their environment in order to sense the climate of opinion. They make use of their Perceptions of the climate of opinion to determine what opinions can be expressed in public and what opinions must be expressed in public. When their private opinions cannot be expressed in public due to a fear of isolation, they remain silent. As others go through this same process, there is a change in the climate of opinion; the interpersonal environment appears to be even more one-sided as members of the Perceived minority fail to speak out in favor of their position. This produces changes in Perceptions of the climate of opinion, which reinforces minority members’ unwillingness to speak out. The process continues until one position becomes dominant. In addition to changes in the climate of opinion, this process can change private opinions for those who are undecided or weakly committed to their viewpoint.

The role of the mass media in the spiral of silence is to serve as an indicator of the climate of opinion. Noelle-Neumann has identified the New York Times, the Washington Post, and possibly the major television news networks in the United States as the “trend-setting” mass media because they tend to set the agenda for other news media outlets such as local newspapers. People sense public opinion via the trend-setting mass media, which Noelle-Neumann has claimed are consonant (all the same), ubiquitous (omnipresent), and cumulative (effects accumulate with repetition) and therefore not susceptible to selective exposure, attention, or recall.

According to the spiral of silence, the mass media serve as agents of social control. They convey information—although not necessarily accurate information—about the norms of society and thus opinions that may be expressed without fear of isolation. These clues may be portrayed in several different ways, including but not limited to camera angles (Noelle-Neumann, 1993), general statements about public opinion by reporters, invocation of social norms as boundaries of mainstream public opinion, actions in relation to community laws, “man-on-the-street” interviews (McLeod & Hertog, 1992), and reports of opinion polls (Salmon & Kline, 1985).

The literature has provided some evidence for the role of news media use in Perceptions of public opinion. Glynn (1987) found that those who were frequent users of mass media tended to perceive that they were dissimilar from their neighbors, whereas those who were high in interpersonal communication tended to perceive that they were similar to their neighbors. Similarly, Salmon and Neuwirth (1990) found that media message discrimination (remembering information about the relevant issue appearing in the news media) was negatively associated with perceiving one’s own opinion as being in the majority of community opinion. That is, the more a Person remembered about news media coverage of the issue (abortion in this case), the more likely the Person was to believe he or she was in the minority. Eveland and his associates (1995) also found that during the Persian Gulf War, television news viewing and, to a lesser extent, radio news use were related to the Perception that most people supported the war. This Perception was consistent with news media coverage of public opinion about the war during this time, which portrayed consonant support for the war.

On the one hand, evidence that media use can lead to misperceptions of public opinion can be interpreted as support for Noelle-Neumann’s hypothesis that media portrayals of public opinion are biased. On the other hand, it reveals that people are not necessarily capable of using their quasi-statistical sense to develop accurate Perceptions of the distribution of opinion in society. Noelle-Neumann has sidestepped this criticism of her theory by arguing that people are not necessarily able to accurately perceive the climate of opinion at any one time but that they are capable of sensing changes in the climate such that they can identify which side is gaining or losing favor. This question has rarely been examined in the literature, however, because most studies continue to be cross-sectional despite early pleas for more sophisticated designs (Glynn & McLeod, 1983). Consistent with Noelle-Neumann’s claims, however, Shamir (1995) found that Perceptions of the future trend did track along with aggregate opinions over several years. But the Perceptions of majority opinion followed aggregate opinion only when an election made the current state of opinion abundantly clear.

While Shamir’s (1995) evidence appears to support Noelle-Neumann’s theory, the study also suggested that these accurate future trend Perceptions did not derive from media coverage of opinions or from interpersonal discussion as Noelle-Neumann has claimed. Instead, Perceptions of future trends came from media reports of objective events that would presumably influence public opinion. This is consistent with Gunther’s (1998) Persuasive press inference hypothesis (to be discussed in the next section). Perceptions of the current climate (which were inaccurate in the aggregate), on the other hand, were derived by extrapolation from “known distributions of opinion” such as party membership and ideological distributions. Thus, these findings shed some doubt on Noelle-Neumann’s version of the spiral of silence.

However, it should be noted that known distributions of opinion must come from somewhere. The reason why these opinion distributions are “known,” I would submit, is that they are covered in the news media via polls and other means. Therefore, while the distributions may be “known” (i.e., in the mind) at the time they are used to estimate other opinion distributions, they are known only because they were at some point learned via the news media reports of public opinion polls, the party controlling the government, election results, and so on.

As already noted, the spiral of silence suggests that a quasi-statistical sense allows individuals to use information in the environment—primarily news media and interpersonal communication—to shape Perceptions of the climate of opinion. Based on the spiral of silence, misperceptions of the state of public opinion could be caused by several sources. One of the key explanations has been labeled the “dual climate of opinion” (Noelle-Neumann, 1993). This explanation centers on situations in which the climate of opinion generated through interpersonal expression of opinions is inconsistent with the climate of opinion portrayed in the trend-setting mass media. For example, if media coverage portrayed public opposition to the impeachment of President Bill Clinton during the fall of 1998, but those in support of impeachment appeared to be the majority based on public expressions of opinion, then a dual climate of opinion would exist. In this situation, some might misperceive the climate of opinion to be that portrayed by the mass media. In effect, this explanation for misperceptions of the state of opinion is simple: media bias (or at least inaccurate media coverage).

Alternatively, because those who Perceive themselves to be in the minority are less willing to publicly express their opinions, opinions held by these individuals will become less prominent in public conversations. As this happens, it leads to an incongruity between the expressed climate of opinion and the sum of individual Personal opinions. Social observers will infer that a given viewpoint is in the minority because few individuals are expressing this view publicly, but in fact this unexpressed view may be the privately held view of the majority. President Richard Nixon claimed just this when he spoke of a “silent majority” that supported the conservative cause.

Research on the influence of Perceptions of being in the minority (whether or not this is a misperception) and willingness to speak out publicly from the spiral of silence Perspective has produced mixed results in the United States (for a meta-analysis, see Glynn, Hayes, & Shanahan, 1997). Despite this, there has been enough positive evidence to suggest that something like the spiral of silence does indeed occur. In fact, one of the most vocal American critics of the spiral of silence, Chuck Salmon, acknowledged, “The essence of the model—that individuals’ Perceptions of their environment do have some bearing on their communication and behavior … is incontestable” (Salmon & Moh, 1992, p. 159).

In addition to affecting willingness to speak out, Perceptions of public opinion can eventually influence private opinions, at least among those who initially hold weak opinions. Some evidence has provided support for this notion. For instance, pluralistic ignorance researchers have demonstrated that Perceptions of being in the minority can, over time, produce conformity (Prentice & Miller, 1993; Schroeder & Prentice, 1998). Public opinion research (e.g., Eveland et al., 1995; Glynn & McLeod, 1984, 1985) has also provided some evidence for this component of the spiral of silence hypothesis.

Persuasive Press Inference. The Persuasive press inference (Gunther, 1998) is the most recent of a long line of research on Perceptions of media impact, and it draws heavily on research examining the third-Person Perception and hostile media phenomenon. The Persuasive press inference is the inference that individuals make about the power and reach of news media. In essence, the claim is that individuals observe news content and come to some conclusion about the valence of this content relative to a particular topic.

For example, news of the murder of more than a dozen people in Columbine, Colorado, by two students using semi-automatic weapons and bombs may be evaluated as providing an argument in favor of more strict gun control laws. Some evidence indicates that the hostile media phenomenon would play a role in the development of Perceptions of the direction of media coverage (Gunther & Christen, 1999b). Therefore, those opposed to gun control would likely perceive the content to be even more slanted in favor of gun control. In general, viewers would assume that the coverage to which they are personally exposed is representative of the population of news media coverage. They would also make assessments about the extent of potential impact from this content. Finally, using an existing Perceived opinion baseline, they would update their Perceptions of public opinion to include Perceived effects of the media coverage in the direction of the slant that they perceive. The research to date in both experimental and survey contexts (Gunther, 1998; Gunther & Christen, 1999a, 1999b) generally supports the Persuasive press inference. This support holds, in most cases, even after controlling for the effects of opinion projection on estimates of overall public opinion.

The Persuasive press inference suggests that misperceptions of public opinion may be due to misperceptions of mass media impact on public opinion. Specifically, Gunther (1998) argued that people will overestimate the impact of news media coverage on public opinion and that, therefore, estimates of public opinion will be inaccurate. The underlying mechanisms that cause misperceptions of the impact of mass media, however, are not clearly defined from this admittedly new Perspective, although the theory draws on the existing research on third-Person Perceptions (Gunther & Christen, 1999b).

Cultivation. The cultural indicators approach to the study of mass media effects provides a well-rounded theoretical Perspective from which to examine more general Perceptions of social reality. This section, where not referenced, draws heavily from the writings of the key members of the cultural indicators group that originated at the University of Pennsylvania, including George Gerbner, Michael Morgan, Larry Gross, and Nancy Signorielli (e.g., Gerbner, Gross, Morgan, & Signorielli, 1980, 1994; Morgan & Shanahan, 1997; Signorielli & Morgan, 1990, 1996).

The cultural indicators approach has three main components. Media message analysis is the component of the cultural indicators approach that examines the content of television. Other media forms—newspapers, radio, magazines, and the like—are outside the domain of cultivation theory as described by its originators. The most prominent finding of the media message analysis is that television is full of violence and mayhem. Research has also consistently demonstrated that the characters and situations shown on television are different from those in the real world. For example, television characters are more likely to be male, White, and young as compared to other segments of the U.S. population. This disparity between the television world and the real world is a linchpin of the theory.

Institutional process analysis is the component of the cultural indicators approach that focuses on how the media messages that exist are chosen. Specifically, this component attempts to answer the question, “Why are the media messages we have the way they are?” Researchers have pointed to the roles of the profit motive in commercial television as well as the culture and ownership structure of the television industry as potential answers to this question.

Cultivation analysis seeks to connect the content of television to the public’s Perceptions of social reality. Because media message analysis indicates that the content of television is not an accurate portrayal of the real world, people who use the television world to shape their Perceptions of the real world will come away with a serious misunderstanding. Cultivation analysis suggests that this is exactly what happens. Specifically, the most prominent claim of cultivation researchers is that the violent world of television has created a “mean world syndrome” in which heavy viewers believe that crime and violence are much more prevalent than they are in reality and that most people cannot be trusted.

In addition to Perceptions of a mean and dangerous world, other misperceptions of social reality have been linked to television viewing. Cultivation researchers have examined the role of television in forming stereotypes of social groups and behaviors, including Perceptions of social out-groups and minorities (Gandy & Baron, 1998; Kiecolt & Sayles, 1988), Perceptions of professionals such as lawyers and doctors (Pfau, Mullen, Deidrich, & Garrow, 1995; Pfau, Mullen, & Garrow, 1995), and Perceptions of sex roles and marriage among young people (Morgan & Rothschild, 1983; Rosenwasser, Lingenfelter, & Harrington, 1989; Signorielli, 1991; Signorielli & Lears, 1992). Generally speaking, television use tends to produce stereotypical Perceptions of these groups and behaviors.

Beginning during the early 1980s, researchers began to directly challenge the findings of the original cultivation researchers, sometimes reanalyzing the same data but producing conflicting results (e.g., Hirsch, 1980, 1981a, 1981b; Hughes, 1980; for a more recent critique, see Potter, 1994). These researchers demonstrated that simultaneous controls for demographic characteristics substantially reduced or eliminated relationships between television viewing and Perceptions of social reality. They criticized the underpinnings of the theory, including the assumptions that television viewing is nonselective and television viewing’s effects on social reality Perceptions are linear. Many researchers have also clearly demonstrated that exposure to different genres of television, including television news, has different effects (e.g., Hawkins & Pingree, 1981; McLeod et al., 1995; O’Keefe, 1984; O’Keefe & Reid-Nash, 1987; Potter & Chang, 1990; Rubin, Perse, & Taylor, 1988).

This suggests that overall television viewing is unlikely to be the true culprit. The early criticisms and their corresponding rebuttals in major communication journals—sometimes known as the “cultivation battles”—led the original researchers to make several revisions to the theory. However, they remained steadfast in most of their major claims and assumptions.

The two major changes to cultivation theory were labeled “mainstreaming” and “resonance.” These revised versions of cultivation were designed to account for nonsignificant and contradictory findings by specifying the potential for cultivation effects to vary across social categories. Both mainstreaming and resonance thus suggest interactions between background characteristics or Personal experience and the effects of television viewing, but the forms of these interactions differ.

Mainstreaming suggests that television can move heavy viewers toward a mainstream view or Perception by having different effects on viewers with different background characteristics—a prediction of a “convergent” interaction. In so doing, for instance, heavy television viewers from the South and Northeast would become more similar in their views on morality despite the fact that light television viewers from both regions were very different in their views on morality. Another means of testing this hypothesis suggested by some authors (e.g., Morgan, 1986) is to compare the variance in the criterion measure for heavy and light television viewers. The hypothesis would be supported statistically if the variance were significantly less among heavy television viewers than among light television viewers, using Levene’s test for homogeneity of variance as the statistical criterion (McLeod et al., 1995). Some researchers have found evidence consistent with mainstreaming effects (e.g., Shanahan, Morgan, & Stenbjerre, 1997), but a lack of theoretical specification about when main streaming effects should be expected (as opposed to overall cultivation effects), a lack of use of appropriate statistics to test this hypothesis (see Eveland, 1997; McLeod et al., 1995), and confusion about what the mainstreaming hypothesis really means make generalizations about support difficult.

The resonance hypothesis also predicts an interaction between background characteristics and television viewing such that effects of television are greater when its content is consistent with the real-life situation of the viewer. Thus, unlike mainstreaming (which predicts a convergent interaction), resonance predicts a “contributory” interaction (McLeod & Reeves, 1980). That is, the impact of television viewing would be greater among groups of individuals for whom the television messages “resonate,” thus leading Perceptions of these groups to be even more different from other heavy viewers than light viewers are from other light viewers. Once again, while some studies have found evidence for resonance (Gerbner et al., 1980), there is a serious lack of a priori specification of what situations warrant resonance versus mainstreaming versus overall cultivation predictions. When predictions are vague, both finding significant differences across subgroups and not finding significant differences across subgroups can be interpreted as consistent with the theory. This has led some researchers to argue that the mainstreaming and resonance hypotheses, as formulated by Gerbner et al. (1980), make cultivation un-falsifiable (Hirsch, 1981b).

The surface explanation for misperceptions as demonstrated by cultivation research would point to the results of the media message analysis component of the cultural indicators program. These content analyses, primarily conducted by the original cultivation researchers (e.g., Signorielli, 1981, 1986, 1990), have consistently demonstrated that television misrepresents social reality along a number of dimensions. Television programs suggest that crime is more likely to be violent than it is in reality (e.g., Oliver, 1994). Television also misrepresents the distribution of people along various demographic characteristics such as gender, ethnicity, age, and marital status (Signorielli, 1981). Thus, the underlying claim of the cultivation hypothesis is that these (mis)representations of reality on television are learned by viewers (particularly heavy viewers) and used in judgments about the real social world. If television provided an accurate description of the real world, then cultivation would change from an antisocial effect leading to misperceptions to a prosocial effect of simple learning about the world.

Despite scores of empirical studies of cultivation (for a qualitative and quantitative review, see Morgan & Shanahan, 1997), early researchers made few attempts to address the cognitive processes through which cultivation effects occur. One prominent exception was the program of research by Hawkins and Pingree (e.g., Hawkins & Pingree, 1982, 1990). A key distinction advanced by these authors is that there are two types of cultivation Perceptions: first order and second order.

First-order (also called demographic) beliefs are estimates of the frequency or probabilities or characterization of events, and one can clearly distinguish a “television” answer from a “real-world” answer…. Second-order beliefs, such as fear of walking alone or mistrusting strangers, have no quantifiable referent in television content, but one can argue that the prevalence of violence between strangers implies that one should fear and mistrust strangers. (Hawkins & Pingree, 1990, p. 49, emphases in original)

Hawkins and Pingree argued that because the correlations between first-order and second-order cultivation beliefs are typically small to nonexistent (but see Potter, 1991), and because research indicates that the relationship between television viewing and second-order beliefs is often reduced to nonsignificance when simultaneous controls are applied (but see Potter, 1991), it is possible that these two types of cultivation effects are based on entirely different psychological processes. However, despite some basic suggestions for future research to determine the processes involved in these cultivation Perceptions, Hawkins and Pingree were unable to offer solid empirical evidence for any particular psychological process.

Following Hawkins and Pingree (1982, 1990), researchers have recently made more explicit efforts to understand the underlying processes that drive the cultivation effect (e.g., Mares, 1996; Shapiro, 1991; Tapper, 1995). For example, Shapiro (1991) argued for a model of retrieval, weighting, and balancing of specific events from episodic memory as the cognitive process behind social reality Perceptions. As part of the weighting and balancing process, individuals were thought to make use of factors such as the source of the memory (e.g., direct experience, television, newspapers). In a related vein, Mares (1996) argued that source confusions—the tendency for individuals to think that events from entertainment actually came from the news—were related to the cultivation effect. Specifically, she argued that those who tend to mistake entertainment sources to be news sources in recall tests will evidence stronger cultivation effects, while those who tend to mistake news sources as entertainment sources will demonstrate weaker cultivation effects.

Probably the most empirically successful attempt to explain the mechanisms behind the cultivation effect can be found in the work of L. J. Shrum in his “heuristic model of cultivation effects” (Shrum, 1997, 1998). Shrum and O’Guinn (1993; see also Shrum, 1995, 1996) suggested that first-order (but not second-order) cultivation effects may be explained by combining insights from research on construct accessibility and the availability heuristic. Construct accessibility refers to the ability to easily access information from memory. Information that has been recently used, and information that is frequently used, both are more accessible in memory and thus are more likely to be used in making judgments. The availability heuristic suggests that because individuals are cognitive misers, they normally do not make thorough searches of memory when faced with the need to make decisions. Instead, they use cognitive shortcuts that often achieve acceptable results with significantly less effort. The availability heuristic states that information that is found to be easily accessible (i.e., information that comes easily to mind) is believed to be common and typical.

Given this, Shrum and O’Guinn (1993) provided the following account of cultivation. When asked to make a judgment about the frequency of some event, such as the relative amount of violence, individuals use the availability heuristic. The information that is most accessible is the information that has been recently and frequently activated. Given the high rate of violence on television, heavy television viewers will have instances of violence easily accessible. Thus, television viewing should contribute to higher (and generally inaccurate) estimates of the rate of violence in society. The limited research evidence for this model so far generally supports these assertions (Shrum, 1996; Shrum & O’Guinn, 1993).

As Hawkins and Pingree (1990) correctly noted, “First-order beliefs are trivial if there is no further connection to other beliefs or behavior” (p. 45). What they meant is that if there is little relationship between Perceptions of the demographics of the real world and things such as fear, interpersonal mistrust, racism, and so forth, then why should we care about the impact of television viewing on Perceptions of social reality? Unfortunately, there appears to be only a weak relationship between first-order and second-order cultivation beliefs (Hawkins & Pingree, 1982, 1990; but see Potter, 1991), so the socially relevant impact of first-order cultivation beliefs is unclear.

However, the admittedly limited direct impact of television on second-order cultivation beliefs does have important implications. Gerbner, Gross, Signorielli, Morgan, and Jackson-Beeck (1979) argued that the images of violence on television cultivate Perceptions of fear of crime. What is most important about this, they argued, is that those who hold these mean world Perceptions “may grow up demanding protection and even welcoming repression in the name of security” (p. 196). There is in fact some recent evidence that this claim is true (McLeod et al., 1996). McLeod and his students demonstrated that fear of crime is a significant and strong predictor of support for punitive measures to reduce crime (e.g., the death penalty, “three strikes” laws), whereas it is unrelated to support for preventive measures such as spending money to create jobs and requiring rehabilitation for prisoners. These relationships (and lack thereof) held up at several levels of control, including controlling demographics, political ideology, modes of thought (similar to authoritarianism), media use, and information processing. However, it should be noted that in this study it was not television generally, but instead exposure and attention to crime news in newspapers and on television, that was hypothesized and found to be driving fear of crime. The impact of time spent watching television more broadly was not examined.

Integration of Perspectives on Misperceptions of Social Reality

It should be evident to the reader after this review of each of the preceding areas of research on Perceptions of social reality that many connections among these areas exist. Scholars studying Perceptions of social reality should attempt to integrate these different programs of research by searching for areas of agreement and disagreement among theories and Perspectives that all focus in some way on misperceptions of the world around us, explanations for these observed misperceptions, or socially relevant implications of these misperceptions. Where agreement across threads of research is found, integration of research programs can take place and contribute to theoretical parsimony and cross-disciplinary collaboration. Where disagreement exists, competing hypotheses may be developed and tested, leading to a further reduction in the number of theories and potentially a greater understanding of misperceptions of social reality.

Some scholars have already been working to integrate at least two or three of these areas and should be congratulated for doing so (e.g., Edelstein, 1988; Glynn et al., 1995; Gunther, 1998; Mutz, 1998). As another step toward facilitating this integration, Tables 31.1, 31.2, and 31.3 lay out some initial dimensions on which these Perspectives share foci, explanations, and implications.

Table 31.1 Relevant Perceptions about Social Reality Relevant Perceptions about Social Reality

Content of Mass Media	Effects of Mass Media	Characteristics of Others and Society
Hostile media phenomenon	Third-Person Perceptions	Pluralistic ignorance
	Persuasive press inference	Cultivation
		Spiral of silence
		Persuasive press inference

The most obvious connection that can be made between these areas of research on social reality Perceptions is on the basis of their common foci (see Table 31.1). Pluralistic ignorance research, the Persuasive press inference, the spiral of silence, and cultivation all deal, at least in part, with how individuals misperceive characteristics of others and of society. However, the spiral of silence and the Persuasive press inference are more closely related to pluralistic ignorance research than to cultivation because they generally focus on Perceptions of the opinions of others. Research on the third-Person Perception (and to some extent the Persuasive press inference) is concerned with misperceptions of the impact of mass media messages on the public. The Persuasive press inference in some ways provides an explanation for why the third-Person Perception can lead to pluralistic ignorance, that is, by updating Perceptions of public opinion when media impact is perceived. Finally, the hostile media phenomenon is focused on the content, instead of the effects, of mass media; therefore, it is probably not well-connected to pluralistic ignorance but can provide some help in understanding the third-Person Perception and the Persuasive press inference. Given all of these (and potentially other) connections, it only makes sense that interested researchers work to integrate findings from each of the traditions both within and outside of communication.

Table 31.2 reveals that, as might be expected, there is some disagreement about the causes of misperceptions across the areas reviewed in this chapter. Many of these differences are, in part, the result of research originating in different fields. Theories originating in communication, such as cultivation, the Persuasive press inference, and the spiral of silence, tend to rely on the media and interpersonal communication as at least part of the explanation for misperceptions of public opinion. Specifically, these theories suggest that when the content of communication does not mirror reality, it can produce misperceptions of reality. By the same token, if communication were an accurate representation of the true world (i.e., if we all told the truth and if both news and entertainment content in the media portrayed the world as it is), then communication could lead to accurate Perceptions of reality. However, particularly for mass communication—even news media— portraying the world accurately is unlikely to happen because most people want to experience a world that is different from the one they experience every day. News, for instance, does not tell us what is normal about our world but instead stresses that which is atypical such as natural disasters, terrorist attacks, and mass murders.

Table 31.2 Explanations for Misperceptions of Social Reality

Inaccurate Media Portrayals	Inaccurate Interpersonal Portrayals	Motivational or Cognitive Biases
Cultivation	Spiral of silence	Third-Person Perceptions
Spiral of silence	Pluralistic ignorance	Cultivation
Persuasive press inference		Hostile media phenomenon
		Pluralistic Ignorance
		Persuasive press inference

Theories going through their infancy in psychology tend to focus on psychological processes such as motivational or cognitive biases in Perception. In the case of motivational explanations, it is suggested that individuals have either a conscious or an unconscious desire to see themselves as better than others. Perceptions are thus shaded by this need. Cognitive error explanations rely on research demonstrating that individuals use heuristics to simplify decision-making and inference processes. These heuristics sometimes lead to errors, including errors in Perceptions of social reality.

Despite differences, there has been a trend toward cross-disciplinary similarities in explanations during recent years. For example, much of the current theorizing about cultivation focuses on cognitive processes, specifically biases introduced by heuristics (e.g., Shrum, 1995). Similarly, explanations for the third-Person Perception have drawn heavily from the attribution literature that is the basis of much of pluralistic ignorance (e.g., Eveland et al., 1999; Gunther, 1991). Third-Person Perception researchers have also struggled with the same problem of distinguishing motivational from cognitive error explanations that has plagued theoretical advancement of research on pluralistic ignorance. Nevertheless, it appears that, more often than not, motivational biases are suggested as explanations for third-Person Perceptions.

The diffusion of theoretical explanations across disciplinary lines tends to be unidirectional in the case of communication. That is, while communication researchers have begun to borrow theory related to social reality Perceptions from the level fields of psychology and sociology, there has been relatively little movement of theory from communication to researchers studying pluralistic ignorance outside of communication. One exception is that scholars studying pluralistic ignorance have acknowledged the potential role of interpersonal communication (both verbal and nonverbal) in contributing to misperceptions of the opinions of others (e.g., Korte, 1972; Prentice & Miller, 1993). Specifically, of the four explanations for pluralistic ignorance provided by Prentice and Miller (1993) and described in an earlier section— impression management, cognitive errors, differential interpretation, and differential encoding—at least three involve either interpersonal or mass communication, in conjunction with psychological processes, as central components. In fact, the differential interpretation explanation is wholly consistent with the interpersonal component of the spiral of silence, albeit much less fully developed than Noelle-Neumann’s theory.

Table 31.3 Effects of Misperceptions

Communication	Attitudes	Behaviors
Spiral of silence	Cultivation	Cultivation
Third-Person Perception	Spiral of silence	Pluralistic ignorance

While implications are often suggested but rarely examined empirically in this literature (exceptions are the spiral of silence and the third-Person effect), it is also instructive to compare the claims for the ultimate social relevance of misperceptions of social reality (see Table 31.3). As might be expected, communication theories of social reality Perceptions have, in part, focused on the impact on further communication. The spiral of silence suggests self-censorship of opinions deemed to be in the minority, whereas third-Person effect researchers have demonstrated support for societal-level censorship in response to third-Person Perceptions. Researchers have also posited that misperceptions of public opinion can potentially change opinions on that same topic (e.g., the spiral of silence, pluralistic ignorance) or more general attitudes such as beliefs about the need for social control (e.g., cultivation). Finally, some theories describe a more ambiguous impact on behaviors due to misperceptions of public opinion.

Integration with Theories of Persuasion

We know that the empirical link between Perceptions or beliefs and attitudes and behaviors is often limited by a number of factors. Despite this, research on social reality Perceptions across a number of fields has repeatedly demonstrated that one’s Perceptions of the world are at least a partial determinant of attitudes and behaviors (e.g., Glynn et al., 1997; Hetherington, 1996; McLeod et al., 1996, 1997; Perkins & Wechsler, 1996; Prentice & Miller, 1993). This is consistent with a number of theories of Persuasion that may be instrumental in any theorizing about Perceptions of social reality.

By now, the reader with a background in Persuasion has probably noted a number of theories of Persuasion that would suggest how Perceptions of social reality as described in this chapter would be used in an individual’s construction of attitudes and behaviors. Many theories of Persuasion suggest some form of a link between Perceptions and opinions and behaviors or are in some other way concerned with Perceptions of social reality. For illustrative purposes, I discuss how the research reviewed in this chapter may be integrated with several theories of Persuasion both to advance social reality Perception research and to extend Persuasion research.

Balance and Coorientation Theories

Balance theory (Heider, 1958; Monsour, Harvey, & Betty, 1997) and coorientation theory (McLeod & Chaffee, 1973; Newcomb, 1953) both deal with the interrelationships among an individual, his or her Perceptions of other individuals (or groups), and evaluations of objects. Balance theory is more closely related to Persuasion, while coorientation theory is more closely related to communication. However, both theories are centrally focused on the role of Perceptions of others as an important component of how individuals respond to objects in their social environment. As such, both theories can draw important insights from research on Perceptions of social reality.

Briefly, balance theory postulates that an individual (p) has an evaluation of both an object (x) and another individual (o). The individual also perceives the relationship between the object and the other Person. These three evaluations may be in balance or out of balance. If the triad is out of balance (e.g., the individual likes both the object and the other but Perceives that the other does not like the object), there is a motivation to bring the system into balance. This can be done by changing any one of the three relationships. Change can occur if the individual (a) decides that he or she does not like the object (to bring the individual’s attitude in line with a liked other), (b) decides that he or she does not like the other (to bring the individual’s attitude about the other in line with his or her attitude about the object disliked by the other), or (c) he or she may change the Perception that the other dislikes the object. Thus, balance theory can be used to explain both how Perceptions of the opinions of others can potentially lead to attitude change and how Perceptions of the attitudes of others may be inaccurate due to a motivation to maintain balance.

From the Perspective of coorientation theory, McLeod and Chaffee (1973) argued that interpersonal communication can lead to accuracy in Perceptions of others. This, they claimed, “seems an ideal criterion for communication in that it is (theoretically, at least) achievable through communication alone” (p. 487). However, they acknowledged that this is not always the case (communication often is used to decrease accuracy), nor should it be given that accuracy is not always the most productive condition for social relationships.

Mass communication could, theoretically, also increase accuracy in Perceptions of others, especially if the other is a social group as opposed to an individual. Indeed, researchers have tested the coorientation model in the context of social groups (e.g., Grunig & Stamm, 1973). However, mass communication is more likely to decrease accuracy through its role in the development of pluralistic ignorance as described previously. This was first noted decades ago (Tichenor & Wackman, 1973), but little connection has been made since then between coorientation (or balance theory) and research on social reality Perceptions in mass communication. Despite this lack of follow-up, progress in social reality Perception research could do much to inform Persuasion research from the balance theory Perspective, and social reality Perception research could benefit from a review of the literature from balance theory and coorientation.

Theory of Reasoned Action

One classic Persuasion theory that is dependent on Perceptions of social reality is the theory of reasoned action (Fishbein & Ajzen, 1975; Sutton, 1998). The theory of reasoned action predicts that behaviors are most proximally determined by behavioral intentions. Behavioral intentions, in turn, are determined by attitudes toward the behavior and “subjective norms.” A subjective norm has two components: normative beliefs and the motivation to comply. The normative belief is a Perception of what opinions or behaviors are considered acceptable or favored (or unacceptable or disfavored) by important others along with a “motivation to comply.”

Research on the construction of social reality could make a substantial contribution to the theory of reasoned action by providing a solid basis of research for the development of normative beliefs and the potential impact of motivations to comply on behavioral intentions. Therefore, those interested in Persuasion might consider integrating research described in this chapter pertaining to the development of Perceptions of social reality with the theory of reasoned action. Specifically, ideas of how Perceptions of public opinion are shaped based on research in pluralistic ignorance, the spiral of silence, and the Persuasive press inference would fit nicely into the theory of reasoned action to demonstrate one potential avenue for attitude change and behavior modification.

In addition, the spiral of silence’s concept of fear of isolation appears to be closely related to the theory of reasoned action’s notion of motivation to comply. Specifically, the spiral of silence suggests that individuals have an innate fear of social isolation and that they engage—or choose not to engage—in communication behaviors based on this fear of isolation. Thus, fear of isolation provides at least one motivation to comply, although there are likely others as well. However, this connection between the spiral of silence and the theory of reasoned action again indicates that research on social reality Perceptions may be able to contribute both theoretical advances and a substantial base of research toward a theory of Persuasion.

Cognitive Response Theory

Cognitive response theory (Greenwald, 1968; Petty & Cacioppo, 1981) claims that attitude change in response to Persuasive messages takes place through an internal process of argument and counterargument generation. That is, when exposed to a Persuasive communication, individuals will be persuaded not by the message itself but rather by the arguments they are able to generate in favor of and opposition to the position advocated in the message. Thus, it is the individual’s cognitive response to the message, and not the message itself, that is the ultimate source of Persuasion.

Advancing the cognitive response model, Mutz (1992, 1998) argued that when individuals are presented with information about the state of public opinion from poll information (in our case, this could simply be Perceptions of public opinion), they sometimes go through a process of cognitive response that results in attitude change. Faced with a belief (potentially created through media poll reports or other depictions of public opinion as described by the spiral of silence) that a majority of the public supports a certain policy proposal, individuals produce arguments to explain and understand this position. They also develop counterarguments about why the majority position might be wrong. In the end, as with the cognitive response model more generally, it is the arguments generated in response to the communication, and not the communication itself, that determines whether a change in attitude or behavior will take place. By focusing on the role of Perceptions of social reality in the form of public opinion, Mutz’s extension of the cognitive response model provides an important linkage between Persuasion research and both pluralistic ignorance research in psychology and the spiral of silence model.

Credibility and Social Reality Perceptions

A key construct that can influence the Persuasive potential of a message is the Perception of the credibility of the source of the message (Hovland, Janis, & Kelley, 1953; Slater & Rouner, 1996). In his study of Perceptions of news media bias, Gunther (1992) argued that credibility Perceptions may be determined more by the characteristics of the individual than by the characteristics of the source itself. For example, he found that the best predictor of Perceptions of biased media coverage of social groups was membership in those social groups, which is consistent with the hostile media phenomenon. Given the centrality of the credibility construct in Persuasion, and given that media credibility Perceptions are susceptible to bias through the hostile media phenomenon, there is a need for integration of these strains of research. Specifically, Persuasion researchers concerned about the credibility of different media sources should be sure to take into account the fact that those who are highly involved, such as members of certain social groups, may Perceive the content to be biased even when the source appears to be objectively credible.

There is an additional linkage between credibility research and social reality Perceptions. The “sleePer effect” described in early research on source credibility and Persuasion (Hovland et al., 1953; Pratkanis, Leippe, Greenwald, & Baumgardner, 1988) may also be at play in media influence on Perceptions of social reality. When a message from a source lacking credibility has a greater Persuasive effect after a delay than immediately after exposure, this is labeled a sleeper effect. One explanation for sleeper effects is that, over time, the message and source are dissociated. When this occurs, the individual is unable to discount the message based on the source and thus is more likely to accept the message.

Cultivation researchers have used this basic groundwork to explain the impact of televised entertainment content on Perceptions of social reality. Mares (1996) argued that the impact of television entertainment programming on social reality Perceptions is greater among those individuals who regularly misremember the source of information to be from a newscast instead of an entertainment show. Basically, Mares was interested in tapping an individual difference in susceptibility to a sleeper effect. Her reasoning, consistent with the sleeper effect, was that those individuals who can accurately link the source of a message with its content are less likely to be influenced by entertainment media because they are able to discount the source. However, over time, some individuals are more likely to misremember the source such that information garnered from entertainment television is thought to originate from the more credible news genre. This confusion of source, and thus source credibility, leads to a greater impact on Perceptions of social reality. Mares found support for this interpretation of the cultivation effect.

Applications to Communication Campaigns

Miller and McFarland (1987) provided a concise statement of the general fear expressed by many researchers about the social consequences of misperceptions of social reality: “Pluralistic ignorance can lead a group to act in a manner that is inconsistent with the inclinations of its members” (p. 304). One avenue that campaigners may pursue is to attempt to alter Perceptions of social reality to more closely match reality. As I have already described, research indicates that Perceptions of social reality may be influenced by exposure to both entertainment media and news media. Furthermore, many different types of Perceptions—including Perceptions of public opinion, of the behavior of others, and of social conditions—are susceptible to influence.

Much of the research on Perceptions of social reality described in this chapter places the finger of blame on the media for inaccurate Perceptions of social reality. It has also taken a relatively deterministic Perspective regarding the content of the media. That is, the existing research takes media content as a starting point to determine its effects. However, a campaigner would likely examine media content from another Perspective: How might media content be influenced to produce a more desirable outcome?

One means of exerting influence over media content is through public pressure on networks (Weinraub, 1999). Protests may be held, or a broad-based boycott may be attempted. Similarly, social action campaigners may privately lobby television producers to adjust their content for some socially desirable purpose and then publicly reward the producers when they respond appropriately (Montgomery, 1989). Unfortunately, the truly successful application of these types of influence is likely to be relatively rare in the United States because the creative forces behind television content are typically reticent to give in to attempts at control over their creative product. However, campaigners may have greater influence on noncommercial stations such as PBS or in contexts where television is viewed more as a public commodity such as in developing or socialist countries.

Influence over news media content, on the other hand, is a major focus and goal of public relations practitioners and political consultants. The impact of public relations on news content takes place through the framing of press releases, press conferences, and on-air interviews. Presidential administrations attempt to influence which issues are considered important to the public by maintaining a consistent theme among all members of the administration, sometimes called the “line of the day” (Maltese, 1992).

Provided that they are able to exert some influence over media content, there are a number of ways in which campaigners could use the Persuasive power of entertainment and news media to influence public beliefs about social reality. In what follows, I describe three areas in which campaigners might wish to apply the research on communication and social reality Perceptions to real-world issues. These issues are health and development campaigns, changing racial and sexual stereotypes, and political campaigns.

Health and Development Campaigns

Social reality Perceptions play a significant role in a number of health-related concerns. For instance, a number of social scientists (e.g., Marks, Graham, & Hansen, 1992; Perkins & Wechsler, 1996; Prentice & Miller, 1993; Schroeder & Prentice, 1998) have demonstrated that Perceptions of social reality in the form of Perceived norms have important implications for the use of alcohol by adolescents and college students. Research has consistently demonstrated that “one of the most consistent predictors of an adolescent’s alcohol use is Perceived alcohol use by his or her peers” (Schroeder & Prentice, 1998, p. 2151). Importantly, longitudinal research (Marks et al., 1992) has demonstrated that the social conformity effect (changing one’s own behaviors to fit Perceptions of others) is stronger than the social projection effect (changing one’s Perceptions of others to fit one’s own behavior). The link between Perceived norms and behavior has been replicated for other health-related behaviors, including sexual experiences and smoking among young adults (e.g., Botvin, Botvin, Baker, Dusenbury, & Goldberg, 1992; Cohen & Shotland, 1996).

Any health campaign designed to reduce alcohol use, smoking, or risky sexual behaviors by adolescents and college students would be well-served by attempting to alter Perceptions of the social norm regarding the behavior. As Montgomery (1989) detailed, there is a substantial history of groups attempting to influence American television content for public health purposes. For example, characters on entertainment television programs can regularly engage in safe sex as opposed to unprotected sex or can abstain from sexual activity altogether.

One example of how change might occur in public Perceptions of social norms regarding health behaviors can be drawn from the 1990s teen soap opera Beverly Hills 90210. For several years of this series, the character “Donna” abstained from sexual intercourse while her high school and college peers were sexually active. For a truly significant impact, more than a single character would probably be necessary to influence Perceptions of the number of individuals who remain virgins through late high school or early college.

Another context in which television content could be influenced would be in automotive safety. For example, based on outside pressures, characters on several popular television series during the 1980s, including The A-Team, began to visibly buckle their seat belts each time they entered their cars (Geller, 1989). This change in behavior could have led viewers to perceive a change in public norms regarding the use of safety belts and to adjust their own behaviors accordingly. Similarly, if instances of drinking and driving are followed by accidents and other negative consequences, such as the drunk driving accident of the character “Bailey” on Party of Five, regular viewers may become likely to associate these events occurring together frequently. A related approach would be to portray more occurrences of the use of designated drivers in situations where a lead character is drinking before driving.

News media content can also be influenced to more accurately portray the frequency of healthy and unhealthy behaviors. Surveys can be publicized that report the frequency of, for instance, unprotected sex or the number of sexual partners by a given age. These figures can be contrasted with the nature of sex portrayals on television. Given the common misperception of drinking norms on college campuses, college newspapers might consider reporting statistics on the actual occurrence of binge drinking on campus to help correct students’ erroneous Perceptions.

A final approach to influencing social reality Perceptions on health issues is to institute a form of media literacy training. In general, this type of training is aimed at children with the goal of helping them to better understand the content of the media and to increase skepticism of many of the images and inferences in media content. There is some evidence that this type of literacy training is successful in limiting the influence of television viewing on Perceptions of alcohol norms (Austin & Johnson, 1997).

In Third World countries, development messages have been included in soap operas to introduce or change social norms about a number of issues (e.g., Kottak, 1990; Singhal & Rogers, 1989). For example, if campaigners hope to reduce the infant mortality rate in a developing country, they might develop a drama or soap opera focusing on people with whom the target audience could easily identify. Characters in the soap opera would explicitly and frequently engage in the appropriate behaviors (e.g., use the proper medicine when babies had diarrhea) and be positively rewarded for doing so. Those who did not do so would receive negative social feedback or experience negative consequences. Over time, these portrayals could cultivate the Perception that the health behaviors are normative and thus should be followed as standard practice because they must be appropriate and effective. If nothing else, individuals may simply engage in these behaviors to avoid social isolation. This same logic could be applied to the use of new agricultural methods or other innovative practices.

Racial and Sexual Stereotypes

In addition to changing behaviors, media content can be developed to produce counter-stereotypical beliefs and engender tolerance toward minority groups. Television news organizations can change their patterns of displaying minority criminal suspects in handcuffs. Newspapers can refrain from mentioning the race of suspects, especially because race is rarely a relevant factor in crime news stories. Entertainment shows can portray minority characters in prominent, powerful, and prosocial roles (e.g., as doctors and lawyers) and can limit the frequency with which they portray minorities stereotypically.

There is evidence that change in the typical demographics of television is possible. A number of minority groups expressed outrage at the small number of minority characters being presented in major roles in prime-time television during the Fall 1999 season. A “brownout,” in which minorities would boycott network television to demonstrate their displeasure as well as their importance, was threatened. In response, producers promised to add more minority characters (Weinraub, 1999).

Changes in media content to reflect greater diversity of jobs and activities among women can help to reduce stereotypical beliefs of young children regarding sex roles. For example, one episode of the Freestyle program on PBS (LaRose, 1989) showed a young female character engaging in counterstereotypical behavior (playing football) as part of a campaign to reduce sex role stereotyping among children. More generally, female characters on television can be shown to be just as important—and capable of the same feats—as male characters instead of being portrayed as simply victims, “prizes,” or bystanders. When these nontraditional gender roles are presented on television, there is some evidence that they may lead to a reduction in children’s stereotypical Perceptions about gender roles (e.g., Corder-Bolz, 1980; Rosenwasser et al., 1989).

Political Campaigns

The impact of Perceptions of public opinion on voter attitudes and behavior, especially during primary campaigns, is an important area of research in political science (e.g., Bartels, 1988; Ceci & Kain, 1982; Joslyn, 1997; Mehrabian, 1998; Mutz, 1998; Nadeau, Niemi, & Amato, 1994; West, 1991) and a potential avenue for Persuasive impact. The Persuasive press inference suggests that individuals will infer public opinion from the slant of news media coverage. And political communication research has often found either that voters tend to side with the Perceived winner (bandwagon effects) or that they give renewed effort for a candidate Perceived to be losing ground (underdog effects). In addition, the spiral of silence (Noelle-Neumann, 1993) suggests that Perceiving public opinion to be incongruent with one’s own opinion can further shape public opinion indirectly by leading members of the minority to refrain from expressing their opinions publicly. Equally important is the potential impact of Perceived public support for presidents in their dealings with legislators and other politicians. As Kernell (1986) noted, “A president with a strong reputation does better in his dealings largely because others expect fewer concessions from him…. Saddled with a weak reputation, conversely, a president must work harder, because others expect him to bargain less effectively” (p. 146). Reputation, of course, is enhanced by the Perception that the politician has strong public support.

Combining these theories and findings has important implications that have not been lost on political campaign managers. By attempting to alter Perceptions of candidate support via “spin” that is then conveyed in news media coverage, candidates can potentially increase their standings in the polls as well as their chances of receiving monetary contributions and volunteer efforts (Henshel & Johnston, 1987). Although candidates can provide bogus poll findings to the news media, poll findings are not the only means of influencing news media coverage. “Perhaps more than the polling results, it is the interpretation of the results that informs and persuades” (DeRosa & Tedesco, 1999, p. 67). Therefore, candidate representatives do their best to portray their candidates as being successful and influencing public opinion by, for instance, criticizing polls when their candidates are trailing (DeRosa & Tedesco, 1999), trumpeting the poll findings when their candidates are leading, and claiming that their candidates “won” debates (Lemert, Wanta, & Lee, 1999). Significantly, debate verdicts have been repeatedly shown to have a significant impact on Perceptions of which candidates won the debates as well as candidate image, above and beyond the impact of debate exposure itself (Lemert, 1996; Lemert, Elliott, Bernstein, Rosenberg, & Nestvold, 1991). Unfortunately for campaign workers, however, their ability to influence debate verdicts has been reduced during recent years because they are less likely to be invited on debate specials and newscasts (Lemert et al., 1999).

Conclusions

It has often been said that “Perception is reality.” Our Perceptions of the world and people around us, and not their “real” referents, are what drive our opinions, underlying attitudes, and everyday behavior. As such, Perceptions should be, and are, important variables in many theories of Persuasion.

Many forms of mass media are designed to shape our Perceptions of the world. Television advertisements tell us that we will be ostracized if we smell bad, so we should buy Right Guard deodorant. We are led to believe that we can be a hip member of the Gen-X culture if we just “do the Dew” by drinking Mountain Dew. Many advertisements encourage us to join the majority by using a particular product. However, these are not the only messages that can have a Persuasive impact on our beliefs about the world and people around us. Entertainment television, although not typically an intentionally persuasive form of content, has been demonstrated to have at least some impact on our Perceptions of social reality. News media content, while generally not intended by its producers to persuade us, is intended to inform us about the world around us. When the information we derive from the news media is accurate, this is considered a prosocial learning outcome. However, often individuals use news media content to infer facts about the world, and often these inferences are inaccurate. So, while the intention to Persuade is not commonly associated with news and entertainment media, they often have a Persuasive impact nonetheless.

The study of misperceptions of social reality has a long history in a number of related fields. However, for most of this history, the research on social reality Perceptions has proceeded without important integration across theoretical Perspectives or disciplinary boundaries. It is my hope that this review of the research will assist researchers in studying Persuasion to integrate and advance both Persuasion theory and the study of how entertainment and news media, in conjunction with motivational and cognitive processes, can influence our Perceptions—and, more important, our misperceptions—of social reality.