From Single to Multiple to Metacognitive Processes

This article provides a brief overview of major developments in the history of contemporary persuasion theory. The first intuitive and empirical approaches to persuasion were guided by main-effect questions (e.g., are experts more persuasive than nonexperts?). Furthermore, researchers focused on only one process by which variables (e.g., emotion, source credibility) would have an impact (e.g., emotion affected attitudes by classical conditioning). As data began to accumulate, so many new theories and effects were uncovered that the discipline faced collapse from the numerous inconsistencies evident. In response to the reigning confusion of the previous era, contemporary multiprocess theories were proposed (e.g., the elaboration likelihood model). According to these more integrative approaches, any one variable could affect attitudes by different processes in different situations and thereby sometimes produce opposite effects. Finally, we describe the role of a recently discovered new contributor to persuasion: self-validation. Unlike previous mechanisms that focus on primary cognition, this new process emphasizes secondary or meta-cognition. Persuasion is everywhere, playing an essential role in politics, religion, psychotherapy, education, and day-to-day social interactions. Given that people attempt to influence others and are also targets of influence, they have learned something about how persuasion works thorough trial and error. In contrast to this intuitive persuasion knowledge and the advice available from many popular books on the subject, scholars in disciplines such as psychology, communications, political science, marketing, and advertising have systematically studied persuasion for many years. In this article, we review a contemporary social psychological perspective on persuasion with an emphasis on explicating the psychological processes that account for how variables such as credible sources, a person’s emotions, and others produce attitude change. In describing the basic mechanisms underlying persuasion, we will provide a brief overview of social psychology’s historical contribution to this area of research, describe the evolution from main effect and single processes approaches to contemporary multiprocess and system theories, outline a general framework that articulates the key processes of persuasion, and highlight a recently discovered new mechanism of persuasion—called self-validation—that ties together the operation of a diverse set of variables. We use the term persuasion quite broadly to refer to any procedure with the potential to change someone’s mind. Although persuasion can be used to change many things such as a person’s specific beliefs (e.g., that wine is good for one’s health), the most common target of persuasion in the psychological literature is a person’s attitudes. Attitudes refer to general evaluations people have regarding other people, places, objects, and issues. Attitudes are studied as the primary object of influence because of their presumed guiding influence on choice and action. That is, all things being equal, people will decide to buy the product they like the most, attend the university they evaluate most favorably, and vote for the candidate they approve of most strongly. EARLIEST IDEAS ABOUT PERSUASION Scholarly speculation about persuasion has a long history and has gone through a number of distinct eras (see Petty, 1997). As with many phenomena, the first phase involved asking simple main effect questions about single variables. This approach is evident beginning with the ancient Greeks (e.g., Aristotle’s Rhetoric) and continuing through the Oratoria from the Italian Renaissance (see McGuire, 1969). These early approaches to persuasion attempted to answer several questions. Are experts more persuasive than nonexperts? Is it better to present people with logical arguments or with appeals based on emotion? Is fear a good emotional tool or is it counterproductive? Humans have a longstanding curiosity about such questions, and contemporary scholars continue to study these issues as well. Modern-day behavioral experiments on persuasion were initiated in the early 1900s (see Murphy, Murphy, & Newcomb, 1937). These initial empirical approaches were often guided by Address correspondence to Richard E. Petty, Department of Psychology, Ohio State University, 1835 Neil Avenue, Columbus, OH 43210; e-mail: [email protected]. PERSPECTIVES ON PSYCHOLOGICAL SCIENCE Volume 3—Number 2 137 Copyright r 2008 Association for Psychological Science the same main-effect type questions that inspired the early philosophers. According to the main-effect view, any one variable (e.g., an expert source, a happy emotional state) was likely to have just one effect on persuasion—either enhancing or reducing it. Early theories of persuasion also suggested that there was likely only one mechanism or process responsible for whichever outcome was produced. For example, a happy emotion might increase persuasion because of classical conditioning. One of the earliest and influential general theories of persuasion in the modern era was based on learning theory principles (Hovland, Janis, & Kelley, 1953). Briefly described, this ‘‘Yale’’ approach held that anything that facilitated attending to, comprehending, and learning the contents of a persuasive message would be good for attitude change and that anything that would disrupt those learning processes would be bad. For example, distracting someone from the message was predicted to reduce persuasion because it would interfere with learning the message. Similarly, according to this theory, providing a person with a credible source would be good because it would motivate people to learn the message. Although the single effect and single process assumptions provided a reasonable beginning to the field, it was not long before complications arose. First, any one variable (e.g., an expert source, fear) was shown to be good for persuasion in some studies but was found to be detrimental in others. Also, there was no compelling support for a single mechanism by which persuasion worked. Finally, researchers have struggled for decades to determine why attitude changes sometimes seemed to be relatively durable and impactful (e.g., guiding behavior) but were rather transitory and inconsequential at other times (e.g., Fazio, 1986). Thus, theories evolved to account for multiple effects, processes, and consequences. DUALITY IN EARLY THEORIES OF PERSUASION As studies on persuasion accumulated and the single effect, process, and consequence assumptions were challenged, theories of influence became more complex. Researchers either modified the early theories or developed new ones to account for the emerging data. One recurring theme involved an underlying duality in persuasion processes. The idea that there are two fundamental types of persuasion can be traced at least to Aristotle, who highlighted a distinction between persuasion involving emotion (passion) versus persuasion involving reason. Furthermore, the notion of an underlying duality in judgment and behavior (e.g., acting or deciding based on one’s first impulse versus a more deliberative consideration) is a recurrent theme in psychology since Freud (1923/1962; see Carver, 2005, for a review). So it is not surprising that a duality emerged in persuasion theory as well. Perhaps the most important initial example of this duality in persuasion theory was in the Hovland group’s eventual distinction between persuasion based on learning simple augmenting or discounting cues versus persuasion based on learning the message arguments (e.g., Kelman & Hovland, 1953). The key idea was that, separate from the impact of learning the substantive arguments in a persuasive message (the initial focus of their theory), various simple cues (such as high or low credibility sources) could independently augment (or discount) the amount of influence that took place based on the message alone. These orthogonal cue and argument-learning effects were thought to operate simultaneously. That is, people could learn to associate both simple cues and complex arguments with a message conclusion. Furthermore, the impact of each on acceptance of the message conclusion was unique, and because each type of learning was independent, each type of learning had its own forgetting curve. The duality in this theory was primarily one of content: cues versus arguments. The same fundamental process (learning) operated on each content, though learning simple cues would presumably require less cognitive effort than would learning complex message arguments. Nevertheless, separating cue learning from argument learning and making them both independent contributors to persuasion has allowed the theory to explain some novel persuasion phenomena such as how variables could affect persuasion in the absence of affecting message learning (which was not possible in the original theory) and how initial resistance to a message could change over time into acceptance, such as when people forget a negative message source faster than they forget favorable message arguments (the sleeper effect; see Weiss, 1953). In another influential early framework, Kelman (1958) introduced a process distinction that was tied to particular content. Specifically, Kelman distinguished between two kinds of persuasion: internalization (acceptance of the message arguments) versus identification (agreeing because one likes the message source). In Kelman’s framework, certain variables (e.g., high source expertise) induced agreement because they enhanced acceptance of the message arguments, whereas other variables (e.g., high source attractiveness), when paired with the same message, induced agreement because of identification with the message source. In both cases, the attitude change was real; in the former case, the change would persist in the absence of the source, whereas in the latter case, attitude change depended on one’s continued liking of the source. The distinction that Kelman (1958) introduced had some parallels to Hovland and colleagues’ earlier cue versus arguments distinction; however, rather than being a content distinction to which the same process (learning) was applied, Kelman argued that there were two different processes (accepting arguments, liking sources) tied to different sets of content. That is, some variables (source expertise) were associated Kelman (1958) also distinguished compliance as a form of influence in which there is no internal change and the individual is merely going along with a powerful other. 138 Volume 3—Number 2 Persuasion Processes