Framing as Media Effects Framing as a Theory of Media Effects by Dietram A. Scheufele Research on framing is characterized by theoretical and empirical vagueness. This is due, in part, to the lack of a commonly shared theoretical model underlying framing research. Conceptual problems translate into operational problems, limiting the comparability of instruments and results. In this paper I systematize the fragmented approaches to framing in political communication and integrate them into a comprehensive model. I classify previous approaches to framing research along two dimensions: the type of frame examined (media frames vs. audience frames) and the way frames are operationalized (independent variable or dependent variable). I develop a process model of framing, identifying four key processes that should be addressed in future research: frame building, frame setting, individuallevel processes of framing, and a feedback loop from audiences to journalists. Entman (1993) referred to framing as “a scattered conceptualization” (p.␣ 51), with previous studies lacking clear conceptual definitions and relying on context-specific, rather than generally applicable operationalizations. Brosius and Eps (1995) went even further, positing that framing is not a clearly explicated and generally applicable concept, but only a metaphor that cannot be directly translated into research questions. Partly because of these vague conceptualizations, the term framing has been used repeatedly to label similar but distinctly different approaches. For example, Wicks (1992) identified subtle but distinct differences between various concepts of cognitive categorization. Hamill and Lodge (1986) and Lodge & Hamill (1986) saw only a terminological difference between concepts like frame, script, or schema. At the same time, studies have operationalized framing in combination with other concepts such as agenda setting or priming (Iyengar & Kinder, 1987). More recently, McCombs, Shaw, and Weaver (1997) suggested that not only are agenda setting and framing effects related, framing is, in fact, an extension of agenda setting. They used the term second-level agenda-setting to describe the impact of the salience of characteristics of media coverage on audiences’ interpretation of these news stories. Perhaps as a result of these terminological and conceptual Dietram A. Scheufele (MA, University of Wisconsin-Madison, 1997) is a doctoral candidate at the Mass Communications Research Center, University of Wisconsin-Madison. His research interests include public opinion, media effects, and research methods. He thanks Patricia Moy, Jack M. McLeod, and Diana C. Mutz for their helpful comments. An earlier version was presented at the convention of the Association for Education in Journalism and Mass Communication, Anaheim, CA, August 1996. Copyright © 1999 International Communication Association 103 Journal of Communication, Winter 1999 inconsistencies, other studies have referred to agenda setting, priming, and framing without differentiation (e.g., Popkin, 1994). Therefore, additional research demonstrating framing effects for particular media or in specific content areas is of limited use to the field. Rather, research should address framing from a more metatheoretical perspective. In other words, how can framing be used to broaden our understanding of media effects? Is it possible to categorize framing research by key inputs, processes, or outcomes? Finally, what are the theoretical and methodological implications for future studies on framing effects? This essay proceeds in three steps to answer these questions. In the first part, framing as a concept is embedded in the larger context of media effects research, and its theoretical premises are outlined. In the second part, I develop a typology of framing research that classifies the applications of framing in media effects research along two dimensions: media versus audience frames, on the one hand, and frames as independent versus dependent variables, on the other. This typology responds to Entman’s (1993) call for the development of a consistent concept of framing, a “common understanding [that] might help constitute framing as research paradigm” (p. 56). By integrating the various, atomistic approaches to framing, the typology serves as a tool for theory building, thus contributing, as Entman argued, “to social theory in the largest sense” (p. 58). Based on this typology, previous studies on framing are evaluated with respect to their conceptual and operational contributions to framing as a concept in media effects research.1 In the third part of the essay, I develop a process model of framing, addressing deficits of previous studies and suggesting guidelines for future research in framing in the area of mass media effects. Framing, Mass Media, and Audiences This section has two goals. First, framing needs to be differentiated from other closely related concepts in mass media effects research. This differentiation requires examining framing analysis in the larger historical context of media effects research. Second, a general conceptual definition of framing needs to be developed. This endeavor involves identifying theoretical premises common to all conceptualizations of framing, and, based on these premises, developing a definition of framing generally applicable to media effects research. Framing as the Construction of Social Reality “The entire study of mass communication,” McQuail (1994) wrote, “is based on the premise that the media have significant effects” (p. 327). This diagnosis, however, must be understood as the temporary result of a scholarly discussion that has 1 The intention here is not to provide an exhaustive overview of existing operationalizations, but to use existing research to illustrate the theoretical model outlined in this essay. Space constraints force me to limit my examination to media effects in the area of political communication. This does not mean, however, that the typology developed cannot be applied to other areas. 104 Framing as Media Effects been characterized by significant changes in paradigms over the past decades.2 According to McQuail, the history of research on media effects can be divided into four stages. The first stage, from the turn of the 20th century to the late 1930s, was dominated by an experience with strategic propaganda during World War I, which led to a growing fear of the influence of media messages on attitudes. The second stage, which ended in the late 1960s, revised the paradigm of strong media effects. Personal influence was considered to be the main influence on attitude change. Klapper (1960) summed up the findings: Campaigns do not influence people; their major effect is the reinforcement of existing attitudes. Even for those who actually do change their mind, the effects are minimal. The third stage, beginning in the 1970s, was dominated by the search for new strong media effects (NoelleNeumann, 1973). The focus of research shifted from attitude change, as found in the Columbia studies, to more cognitive effects of mass media (Beniger & Gusek, 1995). The fourth and present stage, started in the early 1980s, is characterized by “social constructivism.” The description of media and recipients in this stage combines elements of both strong and limited effects of mass media. On the one hand, mass media have a strong impact by constructing social reality, that is, “by framing images of reality . . . in a predictable and patterned way” (McQuail, 1994, p. 331). On the other hand, media effects are limited by an interaction between mass media and recipients. “Media discourse is part of the process by which individuals construct meaning, and public opinion is part of the process by which journalists . . . develop and crystallize meaning in public discourse” (Gamson & Modigliani, 1989, p. 2; see also McLeod, Kosicki, Pan, & Allen, 1987). Within the realm of political communication, framing has to be defined and operationalized on the basis of this social constructivism. Mass media actively set the frames of reference that readers or viewers use to interpret and discuss public events (Tuchman, 1978, p.␣ ix). According to Neuman, Just, and Crigler (1992) “They give the story a ‘spin,’ . . . taking into account their organizational and modality constraints, professional judgments, and certain judgments about the audience” (p. 120). At the same time, people’s information processing and interpretation are influenced by preexisting meaning structures or schemas. Three dimensions of news processing have been identified (Kosicki & McLeod, 1990). Active processing refers to an individual seeking out additional sources based on the assumption that mass-mediated information in general is incomplete, slanted, or in other ways colored by the intentions of the communicator. Reflective integrators ponder or think about information they gather from mass media, or they talk to others about what they have learned from mass media to understand fully what they have learned. Finally, selective scanners use mass media only to seek information relevant to them. They skim over or ignore irrelevant or uninteresting content. In sum, according to a constructivist media effects model, audiences rely on “a version of reality built from personal experience, interaction with peers, and interpreted selections from the mass media” (Neuman et al., 1992, p. 120). 2 McQuail’s (1987) book itself is an indicator of this change in paradigms. In the second edition the author talked about the agreement “that there are effects from the media” (p. 251). The third edition referred in the same context to “significant effects” of mass media (McQuail, 1994, p. 327). 105 Journal of Communication, Winter 1999 This interactive model of construction of reality has important implications for conceptualizing framing as a theory of media effects. An analysis of the roles that audiences and mass media play in this constructivist approach requires research on various levels of analysis. Linking macrolevels and microlevels of analysis is not new and has been formulated as a postulate in other disciplines such as sociology (e.g., Coleman, 1987; Luhmann, 1995), social psychology (e.g., Doise, 1986) and political psychology (e.g., Eulau, 1977, 1986). For mass communication, multilevel analyses can be systematized by using a metatheoretical model for between-level and within-level analyses (McLeod & Pan, 1989; McLeod, Pan, & Rucinski, 1995; Pan & McLeod, 1991). Gamson (1992a) implicitly called for the application of this model for framing research when he noted the lack of theories examining “the interplay between two levels—between individuals who operate actively in the construction of meaning and socio-cultural processes that offer meanings that are frequently contested” (p.␣ 67). Media Versus Individual Frames Because frames have to be considered schemes for both presenting and comprehending news, two concepts of framing can be specified: media frames and individual frames. This terminological and conceptual distinction follows the Kinder and Sanders (1990) suggestion that frames serve both as “devices embedded in political discourse,” which is equivalent to the concept of media frames, and as “internal structures of the mind,” which is equivalent to individual frames (p. 74). The definitions by Gitlin (1980) and Entman (1991) are more specifically related to the field of political communication. According to Gitlin (1980), frames, “largely unspoken and unacknowledged, organize the world both for journalists who report it and, in some important degree, for us who rely on their reports” (p. 7). Similarly, Entman (1991) differentiated individual frames as “information-processing schemata” of individuals and media frames as “attributes of the news itself” (p. 7). All these researchers decomposed framing into media and audience frames and linkages between them. Friedland and Zhong (1996) summarized the perspective that all these studies share the belief that frames serve as “the bridge between . . . larger social and cultural realms and everyday understandings of social interaction” (p. 13). Consequently, a concept explication of framing must take into account both kinds of frames and link them consistently. Media frames. Gamson and Modigliani (1987) conceptually defined a media frame as “a central organizing idea or story line that provides meaning to an unfolding strip of events . . . The frame suggests what the controversy is about, the essence of the issue” (p. 143). Viewing media or news frames as necessary to turn meaningless and nonrecognizable happenings into a discernible event, Tuchman (1978) offered a similar definition for media frames: “The news frame organizes everyday reality and the news frame is part and parcel of everyday reality . . . [it] is an essential feature of news” (p. 193). Media frames also serve as working routines for journalists that allow the journalists to quickly identify and classify information and “to package it for efficient relay to their audiences” (Gitlin, 1980, p.␣ 7). This concept of media framing can include the intent of the sender, but the motives can also be unconscious ones (Gamson, 1989). 106 Framing as Media Effects Entman (1993) offered a more detailed explanation of how media provide audiences with schemas for interpreting events. For him, essential factors are selection and salience: “To frame is to select some aspects of a perceived reality and make them more salient in a communicating text, in such a way as to promote a particular problem definition, causal interpretation, moral evaluation, and/or treatment recommendation” (p. 52).3 The framing and presentation of events and news in the mass media can thus systematically affect how recipients of the news come to understand these events (Price, Tewksbury, & Powers, 1995, p. 4). Individual frames. Individual frames are defined as “mentally stored clusters of ideas that guide individuals’ processing of information” (Entman, 1993, p. 53). Two frames of reference can be used to interpret and process information: global and long-term political views and short-term, issue-related frames of reference. Whereas global political views are a result of certain personal characteristics of individuals and have a rather limited influence on the perception and interpretation of political problems (Kinder, 1983, p. 414),4 short-term, issue-related frames of reference can have a significant impact on perceiving, organizing, and interpreting incoming information and on drawing inferences from that information (Pan & Kosicki, 1993, p. 56). Similarly, McLeod et al. (1987) employed the concept of individual frames to describe how audiences make sense of political news. They conceptually defined individual frames as cognitive devices that “operate as non-hierarchical categories that serve as forms of major headings into which any future news content can be filed” (p. 10). Frames as Independent and Dependent Variables In addition to classifying studies with respect to their focus on media or audience frames, framing research can be broken down into research examining frames as independent or dependent variables. Studies of frames as dependent variables have examined the role of various factors in influencing the creation or modification of frames. At the media level, journalists’ framing of an issue may be influenced by several social-structural or organizational variables (e.g., Shoemaker & Reese, 1996) and by individual or ideological variables (e.g., Tuchman, 1978). At the audience level, frames as the dependent variable are examined mostly as direct outcomes of the way mass media frame an issue (e.g., Price, Tewksbury, & Powers, 1995, 1996). Studies in which frames serve as independent variables typically are more interested in the effects of framing. In the case of media frames, the most logical outcome is a link to audience frames. In the case of individual frames, the ques3 Two aspects of Entman’s (1993) definition are especially important for differentiating framing as a media effect from approaches like agenda setting or gatekeeping: selection and salience. Whereas research on gatekeeping (e.g., White, 1950; Whitney & Becker, 1982) and agenda setting (e.g., McCombs & Shaw, 1972) has commonly examined the selection and salience of issues, Entman’s (1993) definition of framing referred to the selection and salience of particular aspects of an issue rather than to the issue itself. 4 Research has identified six personal characteristics that influence global political views of individuals: personality, self-interest, leadership, group identification, values, and inferences from history (see Kinder, 1983, p. 401; Kinder & Sears, 1985, p. 671). 107 Journal of Communication, Winter 1999 tion becomes, does individual framing of issues influence evaluations of issues or political actors? Also, does the way individuals frame issues for themselves have an impact on their willingness to engage in political action or participation? A Typology of Framing Based on the two dimensions—research examining media versus audience frames and frames as independent versus dependent variables—a four-cell typology can be constructed that allows for a classification of previous studies in this area with respect to their main focus (see Table 1). The typology is valuable in three respects. First, it classifies existing research on framing with respect to the way in which it has conceptualized frames and the relationships between frames and other variables. Specifically, it permits a direct comparison of findings both within cells (i.e., consistency across different studies of essentially the same phenomenon) and between cells (i.e., compatibility of processes at different levels of framing). Second, the typology provides information on how well previous studies have answered questions pertinent to each cell. More specifically, several research questions have to be answered. With respect to media frames as dependent variable, we should ask: RQ1: What factors influence the way journalists or other societal groups frame certain issues? RQ2: How do these processes work and, as a result, what are the frames that journalists use? With respect to media frames as independent variable, we should ask: RQ3: What kinds of media frames influence the audience’s perception of certain issues, and how does this process work? With respect to individual frames as dependent variable, we should ask: RQ4: Which factors influence the establishment of individual frames of reference, or are individual frames simply replications of media frames? RQ5: How can the audience member play an active role in constructing meaning or resisting media frames? With respect to individual frames as independent variable, we should ask: RQ6: How do individual frames influence individual perception of issues? Third, and closely related to the previous two points, the typology goes beyond hypothesis testing in relatively isolated or eclectic studies in different disciplines to develop a “common understanding of the concept of framing” (Entman, 1993, p. 56). In this sense, the four-cell typology can serve as a tool for theory 108 Framing as Media Effects Table 1. Typology of Framing Research Studies Examining Frames as . . . Dependent Variables Independent Variables Media Frames Tuchman (1978) Bennett (1991) Edelman (1993) Pan and Kosicki (1993) Entman (1993) Huang (1996) Individual Frames Iyengar (1987, 1989, 1991) Gamson (1992b) Price et al. (1995, 1996, 1997) Huang (1996) Snow et al. (1986) Snow and Bedford (1988, 1992) Entman and Rojecki (1993) Nelson et al. (1997) Studies Examining Frames as . . . building by providing a common set of conceptual definitions and theoretical statements about between-level and within-level relationships. Applying the Typology: Previous Approaches to Framing Classifying and evaluating previous framing studies based on this typology is not a clear-cut task. Certain studies can fall into more than one cell. Several experimental studies examined a relationship between media and audience frames by manipulating the independent variable and measuring the dependent variable (e.g., Iyengar, 1991; Price et al., 1995, 1996). Similarly, Pan and Kosicki (1993) conceptualized a relationship between media frames as the independent variable and audience frames as the dependent variable. Their analyses, however, provided operationalizations for only one type of frame. Although these studies make important contributions to framing theory, they raise problems with respect to where they fit in the typology. For purposes of consistency, all studies are classified according to the operationalized frame. Media Frames as Dependent Variables Although many researchers have examined extrinsic and intrinsic factors influencing the production and selection of news (e.g., Gans, 1979; Shoemaker & Reese, 1996; Tuchman, 1978), no evidence has yet been systematically collected about how various factors impact the structural qualities of news in terms of framing. Based on previous research, at least five factors may potentially influence how journalists frame a given issue: social norms and values, organizational pressures and constraints, pressures of interest groups, journalistic routines, and ideological or political orientations of journalists (e.g., Shoemaker & Reese, 1996; Tuchman, 1978). 109 Journal of Communication, Winter 1999 For an examination of media frames as the dependent variable, these five factors can be grouped into two categories. Van Dijk (1985) raised the question, “why, for instance, . . . [do] news items [have] the kind of thematic or schematic structures we want to study?” (p. 70). He suspected that the way news is framed in the mass media is a result of social and professional routines of journalists. According to Edelman (1993), the choice of frames often is “driven by ideology and prejudice” (p. 232). Gamson and Modigliani (1987) assumed that the formation of frames can be explained by an interaction of journalists’ norms and practices and the influence of interest groups (p. 168). Bennett (1993) and Edelman (1977, 1993) offered a qualitative approach to this aspect of framing. For Edelman, the framing of issues by societal groups is a result of intentional considerations. He provided some evidence in the form of an exploratory qualitative analysis of the news coverage on the Gulf War of 1991. Edelman (1977) concluded that “authorities and pressure groups categorize beliefs in a way that marshals support and opposition to their interests” (p.␣ 51). These groups use mass media to construct opinions and reality, and their societal influence to establish certain frames of reference. Tuchman’s (1978) work on the construction of reality in news media has been cited as conceptualizing and measuring media frames as dependent variables. Her study provided some tentative support for the assumption that organizational structure or media and work routines of journalists have an impact on the way news stories are framed (see Tuchman, 1978, especially chapter 9). Her results, however, were limited in two respects. First, Tuchman’s work lacked a definition of framing that can serve as a guideline for future research in this area. She referred to earlier theoretical work on the general notion of frame analysis, but did not develop a clear conceptual or operational definition of media framing as the dependent variable in her analyses. Second, the dependent variable in her study was “the act of making news” (p. 12) rather than the framing of reality, with the latter being a subdimension of the former. Probably as a result of this broader focus on news production in general, rather than framing in particular, Tuchman’s analysis revealed less about the relationship between news organizations or journalistic norms and framing than could a more narrowly designed study. Media Frames as Independent Variables Conceptualizations of framing developed by Pan and Kosicki (1993), Entman (1993), and Huang (1996) examined media frames as the independent variable. These studies are highly pertinent when it comes to examining framing effects as media effects. In addition to the theoretical statements and links that they provide on a conceptual level, the question arises as to how these studies operationalized the links between media frames as independent variables and audience frames as dependent variables. Studies that have examined media frames as an independent variable can be classified into two groups. Researchers in the first group have conceptually defined media frames as an independent variable having an impact on attitudes, opinions, or individual frames. To a large degree, however, they did not provide empirical data on the links between media frames as inputs and other variables, 110 Framing as Media Effects including audience frames, as outcomes. Pan and Kosicki (1993) and Entman (1993), for example, conceptually predicted a link between media and audience frames, but provided only exploratory analyses of media frames. Using a newspaper article on an antiabortion rally in Wichita, Kansas, as an exemplar, Pan and Kosicki (1993) described the structure of news discourse in general and potential framing devices in particular. They identified four types of structural dimensions of news that influence the formation of frames: (a) syntactic structures, or patterns in the arrangements of words or phrases; (b) script structures, referring to the general newsworthiness of an event as well as the intention to communicate news and events to the audience that transcends their limited sensory experiences; (c) thematic structures, reflecting the tendency of journalists to impose a causal theme on their news stories, either in the form of explicit causal statements or by linking observations to the direct quote of a source; and (d) rhetorical structures, referring to the “the stylistic choices made by journalists in relation to their intended effects” (p.␣ 61). Examining the coverage on the downing of a Korean and an Iranian airplane, Entman (1993) conceptually defined media frames as an independent variable (i.e., as “attributes of the news itself” [p.␣ 7]), influencing both political decisionmaking and public opinion. In content analyses of newspapers, news magazines, and network newscasts, Entman identified five traits of media texts that set a certain frame of reference, and, therefore, have a critical impact on information processing: (a) importance judgments; (b) agency, or the answer to the question (e.g., who did it?); (c) identification with potential victims; (d) categorization, or the choice of labels for the incidents; and (e) generalizations to a broader national context. It is important to note that these studies make valuable contributions to using framing as a theory of media effects. The fact that they generally use data as illustrations rather than as rigid tests of their theoretical arguments is not a shortcoming. Pan and Kosicki (1993) explicitly pointed to the exploratory character of the empirical part of their study and considered it only “an initial step towards analyzing the news discourse process as a whole” (p. 55). The type of data these authors have used is an important characteristic that distinguishes them from a second group of studies. This second group of studies has conceptualized and measured media frames as the independent variable and individual frames as the dependent variable. The most promising approach here is a combination of content analytic data and survey data by Huang (1996). Huang’s study is the only approach conceptualizing and measuring media frames as the independent variable and audience frames as the dependent variable. It is therefore listed in both these cells of the typology. Using the Anita Hill-Clarence Thomas controversy as an exemplar, Huang (1996) analyzed “to what extent media frames are operative in audience frames” (p. 1). She measured frames on both a macro- and micro-level. On the one hand, she analyzed all news stories on the network evening news and in two local newspapers in terms of how they frame the controversy. On the other hand, she tried to identify individual frames among audiences by coding answers to open-ended questions on the Hill-Thomas controversy. To both types of data she applied the 111 Journal of Communication, Winter 1999 same coding scheme to show potential links. Huang’s results showed that media frames not only find their way into audience frames, but that when media and audience frames overlap, the media and the audience accord different weights to those frames. Moreover, in the case of congruent individual and media frames the media and the audience focus on different dimensions of those frames. As Huang wrote, “what was central to the media might be rather peripheral to most respondents” (p. 19). These findings are consistent with Neuman et al.’s (1992) results, which compared depth interviews (coded for individual frames) with results from media content analyses, and revealed “very different priorities” in terms of framing issues (p. 111). Neuman et al. identified a more moral individual framing of certain issues that did not follow the objectified frames employed by journalists.5 Individual Frames as Dependent Variables Similar to the previous group, studies on individual frames as a dependent variable (Gamson, 1992; Iyengar, 1987, 1989, 1991; Price et al., 1995) conceptualized a relationship between media frames as the independent and individual frames as the dependent variable. These studies focused on individual frames as outcomes, given specific types of media frames. Consequently, all these studies measured only the dependent variable and experimentally manipulated media frames as the independent variable. Therefore, they are classified here as studies on individual frames as dependent variable. Iyengar’s (1987, 1989, 1991) content analysis of network television newscasts showed that networks frame newscasts in episodic or thematic terms.6 Episodic newscasts depict public issues in the form of concrete instances or specific events. Thematic newscasts report on a more abstract level in the form of general outcomes. Iyengar’s (1991) analyses also showed that the networks rely extensively on episodic framing. He hypothesized that the type of media framing influences how audience members attribute responsibility. According to results on openended questions in the posttest questionnaire, two dimensions of attribution of responsibility can be differentiated: causal and treatment responsibility (Iyengar, 1987, pp. 818–819; Iyengar, 1991, p. 28). The hypothesized relationship was examined for five issues: crime, terrorism, poverty, unemployment, and racial inequality. Iyengar’s (1991) results showed that a relationship between media frames and audience frames is strongly contingent upon the issue under study. An experimental manipulation of highly salient issues like unemployment, for instance, proved to have little or no impact on individual attribution of responsibility (p. 62). 5 Neuman et al. (1992) offered an explanation for this seemingly paradoxical result. They assumed that there is a strong difference in how private thought and discussion are framed and how public and media discourse is framed. “This disjuncture in public and media frames,” they argued, “demonstrates that alternative frames are out there in the public discourse on issues” (p. 112). 6 Because a clear distinction between thematic and episodic frames is impossible, news stories are “classified on the predominant frame” (Iyengar, 1991, p. 18). It is important to note that little research followed Iyengar’s (1991) seminal work on thematic and episodic frames. 112 Framing as Media Effects Gamson (1992b) operationalized individual frames more qualitatively. He examined the relationship between ideas and symbols used in public discourse on the individual frames “that people are able to construct on many issues” (p. 6). In focus group discussions, he confronted subjects with four different issues: the Arab-Israeli conflict, troubled industry, affirmative action, and nuclear power. The independent variables were “critical discourse moments,” in the form of cartoons and news stories on the four issues (p. 26). Gamson identified three ideal types of formation of frames on the group level: cultural, personal, and integrated. A cultural approach to develop a common individual frame was defined as a group discussion that relies exclusively on media discourse and popular wisdom. A personal approach relies only on experimental knowledge and popular wisdom in framing the issue, but does not integrate media discourse to support it. Integrated discussions use media discourse, popular knowledge, and experimental knowledge to form individual frames in group discussion (Gamson, 1992b, p. 129). Price, Tewksbury, & Powers (1995, 1996, 1997) offered the most elaborate approach to studying individual frames as the dependent variable. They argued that previous studies have focused exclusively on politically relevant outcomes of framing and have not “examined directly the more immediate influence of story frames on readers’ cognitive responses” (Price et al., 1995, p. 5). They examined the influence of certain characteristics of media coverage on “the knowledge activation process, in particular effect of news frames on the applicability of ideas and feelings” (p. 5). Undergraduate students were asked to read news articles about possible cuts in state funding to the university. These articles were experimentally prepared to manipulate various news frames: (a)␣ a conflict frame, in which the conflict between opposing interest groups is described; (b) a human interest frame, operationalized by an article covering the retirement of a state budget director tired of struggling to provide equitable funding for all Michigan universities; and (c) a consequence frame, in which potential financial consequences for all students were mentioned. In a posttest questionnaire the subjects were asked “to write down all thoughts and feelings you had while reading the preceding article, including those thoughts that are not necessarily relevant to the article” (p.␣ 13). Coding of the open-ended question showed that issue frames of news stories had a significant influence on the respondents’ cognitive responses. The most interesting finding is a phenomenon that Price et al. (1995) called “a kind of ‘hydraulic’ pattern, with thoughts of one kind, stimulated by the frame, driving out other possible responses” (p.␣ 23). Similar to Huang (1996), Price et al. (1997) found that individuals’ frames do not exclusively depend on media coverage of an event or issue. Rather, “participants demonstrated a capacity to introduce their own thoughts, going beyond the information provided and drawing out some basic implications on their own” (Price et al., 1997, p. 496). Individual Frames as Independent Variables To find an explicit and direct link between individual frames as the independent variable and individual information processing or political action, one can turn to the social movements literature. Gamson (1985), for example, conceptualized a 113 Journal of Communication, Winter 1999 potential impact of the frames people use to interpret conflicts on the “mobilization for collective action aimed at social change” (p. 620). Snow, Burke Rochford, Worden, and Benford (1986) and Snow and Benford (1988, 1992) noted the importance of master frames and collective action frames for studying cycles of protest.7 Klanderman and Oegema (1987), Klanderman (1988, 1992), and Entman and Rojecki (1993) focused more on how master frames invented by social movements can influence people’s motivation to support these movements and to form consensus. Gerhards and Rucht (1992) tried to synthesize the previous findings into a single model (p. 582), differentiating three types of framing: diagnostic framing (identifying a problem and attributing blame and causality), prognostic framing (specifying what needs to be done), and motivational framing (the “call to arms for engaging in ameliorative or corrective action,” Snow & Benford, 1988, p. 199). If these studies are considered exemplars of research examining individual frames as independent variables, the issue of ecological fallacy arises. In other words, the aggregate-level data used almost exclusively in these studies is of only limited use when examining the potential impact of individual frames on political participation or action. This is not to say that individual-level survey measures of audience frames and individual political action per se would provide stronger evidence of framing effects. Experimental designs in lab settings would be necessary to achieve this. It is safe to say, however, that the link between audience frames and individual action needs to be explored more closely in future research, possibly using different types of research designs. Where Do We Go from Here? In addition to providing a scheme for classifying research on framing as media effects, the typology developed here fulfills a second, and probably more important, function. It helps to explicate framing as a theory of media effects. When examining media effects, however, the focus automatically shifts from a mere description of variables or classification of previous research, as provided by the 4-cell typology, to processes or the links between key variables. A Process Model of Framing I, therefore, developed a process model of framing, breaking important links down into inputs, processes, and outcomes. Figure 1 conceptualizes framing as a continuous process where outcomes of certain processes serve as inputs for subsequent processes.8 More specifically, I examine four processes: frame building; 7 For Snow and Benford (1992), master frames and movement-specific collective action frames perform the same function. They are modes of “punctuation, attribution, and articulation” (p. 138) of certain aspects of social movements. Whereas master frames operate on more macroscopic levels, however, social movement frames have a more conflict-related or movement-specific character (Gerhards & Rucht, 1992, pp.␣ 574–576). 8 I would like to thank an anonymous reviewer for suggesting this idea. 114 Framing as Media Effects Figure 1. A process model of framing research. frame setting; individual-level effects of framing; and a link between individual frames and media frames (i.e., journalists’ and elites’ susceptibility to framing processes). Frame building. As mentioned earlier, there has been several studies on the impact of factors like organizational restraints, professional values of journalists, or their expectations about audiences on news form and content (for an overview, see Shoemaker & Reese, 1996). This research, however, has not determined how media frames are formed or the types of frames that result from this process. Future research, therefore, should address the processes that influence the creation or changes of frames applied by journalists. The term frame building, borrowed from agenda-setting research, seems to capture these processes best. Similar to Cobb and Elder’s (1972) model of agenda building, the key question is what kinds of organizational or structural factors of the media system, or which individual characteristics of journalists, can impact the framing of news content. Gans’s (1979) model of news selection processes (pp. 78–79) and Shoemaker and Reese’s (1996) work on influences on media content suggest at least three potential sources of influence. The first source of influence is journalist-centered influences. Journalists actively construct frames to structure and make sense of incoming information. The formation of frames is moderated by variables such as ideology, attitudes, and professional norms (for an overview, see Donsbach, 1981) and is eventually reflected in the way journalists frame news coverage. The second factor influencing the framing of news is the selection of frames as a result of factors like the type or political orientation of the medium, or what Gans (1979) called “organizational routines.” The third source of influence is external sources of influence (e.g., political actors, authorities, interest groups, and other elites). In 115 Journal of Communication, Winter 1999 this model, news events are covered with “journalists simply holding a mirror to them and reflecting their image to the audience” (Gans, 1979, p. 79). Thus, frames suggested by interest groups or political actors as sound bites are adopted by journalists and incorporated in their coverage of an issue or event. It is rather likely that this frame-building function of mass media has a greater impact for relatively new issues (i.e., issues for which no frames have yet to be established). Frame setting. The second process of interest for framing as a theory of media effects is what I label frame setting. Again, the terminology is similar to McCombs and Shaw’s (1972) idea of agenda setting. Indeed, McCombs, Llamas, Lopez-Escobar, & Rey (1997; also McCombs, Shaw, & Weaver, 1997) have argued that agenda setting and frame setting are based on essentially identical processes. Whereas agenda setting is concerned with the salience of issues, frame setting, or secondlevel agenda setting, as McCombs and his colleagues have labeled it, is concerned with the salience of issue attributes. McCombs, Llamas et al. (1997) wrote, “The first level of agenda setting is . . . the transmission of object salience. The second level of agenda setting is the transmission of attribute salience” (p. 704). Empirical work by Nelson, Clawson, and Oxley (1997) and Nelson and Kinder (1996), however, directly contradicted the theorizing put forth by McCombs and his colleagues. Although Nelson and his colleagues did not completely reject the notion of frame accessibility, or salience, and their role in framing processes, they suggested that perceived importance of specific frames, rather than their salience among audiences, is the key variable. In other words, “frames influence opinions by stressing specific values, facts, and other considerations, endowing them with greater apparent relevance to the issue than they might appear to have under an alternative frame” (Nelson et al., 1997, p. 569). At first glance, this difference between perceived importance and salience of frames seems to be an operational one, with salience and perceived importance representing different ways of measuring essentially the same construct. Beyond operational issues, however, there are conceptual differences between salience of frames and perceived importance of frames. Salience of frames, in this sense, refers to their accessibility, or the “ease in which instances or associations could be brought to mind” (Tversky & Kahneman, 1973, p. 208; see also Hodges & Wilson, 1993; Houston & Fazio, 1989). In other words, how people think about an issue is influenced by the accessibility of frames. The frames that are most accessible are the ones that are most easily available and retrievable from memory (Hastie & Park, 1986; Iyengar, 1990). The perceived importance of frames, in contrast, is the outcome of a more conscious process of information gathering and processing. Consequently, previous research (e.g., McCombs & Shaw, 1972) has assessed perceived importance of issues and frames more directly using self-reports in survey designs. The data collected by Nelson et al. (1997) allowed them to test directly the assumption that perceived importance and accessibility of frames are indeed distinctively different constructs. In addition to measures of perceived importance of frames similar to those employed by McCombs and his associates, Nelson et al. operationalized salience of frames by measuring response latency of answers (for a review, see Bassili 1995). Their results clearly demonstrated the discriminant 116 Framing as Media Effects validity of both the perceived importance of frames and the accessibility of frames. More specifically, causal modeling revealed that various dimensions of perceived importance accounted for major proportions of the variance in framing effects. Salience or accessibility of frames played only a minor role. Future research should address the obvious incompatibility of these competing explanations further. Individual-level effects of framing. Individual-level influences of audience frames on several behavioral, attitudinal, and cognitive variables have been examined using, in most cases, black-box models. In other words, studies examining this link have focused on inputs and outputs (see Figure 1), and, in most cases, neglected the processes linking the key variables. Most research examining individual-level outcomes of framing, has assumed a direct link between media frames and individual-level outcomes. For example, Iyengar (1991) examined the impact of episodic and thematic media framing of issues on attributions of personal or societal responsibility. However, is this relationship mediated by audience framing? This question of whether audiences adopt media frames or the degree to which they use frames similar to media frames in their own information processing, has not been answered. In other words, although making important contributions in describing effects of media framing on behavioral, attitudinal, or cognitive outcomes, these studies provide no explanation as to why and how these two variables are linked to one another. More recent research has started to explore these links in greater detail. Nelson et al. (1997), for example, demonstrated, in a first step, how media frames impact the importance assigned by individuals to various ways of framing an issue. In a second step, they examined the link between these individual-level frames and tolerance for a Ku Klux Klan rally. Journalists as audiences. The final link that deserves more attention than it has received is the link between individual-level variables and media frames. It can be expected that journalists, like their audiences, are cognitive misers. Consequently, they are equally susceptible to the very frames that they use to describe events and issues. Previous research has suggested that this is indeed the case. For example, Rhodebeck (1998) argued that, although the process of framing is commonly conceptualized as a hierarchical process originating from elites, interest groups, or mass media, “there is a reciprocity in framing that the ‘top-down’ depiction omits” (p. 5). Using the term “themes” rather than explicitly referring to “frames,” Fishman (1977, 1980) suggested that, similar to “regular” audiences, journalists are susceptible to frames set by news media. Fishman (1980) demonstrated how news coverage of crimes was framed as “crimes against the elderly” by initially a small number of local media, and how that frame was soon picked up by other journalists and news media. Fishman labeled this phenomenon a “news wave.” Of course, influences of media frames on similar content in other media are a function of many factors that are of only minor interest here, (e.g., professional ethics, time pressures, and differential prestige of news sources). Are journalists’ frames of an issue mostly a function of how elites, interest groups, or other sources frame an issue? Or, do journalists themselves interpret issues based on frames conveyed to them by other news sources? Patterson’s (1993) finding that journal- 117 Journal of Communication, Winter 1999 ists are highly congruent in how they use a horserace frame in election contexts, for example, suggested one of two possible explanations: Either covering elections as competitions between two contenders is a function of journalistic norms and pressures from news directors, or these frames are, once introduced, reproduced constantly though the feedback loop suggested by previous research (e.g., Fishman, 1977, 1980; Rhodebeck, 1998) and shown in Figure 1. Questioning Answers or Answering Questions? The presentation of operationalizations and empirical results shows that the fractured paradigm to which Entman (1993) referred still exists. Even in the area of media effects, the concept of framing is far from being integrated into a consistent theoretical model. As a result of the numerous approaches to framing developed in recent years, the comparability of empirical results obtained in these studies is rather limited. Therefore, future research should integrate previous findings into a consistent model and fill in the missing causal links to develop a complete model of framing in political communication. The classification scheme depicted in Table 1 offers a first step in this direction. It identifies clear subdimensions of framing research in media effects, conceptualizing frames as media or audience frames, on the one hand, and as independent or dependent variables, on the other. Each type of framing is conceptually defined and differentiated from other cells in the typology. A typology like the one proposed is important, considering the seemingly contradictory findings in this area. In other words, are we dealing with an artifact if two researchers claim to examine the same relationship but come to different conclusions, or are there real discrepancies between their results? Based on the typology, it is relatively easy to classify studies and decide if research should be placed in different cells (i.e., if inconsistent results are an artifact, or if approaches should fall into the same category, and the results are indeed contradictory). Beyond classifying research, framing as a theory of media effects needs to be conceptualized as a process model. Rather than focusing on inputs and outcomes, therefore, future research should address the four key links indicated in Figure 1: frame building, frame setting, individual-level framing processes, and feedback from individual-level framing to media framing. In explicating these links, I have raised at least as many questions as I attempted to answer. McCombs and his colleagues’ (e.g., McCombs, Llamas et al., 1997) work on second-level agenda setting, and Nelson et al.’s (1997) work on the salience of individual frames, for example, illustrate nicely how essentially incompatible approaches to the same area of framing research continue to coexist. The process model proposed here is a framework for future research to address systematically unresolved issues in framing research and ultimately integrate atomistic approaches into a coherent theory. References Bassili, J. N. (1995). Response latency and the accessibility of voting intentions: What contributes to accessibility and how it affects vote choice. Personal and Social Psychology Bulletin, 21, 686–695. 118 Framing as Media Effects Beniger, J. R., & Gusek, J. A. (1995). The cognitive revolution in public opinion and communication research. In T. L. Glasser & C. T. Salmon (Eds.), Public opinion and the communication of consent (pp. 217–248). New York: Guilford Press. Bennett, W. L. (1993). Constructing publics and their opinions. Political Communication, 10, 101–120. Brosius, H. B., & Eps, P. (1995). Prototyping through key events: News selection in the case of violence against aliens and asylum seekers in Germany. European Journal of Communication, 10, 391–412. Cobb, R. W., & Elder, C. (1972). Participation in American politics: The dynamics of agenda-building. Boston: Allyn & Bacon. Coleman, J. S. (1987). Microfoundations and macrosocial behavior. In J. C. Alexander, B. Giesen, R. Munch, & N. J. Smelser (Eds.), The micro-macro link (pp. 153–173). Berkeley: University of California Press. Doise, W. (1986). Levels of explanation in social psychology. New York: Cambridge University Press. Donsbach, W. (1981). Legitimacy through competence rather than value judgments: The concept of professionalization reconsidered. Gazette, 27, 47–67. Edelman, M. J. (1977). Political language. New York: Academic Press. Edelman, M. J. (1993). Contestable categories and public opinion. Political Communication, 10, 231–242. Entman, R. M. (1991). Framing U.S. coverage of international news: Contrasts in narratives of the KAL and Iran air incidents. Journal of Communication, 41(4), 6–27. Entman, R. M. (1993). Framing: Towards clarification of a fractured paradigm. Journal of Communication, 43(4), 51–58. Entman, R. M., & Rojecki, A. (1993). Freezing out the public: Elite and media framing of the U.S. antinuclear movement. Political Communication, 10, 155–173. Eulau, H. (1977). Multilevel methods in comparative politics. American Behavioral Scientist, 21, 39–62. Eulau, H. (1986). Politics, self, and society: A theme and variation. Cambridge, MA: Harvard University Press. Fishman, M. (1977). Crime waves as ideology. Social Problems, 1, 531–543. Fishman, M. (1980). Manufacturing the news. Austin: University of Texas Press. Friedland, L. A., & Zhong, M. (1996). International television coverage of Beijing Spring 1989: A comparative approach. Journalism & Mass Communication Monographs, 156. Gamson, W. A. (1985). Goffman’s legacy to political sociology. Theory and Society, 14, 605–621. Gamson, W. A. (1989). News as framing: Comments on Graber. American Behavioral Scientist, 33, 157–166. Gamson, W. A. (1992a). The social psychology of collective action. In A. D. Morris & C. McClurg Mueller (Eds.), Frontiers in social movement theory (pp. 53–76). New Haven, CT: Yale University Press. Gamson, W. A. (1992b). Talking politics. New York: Cambridge University Press. Gamson, W. A., & Modigliani, A. (1987). The changing culture of affirmative action. In R. G. Braungart & M. M. Braungart (Eds.), Research in political sociology (Vol. 3, pp. 137–177). Greenwich, CT: JAI Press. Gamson, W. A., & Modigliani, A. (1989). Media discourse and public opinion on nuclear power: A constructionist approach. American Journal of Sociology, 95, 1–37. Gans, H. (1979). Deciding what’s news. New York: Patheon. 119 Journal of Communication, Winter 1999 Gerhards, J., & Rucht, D. (1992). Mesomobilization: Organizing and framing in two protest campaigns in West Germany. American Journal of Sociology, 98, 555–595. Gitlin, T. (1980). The whole world is watching: Mass media in the making & unmaking of the new left. Berkeley: University of California Press. Hamill, R., & Lodge, M. (1986). Cognitive consequences of political sophistication. In R. R. Lau & D. O. Sears (Eds.), Political cognition (pp. 69–95). Hillsdale, NJ: Erlbaum. Hastie, R., & Park, B. (1986). The relationship between memory and judgment depends on whether the task is memory-based or on-line. Psychological Review, 93, 258–268. Hodges, S. D., & Wilson, T. D. (1993). Effects of analyzing reasons on attitude change: The moderating role of attitude accessibility. Social Cognition, 11, 353–366. Houston, D. A., & Fazio, R. H. (1989). Biased processing as a function of attitude accessibility: Making objective judgments subjectively. Social Cognition, 7, 51–66. Huang, K. S. (1996, May). A comparison between media frames and audience frames: The case of the Hill-Thomas controversy. Paper presented at the annual conference of the International Communication Association, Chicago, IL. Iyengar, S. (1987). Television news and citizens’ explanations of national affairs. American Political Science Review, 81, 815–831. Iyengar, S. (1989). How citizens think about national issues: A matter of responsibility. American Journal of Political Science, 33, 878–900. Iyengar, S. (1990). The accessibility bias in politics: Television news and public opinion. International Journal of Public Opinion Research, 2, 1–15. Iyengar, S. (1991). Is anyone responsible? How television frames political issues. Chicago: University of Chicago Press. Iyengar, S., & Kinder, D. R. (1987). News that matters: Television and American opinion. Chicago: University of Chicago Press. Kinder, D. R. (1983). Diversity and complexity in American public opinion. In A. W. Finifter (Ed.), Political science: The state of the discipline (pp. 389–425). Washington, DC: American Political Science Association. Kinder, D. R., & Sanders, L. M. (1990). Mimicking political debate with survey questions: The case of white opinion on affirmative action for blacks. Social Cognition, 8, 73–103. Kinder, D. R., & Sears, D. O. (1985). Public opinion and political action. In G. Lindzey & E. Aronson (Eds.), Handbook of social psychology (2nd ed., Vol. 2, pp. 173–196). Reading, MA: Addison-Wesley. Klandermans, B. (1988). The formation and mobilization of consensus. In B. Klandermans, H. Kriesi, & S. Tarrow (Eds.), International social movement research. (Vol. 1, pp. 173–196). Greenwich, CT: JAI Press. Klandermans, B. (1992). The social construction of protest and multiorganizational fields. In A. D. Morris & C. McClurg Mueller (Eds.), Frontiers in social movement theory (pp. 77–103). New Haven, CT: Yale University Press. Klandermans, B., & Oegema, D. (1987). Potentials, networks, motivations, and barriers: Steps towards participation in social movements. American Sociological Review, 52, 519–531. Klapper, J. (1960). The effects of mass communication. New York: Free Press. Kosicki, G. M., & McLeod, J. M. (1990). Learning from political news: Effects of media images and information-processing strategies. In S. Kraus (Ed.), Mass communication and political information processing (pp. 69–83). Hillsdale, NJ: Erlbaum. 120 Framing as Media Effects Lodge, M., & Hamill, R. (1986). A partisan schema for political information processing. American Political Science Review, 80, 505–519. Luhmann, N. (1995). Social systems. Stanford, CA: Stanford University Press. McCombs, M., Llamas, J. P., Lopez-Escobar, E., & Rey, F. (1997). Candidate images in Spanish elections: Second-level agenda-setting effects. Journalism & Mass Communication Quarterly, 74, 703–717. McCombs, M. E., & Shaw, D. L. (1972). The agenda-setting function of mass media. Public Opinion Quarterly, 36, 176–187. McCombs, M., Shaw, D. L., & Weaver, D. (1997). Communication and democracy: Exploring the intellectual frontiers in agenda-setting theory. Mahwah, NJ: Erlbaum. McLeod, J. M., Kosicki, G. M., Pan, Z., & Allen, S. G. (1987, August). Audience perspectives on the news: Assessing their complexity and conceptual frames. Paper presented at the annual conference of the Association for Education in Journalism and Mass Communication, San Antonio, TX. McLeod, J. M., & Pan, Z. (1989). Getting levels across and crossing levels. American Behavioral Scientist, 33, 199–202. McLeod, J. M., Pan, Z., & Rucinski, D. (1995). Levels of analysis in public opinion research. In T. L. Glasser & C. T. Salmon (Eds.), Public opinion and the communication of consent (pp. 55–85). New York: Guilford Press. McQuail, D. (1987). Mass communication theory: An introduction (2nd ed.). Thousand Oaks, CA: Sage. McQuail, D. (1994). Mass communication theory: An introduction (3rd ed.). Thousand Oaks, CA: Sage. Nelson, T. E., Clawson, R. A., & Oxley, Z. M. (1997). Media framing of a civil liberties conflict and its effect on tolerance. American Political Science Review, 91, 567–583. Nelson, T. E., & Kinder, D. R. (1996). Issue framing and group-centrism in American public opinion. Journal of Politics, 58, 1055–1078. Neuman, R. W., Just, M. R., Crigler, A. N. (1992). Common knowledge. News and the construction of political meaning. Chicago: University of Chicago Press. Noelle-Neumann, E. (1973). Return to the concept of powerful mass media. Studies of Broadcasting, 9, 67–112. Pan, Z., & Kosicki, G. M. (1993). Framing analysis: An approach to news discourse. Political Communication, 10, 55–75. Pan, Z., & McLeod, J. M. (1991). Multilevel analysis in mass communication research. Communication Research, 18, 140–173. Patterson, T. E. (1993). Out of order. New York: Random House. Popkin, S. L. (1994). The reasoning voter, communication and persuasion in presidential campaigns. (2nd ed.). Chicago: University of Chicago Press. Price, V., Tewksbury, D., & Powers, E. (1995, November). Switching trains of thought: The impact of news frames on readers’ cognitive responses. Paper presented at the annual conference of the Midwest Association for Public Opinion Research, Chicago, IL. Price, V., Tewksbury, D., & Powers, E. (1996, May). Switching trains of thought: The impact of news frames on readers’ cognitive responses. Paper presented at the annual conference of the International Communication Association, Chicago, IL. Price, V., Tewksbury, D., & Powers, E. (1997). Switching trains of thought: The impact of news frames on readers’ cognitive responses. Communication Research, 24, 481–506. 121 Journal of Communication, Winter 1999 Rhodebeck, L. A. (1998, April). Framing policy debates on old age. Paper presented at the annual conference of the Midwest Political Science Association, Chicago, IL. Shoemaker, P. J., & Reese S. D. (1996). Mediating the message (2nd ed.). White Plains, NY: Longman. Snow, D. A., & Benford, R. D. (1988). Ideology, frame resonance, and participant mobilization. In B. Klandermans, H. Kriesi, & S. Tarrow (Eds.), International social movement research (Vol. 1, pp. 197– 217). Greenwich, CT: JAI Press. Snow, D. A., & Benford, R. D. (1992). Master frames and cycles of protest. In A. D. Morris & C. M. Mueller (Eds.), Frontiers in social movement theory (pp. 133–155). New Haven, CT: Yale University Press. Snow, D. A., Burke Rochford, E., Jr., Worden, S. K., & Benford, R. D. (1986). Frame alignment processes, micromobilization, and movement participation. American Sociological Review, 51, 464–481. Tversky, A., & Kahneman, D. (1973). Availability: A heuristic for judging frequency and probability. Cognitive Psychology, 5, 207–222. Tuchman, G. (1978). Making news: A study in the construction of reality. New York: Free Press. van Dijk, T. A. (1985). Structures of news in the press. In T. A. Van Dijk (Ed.), Discourse and communication: New approaches to the analysis of mass media discourse and communication (pp. 69–93). New York: W. de Gruyter. White, D. M. (1950). The “gate keeper”: A case study in the selection of news. Journalism Quarterly, 27, 383–390. Whitney, D. C., & Becker, L. B. (1982). “Keeping the gates” for gatekeepers: The effects of wire news. Journalism Quarterly, 59, 60–65. Wicks, R. H. (1992). Schema theory and measurement in mass communication research: Theoretical and methodological issues in news information processing. Communication Yearbook, 15, 115–145. 122