John Zaller’s “The Nature and Origins of Mass Opinion” is a fantastic addition to the literature on public opinion and the science of examining the motives for people’s views on public policy. When reading Zaller’s book, it’s easy to get tunnel vision and subscribe entirely to his views. While his calculations do provide great evidence for his model (although he does use some rather elaborate statistical gymnastics), his model should nonetheless be taken with a grain of salt since other research provides evidence for other factors at play in public opinion formation.
Zaller makes a couple of overarching points. First off, he re-affirms that communications from elites do have an impact on public opinion. This finding further distances audience studies away from Lazarsfeld’s two-step opinion flow model, a trend that started in the 60s. But, he also found that political awareness determined the impact of elite communication. In effect, more politically aware people tended to get their cues from partisan political elites than people who are basically inattentive.
Secondly, Zaller challenges the basic assumption that voters actually have true preferences on public policies. He presents a model in which individuals have conflicting views on specific issues and “winning” opinion really has more to do with what’s on your mind when being interviewed by a pollster. This revelation should create a pause for students of mass political opinion. Many other models do contain a biased assumption – that everyone thinks about politics as much as the average communications researcher. Zaller’s brings points out what should be the starting point for most public opinion research: Most people just don’t care that much about public policy.
Zaller creates the “Receive-Accept-Sample” model and then tests it using public opinion data from the Vietnam War. The model is broken down into four parts.
The Reception Axiom (A1) posits that “the greater a person’s level of cognitive engagement with an issue, the more likely he or she is to be exposed to and comprehend – in a word, to receive – political messages concerning that issue” (p. 42). In other words, more politically aware individuals will receive political communication better. Zaller notes that the axiom “indicates nothing about the sources of the political communications that shape mass opinion” (42). He therefore avoids directly examining some interesting questions about the role of selective exposure and personal deliberation on public opinion. But, the model still works since all he’s saying is that more attentive people receive communication better – regardless of the source.
The Resistance Axiom (A2) notes that “people tend to resist arguments that are incompatible with their political predispositions, but they do so only to the extent that they possess contextual information necessary to perceive a relationship between the message and their predispositions.” So, prior dispositions have a huge impact on whether new information actually produces a change in opinion. But people without prior dispositions (i.e., the uninformed) will tend uncritically accept any political communication. Zaller backs up this model with “a large body of theory and research concerning political persuasion” (p. 45). His thinking lines up with Converse (1964) who found that informed consumers of political communication tend to “constrain” their views depending on the cues of political elites.
The Accessibility Axiom (A3) states that “the more recently a consideration has been called to mind or thought about, the less time it takes to retrieve that consideration or related considerations from memory and bring them to the top of the head for use” (p. 48). In other words, the mind is like a bucket – what’s on top usually comes out first. So, this axiom also sends shivers down the hopeful students of political communication. Surely, we’re more complex than that? But, Zaller backs up this axiom with a wide body of psychological literature.
The Response Axiom (A4) finds that “individuals answer survey questions by averaging across the considerations that are immediately salient or accessible to them.” This axiom augments A3 in that things on the top of one’s head are probably the one’s that will affect an individual’s decision-making process. The axiom allows “different people to respond to issue questions on the basis of different considerations – one, for example, emphasizing ideological concerns, another gut-level likes and dislikes, and yet another self-interest” (p. 49). In this sense, Zaller’s model is far more nuanced than standard S-R models, which assume that everyone reacts to messages in the same way. (An interesting test of A3 and A4 would be to ask about views on euthanasia and correlating for viewing the film “Million Dollar Baby.” Would views change based on how recently they’ve seen the film?)
After explaining his model, Zaller uses the rest of the book to support it with statistical evidence. In Chapter 6, he finds that when elite discourse favors as given public policy in a one-sided manner, it produces a “mainstream pattern.” The most aware members of the public then subscribe to the consensus most strongly. When elite discourse divides along partisan lines, a polarization of the public inevitably follows. In Chapters 7 and 8, Zaller shows that changes in the relative intensity of opposing messages can also affect attitude change (with more intense messages having greater affect.)
I’ll focus the remainder of this paper on Chapter 9, “Two-sided information flows,” because I see some interesting parallels to both our research on global warming attitudes and the credibility of anonymous sources. In this chapter, Zaller examines how “public opinion can be understood as a response to the relative intensity and stability of opposing flows of liberal and conservative communications” (p. 186).
Zaller uses attitudes toward the Vietnam War as a foil to test his model. He picked the subject because of available data and the war’s shift in public support. In order gauge the coverage of the war, Zaller instructed a research assistant to undergo a content analysis of newspapers and newsmagazines of the era. He had the RA code each article as either “prowar” or “antiwar.” As a veteran content analyzer , I must take umbrage with his methods. Zaller doesn’t mention intercoder reliability, probably because he never tested for it – and I doubt it would have been extremely high. Such a rough estimation sounds more like a guess that an analysis. However, the rough analysis probably gauged the rough outline of Vietnam War coverage over the years acceptably. So, perhaps I shouldn’t quibble. But, it appears we often draw the line on rigorous research methods at different places. On another note, Zaller’s method to obtain a hawk-dove data set is brilliant. (He took data from 1988 survey that asked the question, then applied the results to the 1966 data using attitudes and opinions as an interloper.)
Zaller discovered that the pro- and anti-war message became more intense between 1964 and 1966. Initially, the pro-war message was more intense than the anti-war message, but that advantage disappeared by 1970. The RAS model predicts that the attitude of politically attuned liberals would shift first between 1964 and 1966 because they would be the only one paying attention to the anti-war messages. He found exactly that. Zaller summarizes: “Public attitudes on major issues change in response to changes in the relative intensities of competing streams of political communications, as filtered through the reception-acceptance process” (p. 190). He then goes on to show – via complicated statistical models – that people respond to survey questions on the basis of “whatever considerations are present in their minds and immediately accessible memory” (p. 190). In essence, he finds that message intensity affects people who are inattentive, supporting that part of the RAS model. Some of his data analysis didn’t produce statistical precision. He explains this away with various reasons, but I found myself again wondering about the moving yardstick of research methodology.
After reviewing his statistical results, Zaller sums up by arguing that his theory improves upon Converse in one way. Converse argues that ideology operates as a “constraint.” Zaller points out that if liberals were truly constrained by ideology, then they would’ve remained in support of the war – “as another case of post-World War II liberal internationalism” (p. 209). Since liberals changed their views on the war, Zaller concludes that they must’ve looked toward elite cues, rather than rely upon ideological constraints. The Vietnam War is a rare opportunity, he concludes, to watch a shift in the evolution of mass ideologies.
Exposure to global warming is probably equally restrained by ideology. Perhaps the conservatives holding out skepticism are listening more attentively to the anti-global warming message in a similar fashion. But, unlike the Vietnam War study, the mainstream sources are almost completely ignoring the skeptical elites. So, where are they getting their political communication? Somebody ought to do a study…
Tuesday, February 3, 2009
"Public Opinion" -- Glynn, Herbst, O’Keefe, and Shapiro
Chapter Four of “Public Opinion” by Glynn, Herbst, O’Keefe, and Shapiro offers a great summation of psychological perspectives on the formation of public opinion. The chapter helpfully starts with an overview of the common definitions, including an explanation of the terms “beliefs,” “values,” “attitudes,” and “opinions.”
Beliefs, we are told, are the “cognitive components that make up our understanding of the way things are, that is, the information that individuals have about objects or actions.” They are the building blocks of opinions and attitudes. Often they are widely shared by an entire culture, so beliefs are tough to identify. For instance, Western culture embraces the belief that people are perfectible, that they can improve over time. Individuals in this culture, therefore, find it hard to understand the Indian caste system, in which members cannot rise above their current social status. When shaping public opinion, beliefs must be categorized as the most difficult attribute to influence. Even as an “objective academic,” I struggle to retain neutrality with certain beliefs. For instance, I believe strongly that free speech should always be favored in any battle against speech restriction. While I can defend that belief with evidence (e.g., marketplace of ideas, Milton’s “Areopagitica”) others can defend their opposing beliefs just as easily (e.g., harmony of the state, Hobbes’ “Leviathan.”) Beliefs are simply a representation of that which we assume to be true – a faith, for lack of a better word. Another trait of beliefs (not in the reading) is that people seldom admit that their beliefs are actually beliefs. Instead, they’ll argue that their beliefs are facts – the only possible interpretation of the evidence as laid out. By the way, many academic theorists – while professing to be full of free-thinking, evidence-based argumentation – suffer the worst from the assumption that they’ve found Truth by examining all the facts. Of course, that’s just my belief. (As another aside, I don’t think there’s a difference between a belief and “faith” – although the latter is often a dirty word.)
The authors go on to talk about values, described as representations “of the way things should be.” Values, then, are the goals one strives to obtain in whatever belief-system one operates. The authors then differentiate between two types of values. Terminal values are goals we want to reach (e.g., freedom, equality, peace.) Instrumental values are the means we endorse to reach our goals (e.g., honesty, responsibility, and loyalty.) Values are often, like beliefs, taken for granted.
Attitudes, we are told, “are general and enduring positive or negative feelings about some person, object or issue.” Attitudes have three components: cognitions (or beliefs), affect (or feelings) and behavior. The authors note that attitudes differ from beliefs because they have a strong emotional component. Of course, beliefs carry strong emotional components as well – try explain to a libertarian why rent-control is such a good idea. Indeed, finding the difference between beliefs and attitudes seems fraught with difficulty. The book explains that beliefs are evaluative in nature, but that you “perceive these evaluations as inherent characteristics of the objects of your beliefs.” But, attitudes reflect personal likes and dislikes, “you recognize them as your own orientation toward objects rather than as characteristics that are somehow part of the objects themselves.” Ahh! This makes sense. As I said earlier, beliefs tend to be viewed internally as incontrovertible conclusions brought out by an honest assessment of the facts. But, attitudes people are more likely to understand as simply “their way of looking at things.” This distinction would seem to be an incredibly important point for anyone interested in swaying public opinion. Since attitudes are cognitively conceived as a “personal preference,” they should be more malleable than beliefs, which are representations of Truth. Anyone interested in swaying public opinion, therefore, should first frame the issue as a matter of attitude, not a belief. In fact, views toward homosexuality (at least in the United States) have followed this model. Forty years ago most people shared the belief that homosexuality was wrong. Over the years, people’s feelings about the issue (well, at least some people’s) have softened with most just holding a residual attitude about it – nothing as solid as a belief. Many other social issues have followed a similar path (e.g., civil rights, smoking, corporal punishment, etc.)
That leaves us with opinions. The authors define these as the expression of an attitude “either verbally or through behavior.” The distinction between attitudes and opinions is that attitudes may never be expressed as opinions – “they are deeper inside a person than opinions are.” Attitudes are unwavering, but opinions can change along with circumstances. I love chocolate, but I may hate it after eating too much chocolate cake. Often attitudes and opinions jibe, but sometimes they do not. The book notes some people may believe members of Congress are disreputable but put a lot of faith in their own lawmaker. This tendency is an interesting example and one that I discovered in my lit review of media credibility. Many respondents distrusted the media in general, but gave high marks to their preferred media outlet.
The authors then discuss theories surrounding attitude formation. The earliest research found evidence of conditioning. In classical conditioning, people formed attitudes or beliefs, much like animals, based on cues from their surroundings. In operant conditioning, people act to maximize the positive and minimize the negative consequences of their behavior. For instance, a researcher found he could shape vocal patterns by merely offering positive or negative cues. Although later problems with the behavioral approach have arisen, the authors stress its most important contribution: That people do not think very much about the opinions they express. Probably the understatement of the year.
The authors do list at least five problems with the behavioral approach. The model doesn’t account for why some opinions are more thought-out than others. The model doesn’t account for a “package” of attitudes and opinions, and its impact on belief formation. The model doesn’t take into account that two people may hold the same belief for two different reasons. (Polls of Obama voters should support this tendency well.) Behavioral theories can’t explain why people don’t always act in concordance with their attitudes. Also, they don’t factor in emotions, which play an important role in formulation of public opinion.
So, what theories work better? Well, we have a bunch of choices. Consistency theories basically focus on the differences between different viewpoints (i.e., attitudes, opinions beliefs, etc.) These theories include balance theory (people want to create a balance in their thinking), congruity theory (weighing things with shades of gray) and cognitive dissonance theory (people tune out information that don’t correspond to their own beliefs.)
Judgmental theories argue that “our past experiences play an important role in the ways that we interpret new information.” These theories go beyond the simple Stimulus-Response models and look for more interactions that influence opinion formation. These include social judgment theory and adaption theory which both hold that individual situations play a part in shaping one’s views.
The cognitive mediation theory that we’re using in our global warming paper fits into the judgmental theory category. Grounded in the O-S-O-R model, it looks at Stimulus (S) and Response (R), but also examines those two O’s. The first is structural, cultural, cognitive and motivational characteristics of the audience (i.e., individual baggage). The second O exams what is likely to happen between the reception of the message and the response of the audience member. The cognitive model suggests that motivations do not have a direct role on media effects. Rather, the model predicts that motivations activate information processing behaviors that then lead to cognitive media effects. In sum, “the cognitive mediating model argues that information processing mediates the influence of gratifications sought on media effects.” If I understand that correctly, the authors are stating that prior attitudes will merely influence how we think about new information presented to us – not directly affect our decisions.
Beliefs, we are told, are the “cognitive components that make up our understanding of the way things are, that is, the information that individuals have about objects or actions.” They are the building blocks of opinions and attitudes. Often they are widely shared by an entire culture, so beliefs are tough to identify. For instance, Western culture embraces the belief that people are perfectible, that they can improve over time. Individuals in this culture, therefore, find it hard to understand the Indian caste system, in which members cannot rise above their current social status. When shaping public opinion, beliefs must be categorized as the most difficult attribute to influence. Even as an “objective academic,” I struggle to retain neutrality with certain beliefs. For instance, I believe strongly that free speech should always be favored in any battle against speech restriction. While I can defend that belief with evidence (e.g., marketplace of ideas, Milton’s “Areopagitica”) others can defend their opposing beliefs just as easily (e.g., harmony of the state, Hobbes’ “Leviathan.”) Beliefs are simply a representation of that which we assume to be true – a faith, for lack of a better word. Another trait of beliefs (not in the reading) is that people seldom admit that their beliefs are actually beliefs. Instead, they’ll argue that their beliefs are facts – the only possible interpretation of the evidence as laid out. By the way, many academic theorists – while professing to be full of free-thinking, evidence-based argumentation – suffer the worst from the assumption that they’ve found Truth by examining all the facts. Of course, that’s just my belief. (As another aside, I don’t think there’s a difference between a belief and “faith” – although the latter is often a dirty word.)
The authors go on to talk about values, described as representations “of the way things should be.” Values, then, are the goals one strives to obtain in whatever belief-system one operates. The authors then differentiate between two types of values. Terminal values are goals we want to reach (e.g., freedom, equality, peace.) Instrumental values are the means we endorse to reach our goals (e.g., honesty, responsibility, and loyalty.) Values are often, like beliefs, taken for granted.
Attitudes, we are told, “are general and enduring positive or negative feelings about some person, object or issue.” Attitudes have three components: cognitions (or beliefs), affect (or feelings) and behavior. The authors note that attitudes differ from beliefs because they have a strong emotional component. Of course, beliefs carry strong emotional components as well – try explain to a libertarian why rent-control is such a good idea. Indeed, finding the difference between beliefs and attitudes seems fraught with difficulty. The book explains that beliefs are evaluative in nature, but that you “perceive these evaluations as inherent characteristics of the objects of your beliefs.” But, attitudes reflect personal likes and dislikes, “you recognize them as your own orientation toward objects rather than as characteristics that are somehow part of the objects themselves.” Ahh! This makes sense. As I said earlier, beliefs tend to be viewed internally as incontrovertible conclusions brought out by an honest assessment of the facts. But, attitudes people are more likely to understand as simply “their way of looking at things.” This distinction would seem to be an incredibly important point for anyone interested in swaying public opinion. Since attitudes are cognitively conceived as a “personal preference,” they should be more malleable than beliefs, which are representations of Truth. Anyone interested in swaying public opinion, therefore, should first frame the issue as a matter of attitude, not a belief. In fact, views toward homosexuality (at least in the United States) have followed this model. Forty years ago most people shared the belief that homosexuality was wrong. Over the years, people’s feelings about the issue (well, at least some people’s) have softened with most just holding a residual attitude about it – nothing as solid as a belief. Many other social issues have followed a similar path (e.g., civil rights, smoking, corporal punishment, etc.)
That leaves us with opinions. The authors define these as the expression of an attitude “either verbally or through behavior.” The distinction between attitudes and opinions is that attitudes may never be expressed as opinions – “they are deeper inside a person than opinions are.” Attitudes are unwavering, but opinions can change along with circumstances. I love chocolate, but I may hate it after eating too much chocolate cake. Often attitudes and opinions jibe, but sometimes they do not. The book notes some people may believe members of Congress are disreputable but put a lot of faith in their own lawmaker. This tendency is an interesting example and one that I discovered in my lit review of media credibility. Many respondents distrusted the media in general, but gave high marks to their preferred media outlet.
The authors then discuss theories surrounding attitude formation. The earliest research found evidence of conditioning. In classical conditioning, people formed attitudes or beliefs, much like animals, based on cues from their surroundings. In operant conditioning, people act to maximize the positive and minimize the negative consequences of their behavior. For instance, a researcher found he could shape vocal patterns by merely offering positive or negative cues. Although later problems with the behavioral approach have arisen, the authors stress its most important contribution: That people do not think very much about the opinions they express. Probably the understatement of the year.
The authors do list at least five problems with the behavioral approach. The model doesn’t account for why some opinions are more thought-out than others. The model doesn’t account for a “package” of attitudes and opinions, and its impact on belief formation. The model doesn’t take into account that two people may hold the same belief for two different reasons. (Polls of Obama voters should support this tendency well.) Behavioral theories can’t explain why people don’t always act in concordance with their attitudes. Also, they don’t factor in emotions, which play an important role in formulation of public opinion.
So, what theories work better? Well, we have a bunch of choices. Consistency theories basically focus on the differences between different viewpoints (i.e., attitudes, opinions beliefs, etc.) These theories include balance theory (people want to create a balance in their thinking), congruity theory (weighing things with shades of gray) and cognitive dissonance theory (people tune out information that don’t correspond to their own beliefs.)
Judgmental theories argue that “our past experiences play an important role in the ways that we interpret new information.” These theories go beyond the simple Stimulus-Response models and look for more interactions that influence opinion formation. These include social judgment theory and adaption theory which both hold that individual situations play a part in shaping one’s views.
The cognitive mediation theory that we’re using in our global warming paper fits into the judgmental theory category. Grounded in the O-S-O-R model, it looks at Stimulus (S) and Response (R), but also examines those two O’s. The first is structural, cultural, cognitive and motivational characteristics of the audience (i.e., individual baggage). The second O exams what is likely to happen between the reception of the message and the response of the audience member. The cognitive model suggests that motivations do not have a direct role on media effects. Rather, the model predicts that motivations activate information processing behaviors that then lead to cognitive media effects. In sum, “the cognitive mediating model argues that information processing mediates the influence of gratifications sought on media effects.” If I understand that correctly, the authors are stating that prior attitudes will merely influence how we think about new information presented to us – not directly affect our decisions.
Subscribe to:
Posts (Atom)
