Ambiguity, Distorted Messages, and Nested Environmental Effects on Political Communication

In this paper we are concerned with the clarity of political signals transmitted through political conversation and the accuracy with which those signals are perceived. The social communication of political information is subject to distortion effects that arise due to skewed expectations on the part of the receiver and ambiguous representations on the part of the sender. Indeed, communication that occurs between two citizens might be distorted either by characteristics of the individuals who are transmitting and receiving messages, or by characteristics of the setting in which the information is being transmitted. We argue that the power of majority opinion is magnified by the inferential devices that citizens use to reach judgments in the face of ambiguous political messages and hence the use of a personal experience heuristic gives rise to a political bias that favors the continued dominance of majority opinion. ow important is political communication among citizens? Political discussion is an efficient vehicle for becoming informed about politics (Downs 1957); it is a widespread activity with influential consequences (Berelson, Lazarsfeld, and McPhee 1954; Huckfeldt and Sprague 1995; Lazarsfeld, Berelson, and Gaudet 1944); and it may be fundamentally important to the vitality of democratic politics. At the same time, relatively few citizens demonstrate highly intense levels of political interest and engagement (Verba, Schlozman, and Brady 1995); politics is only one among many important topics competing for airtime during citizens' conversations (Huckfeldt and Sprague 1995); few of the com Nested Environmental Effects 997 municated opinions are likely to be carefully researched (Sniderman, Brody, and Tetlock 1991); and many of the political messages delivered through social communication consist of offhand comments and the very briefest expressions of sentiment. For all these reasons, a great deal of political discussion and communication is likely to be imbedded in ambiguity and uncertainty. And the uncertainty of ambiguous communication gives rise to distortions based on the characteristics of the person sending the message, the characteristics of the person receiving the message, and the various environments and settings within which the message is interpreted. In this paper we are concerned with the clarity of political signals transmitted through political conversation and the accuracy with which those signals are perceived. The potential for distorted communication arises due to individual and environmental effects operating at several nested and overlapping levels. First, social communication regarding politics is subject to distortion effects that arise both due to skewed expectations on the part of the receiver and ambiguous representations on the part of the sender. Second, political discussion often occurs at the discretion of individual citizens, within the closely defined social environments where they are located, but the distribution of political opinions within these microenvironments depends on the supply of viewpoints available in the larger political community, thereby truncating the distribution of political preferences to which citizens are regularly exposed (Huckfeldt et al. 1995; Huckfeldt and Sprague 1995). Third, the level of distortion in dyadic information flows is contingent not only on the characteristics of the individuals involved in the dyad, but also on expectations that are formed on the basis of individual experience. These experiences, in turn, depend on the immediate circumstances of a person's location within these microenvironments. Finally, in assessing political communication, individuals employ methods of inferential judgment that, while yielding relatively accurate judgments in the aggregate, might give rise to a bias fostering the continued dominance of majority opinion (Miller 1956). In short, our analytic framework examines political communication within a series of nested environmental levels: individuals within dyads, dyads within microenvironments, and microenvironments within larger environments of opinion. We begin by addressing a number of substantive issues: the importance of disagreement among citizens, the factors that give rise to ambiguity and distortion in political communication, and the sources of a majoritarian bias in political communication and cognition. A Bayesian logic of inference is then articulated for examining nested environmental effects on political communication. Finally, based on an empirical analysis of the 1992 election, we assess the levels of political disagreement experienced by a national sample of survey respondents and their discussants, the extent to which perceptions of disagreement are contingent on the externally imposed supply of preferences in the surrounding macroenvironment, the accuracy of political communication between main respondents and R. Huckfeldt, P Beck, R. Dalton, J Levine, and W Morgan their discussants, and sources of distortion in the perceptions of discussants' political preferences. Disagreement, Ambiguity, and the Effectiveness of Communication If political discussion is to play an important role in democratic politics, it must introduce new information and new ideas to citizens, thereby creating disagreement as the inevitable by-product of deliberation (Granovetter 1973; McPhee 1963). If people only talk politics with others holding compatible views-or if they wrongly perceive their social contacts to hold sympathetic views-they may be spared the social and political discomfort of disagreement, but the communication of diverse political preferences is rendered ineffective.' Several factors might be expected to affect the level of political diversity-and hence the exposure to disagreement-across the range of a citizen's social contacts. First, the combination of segregated social groups and polarized political preferences reduces the likelihood of encountering alternative political viewpoints (Huckfeldt and Kohfeld 1989). Second, the level of political homogeneity within a citizen's social space is increased to the extent that citizens employ politically relevant selection criteria in constructing their own patterns of social interaction, thereby locating themselves in politically agreeable microenvironments (Finifter 1974). Finally, to the extent that individuals are either unable or disinclined to censor their patterns of political communication, the distribution of political viewpoints in the larger environment takes on heightened importance as a factor that affects individual exposure to alternative political viewpoints (Huckfeldt et al. 1995; Huckfeldt and Sprague 1995). In short, the level of exposure to disagreement and diversity within patterns of political communication is problematic, both across groups and across individuals, and hence the vitality of democratic discourse is an open question. Just as important, while environmental supply and individual control affect the probabilities of exposure to alternative political viewpoints, exposure alone does not insure the effective communication of political disagreement. In order for politically diverse and disagreeable viewpoints to be communicated effectively, it is also necessary that they be perceived correctly. In this sense, effective communication depends on accurate perception. Of course, inaccurately perceived messages may also be influential, and for many purposes it is the perceived message that is most important for the choices that a citizen makes. But this does not alter the fact that incorrect perception has the potential to obscure and disguise diversity and disagreement at both the individual and aggregate levels. Several different factors serve to obscure rather than clarify the perception and expression of socially communicated political information. First, it may be help- 'None of this is meant to suggest that political disagreement is necessarily or inevitably an unpleasant experience for all individuals. Indeed, for some individuals, the expressive benefits of political discussion may even be increased as a consequence of disagreement (Fiorina 1990)  R. Huckfeldt, P Beck, R. Dalton, J Levine, and W Morgan their discussants, and sources of distortion in the perceptions of discussants' political preferences. Disagreement, Ambiguity, and the Effectiveness of Communication If political discussion is to play an important role in democratic politics, it must introduce new information and new ideas to citizens, thereby creating disagreement as the inevitable by-product of deliberation (Granovetter 1973; McPhee 1963). If people only talk politics with others holding compatible views-or if they wrongly perceive their social contacts to hold sympathetic views-they may be spared the social and political discomfort of disagreement, but the communication of diverse political preferences is rendered ineffective.' Several factors might be expected to affect the level of political diversity-and hence the exposure to disagreement-across the range of a citizen's social contacts. First, the combination of segregated social groups and polarized political preferences reduces the likelihood of encountering alternative political viewpoints (Huckfeldt and Kohfeld 1989). Second, the level of political homogeneity within a citizen's social space is increased to the extent that citizens employ politically relevant selection criteria in constructing their own patterns of social interaction, thereby locating themselves in politically agreeable microenvironments (Finifter 1974). Finally, to the extent that individuals are either unable or disinclined to censor their patterns of political communication, the distribution of political viewpoints in the larger environment takes on heightened importance as a factor that affects individual exposure to alternative political viewpoints (Huckfeldt et al. 1995; Huckfeldt and Sprague 1995). In short, the level of exposure to disagreement and diversity within patterns of political communication is problematic, both across groups and across individuals, and hence the vitality of democratic discourse is an open question. Just as important, while environmental supply and individual control affect the probabilities of exposure to alternative political viewpoints, exposure alone does not insure the effective communication of political disagreement. In order for politically diverse and disagreeable viewpoints to be communicated effectively, it is also necessary that they be perceived correctly. In this sense, effective communication depends on accurate perception. Of course, inaccurately perceived messages may also be influential, and for many purposes it is the perceived message that is most important for the choices that a citizen makes. But this does not alter the fact that incorrect perception has the potential to obscure and disguise diversity and disagreement at both the individual and aggregate levels. Several different factors serve to obscure rather than clarify the perception and expression of socially communicated political information. First, it may be help- 'None of this is meant to suggest that political disagreement is necessarily or inevitably an unpleasant experience for all individuals. Indeed, for some individuals, the expressive benefits of political discussion may even be increased as a consequence of disagreement (Fiorina 1990)R. Huckfeldt, P Beck, R. Dalton, J Levine, and W Morgan their discussants, and sources of distortion in the perceptions of discussants' political preferences. Disagreement, Ambiguity, and the Effectiveness of Communication If political discussion is to play an important role in democratic politics, it must introduce new information and new ideas to citizens, thereby creating disagreement as the inevitable by-product of deliberation (Granovetter 1973; McPhee 1963). If people only talk politics with others holding compatible views-or if they wrongly perceive their social contacts to hold sympathetic views-they may be spared the social and political discomfort of disagreement, but the communication of diverse political preferences is rendered ineffective.' Several factors might be expected to affect the level of political diversity-and hence the exposure to disagreement-across the range of a citizen's social contacts. First, the combination of segregated social groups and polarized political preferences reduces the likelihood of encountering alternative political viewpoints (Huckfeldt and Kohfeld 1989). Second, the level of political homogeneity within a citizen's social space is increased to the extent that citizens employ politically relevant selection criteria in constructing their own patterns of social interaction, thereby locating themselves in politically agreeable microenvironments (Finifter 1974). Finally, to the extent that individuals are either unable or disinclined to censor their patterns of political communication, the distribution of political viewpoints in the larger environment takes on heightened importance as a factor that affects individual exposure to alternative political viewpoints (Huckfeldt et al. 1995; Huckfeldt and Sprague 1995). In short, the level of exposure to disagreement and diversity within patterns of political communication is problematic, both across groups and across individuals, and hence the vitality of democratic discourse is an open question. Just as important, while environmental supply and individual control affect the probabilities of exposure to alternative political viewpoints, exposure alone does not insure the effective communication of political disagreement. In order for politically diverse and disagreeable viewpoints to be communicated effectively, it is also necessary that they be perceived correctly. In this sense, effective communication depends on accurate perception. Of course, inaccurately perceived messages may also be influential, and for many purposes it is the perceived message that is most important for the choices that a citizen makes. But this does not alter the fact that incorrect perception has the potential to obscure and disguise diversity and disagreement at both the individual and aggregate levels. Several different factors serve to obscure rather than clarify the perception and expression of socially communicated political information. First, it may be help- 'None of this is meant to suggest that political disagreement is necessarily or inevitably an unpleasant experience for all individuals. Indeed, for some individuals, the expressive benefits of political discussion may even be increased as a consequence of disagreement (Fiorina 1990)Nested Environmental Effects ful to understand the ambiguous communication of political messages from a strategic standpoint (MacKuen 1990). The costs of social communication sometimes increase when disagreement is present in a relationship: for instance, it is often painful to tell highly opinionated relatives that their preferred presidential candidate is a loser! In such situations, the cost can be reduced by obfuscation if not complete avoidance. The receiver obtains an obscure message because the sender has strategically chosen to engage in political retreat. Hence, ambiguity is increased, and the likelihood of distorted communication is heightened. Second, not all ambiguity is the residue of strategic interaction. A great deal of political conversation is casual and offhand, occurring among citizens whose opinions are only weakly formulated. Thus, at one extreme we might expect ambiguity to increase as a consequence of intensely held opinions that increase the potential for conflict and thereby produce strategic retreat. At the other extreme, we might also expect ambiguity to be generated by weak or nonexistent opinions that generate vague political signals (Fazio 1990; Latane 1981; Petty and Cacioppo 1986). In summary, for a number of systematic reasons, a great deal of political communication is inevitably ambiguous and open to alternative interpretations. When citizens are confronted with ambiguous political messages, the potential for distorted communication is enhanced, and several mechanisms of distortion are well known. In particular, selective perception makes it possible for individuals to avoid, disregard, and transform messages that do not agree with their own preconceptions and viewpoints. While earlier explanations for these patterns of selective perception were anchored in dissonance theories (Festinger 1957; Fiske and Taylor 1991), it is also possible to address these processes on the basis of alternative concepts in the newer arsenal of political cognition research (Ottati and Wyer 1990). For example, individuals who receive messages that do not correspond with preexisting conceptions may simply lack a cognitive structure to incorporate the information, thereby integrating it in an inappropriate manner (Lodge and Hamill 1986). In the present context, they might incorrectly identify the message being communicated through political discussion. Ambiguity, Context, and Cognition Particularly in the context of ambiguous political messages, people are also likely to employ contextually based cognitive shortcuts in evaluating socially communicated information. In their classic articulations of representativeness and availability as judgmental heuristics, Kahneman and Tversky (1973; Tversky and Kahneman 1973, 1974) demonstrate that individuals often make judgments that seem most representative of the evidence, guided by the availability of similar past experiences (see also Kinder 1978; Sniderman, Brody, and Tetlock 1991). In the present context, consider the following evaluations of a coworker's political preference. First: Joe is a good guy; he is a lot like me; I'm voting for Bill Clinton; Joe will probably vote for Clinton too. Alternatively: Joe is a good R. Huckfeldt, P Beck, R. Dalton, J Levine, and W Morgan guy; he is a lot like the other people at work; most of them are voting for Bill Clinton; Joe will probably vote for Clinton too. In either case, expectations regarding Joe's behavior are not seriously called into question by the (perhaps intentionally) obscure message that Joe is sending.2 Mechanisms such as these might be seen in the context of behavioral and hence imperfect implementations of a Bayesian logic with prior information updated by (biased) sample data. When citizens interpret an ambiguous signal sent though social communication, they might evaluate that signal in the context of prior information, but where does such information originate? One answer is that the prior information is based on support levels in the larger environment: people who live in Democratic counties, for example, might reasonably expect their associates to be Democrats. This prior information is combined with sample data, where the sample data are obtained through personal experience-personal experience that arises due to the citizen's own biased sample of social encounters accumulated during everyday routines. Bayesian logic may not come naturally to many citizens, however, and experimental results indicate that people rely heavily on their own vivid experience (the sample data), while they systematically undervalue the prior information-Kahneman and Tversky's (1973) base rate. Several questions quite naturally arise with respect to socially communicated political information: How important are such informal sample data in the evaluation of political communication? How important is the prior information? While cognitive shortcuts may be quite efficient and perform quite well overall, they might also give rise to distortions and communication failures. In terms of political communication between discussion partners, empirical demonstrations of misperception and systematic bias in political cognition are readily available. Citizens are certainly more likely to perceive discussants' preferences accurately if they share the discussants' preferences (Huckfeldt et al. 1995; Huckfeldt and Sprague 1995), but these empirical demonstrations show something else as well. The perception of a discussant's political preference depends on contextual opinion distributions-the perception is biased in the direction of environmental preference distributions quite independently of the discussant's true preference. How should we explain these environmental effects? One explanation focuses on the importance of the external macroenvironment: people who live in Republican or Democratic communities might be more likely to generalize on the basis of these external environments when making judgments regarding a particular political message received through social communication. Consider the 1992 campaign in the context of a traditionally 2A direct parallel is readily drawn between these personal experience heuristics, on the one hand, and the structural equivalence explanations invoked by network theorists, on the other (Burt 1987). In particular, the representativeness argument suggests that people base their social and political cognitions of other individuals on the extent to which these other individuals can be seen to represent particular social categories. According to structural equivalence, one very important form of social influence derives from the extent to which one individual sees another as occupying a similar location in social structure. In both instances, social location drives perception.