Voltar à página inicial da revista Estudos em Comunicação

ScopusErihEbsco Host - Research DatabasesUrlrichs Web - Global Serials DirectoryDOAJ WebLatindex

Interview with Maxwell McCombs

Jan Alyne Barbosa e Silva1


During the week of May, 26 and 30, 2008, Maxwell McCombs had been lecturing at Universidad de Navarra, in the city of Pamplona, Spain, where has been a visiting professor since 1994. In order to sum up his huge curriculum, McCombs is graduated in Journalism at Tulane University, worked as a reporter at New Orleans Times-Picayune, and received his Masters and his Ph.D. from Stanford University. He had also been director of the News Research Center of the American Newspaper Publishers Association for ten years, and, at present, holds the Jesse H. Jones Centennial Chair in Communication at the University of Texas at Austin, where he runs the subjects of Content Analysis, Agenda Setting, Contemporary Trends in Journalism and Political Communication.
Worldwide cited reference and precursor in agenda setting research, he is quite fond of metaphors, has a very refined sense of humor, and method accuracy when it comes to explain the ideas surrounding agenda setting research, which he has been conducting for over 40 years together with collaborators from several different universities and countries. This interview is rather a collection made from the classes he had run in that week, and some hours of conversation we had taken to complement the material. Besides making a comprehensive review about the settings, concepts and domains within agenda-setting research, McCombs talks about important studies conducted within the field, collaborations with scholars and students, criticisms addressed to agenda-setting theory, as well as new challenges and perspectives for research on the World Wide Web.

A comprehensive perspective of agenda setting theory

MM: Agenda-setting started with a very small study in Chapel Hill in North Carolina, in the summer of 1968, during a presidential election in United States, and that small study tested basically one hypothesis that has expanded to a theory which have five distinct stages. The intellectual father of agenda-setting theory is Walter Lippmann, an American Journalist and social commentator, and in 1922, he published an influential book on public opinion. In the opening chapter of Public Opinion has the title: The world outside and the pictures in our heads: the world outside, meaning reality, and the pictures in our heads, meaning the world as we imagine it to be, what we think it is. His thesis was that the news media is the link between the world outside and those pictures in our heads, and he made the important observation that people's behavior is a response, not to the environment as it actually exists, but to the environment as they think it exists. Although Lippmann had never used the term "agenda-setting", the idea he was presenting was essentially what we call now agenda-setting. The field of communication really didn't begin to develop until 1930s, and, particularly for the area of political communication, I usually mark the historical beginning of the field with the Erie County election study, in the United States, in 1940, almost 18 years after Lippmann's book was published. But in Columbia school, Lazarsfeld and his partners took a very different approach from Lippmann's to study mass communication. Lippmann's ideas sat there, people read them, but it didn't really attract a lot of researchers. The Lazarsfeld's school carried out a number o empirical studies in the 1940s and in the 1950s. By roughly 1960, one of Lazarsfeld students, Joseph Klapper, wrote a book called Effects of Mass Communication, which I sometimes say that it's the worst title book I've ever seen in our field, because the argument of the book is essentially that there are no effects of mass communication. And that was where essentially where the field had arrived both theoretically and empirically in about 1960s. There were a number of studies that would find no media effects on attitudes and opinions - which is the focus of Lazarsfeld's research. But persuading people is not the primary purpose of Journalism. The primary purpose of Journalism is to inform. So we thought that, perhaps, there are other kinds of effects, and it was out of that perspective that we drew the idea of agenda setting. Then, I moved from UCLA, in Los Angeles, to North Carolina, where I met Don Shaw, who has been not only my research collaborator, but also my best friend for more than 40 years now, and we began to talk about how we could test Lippmann's idea, and that's when we came out with the metaphorical idea of agendas. So, in the 1968 presidential election, we did a very small test of the idea that has come to be known as agenda-setting. We did not have a huge grant to do this, and we needed to do it in a very strategic way. If there are effects in an election, where are they most likely to be bound? Klapper's explanation of why there were no effects was based on the idea of selective perception, that media did not have effects on people's attitudes and opinions because people put up a psychological defense against being persuaded. If selective perception is the explanation for a lack of effects, we decided to focus our study on undecided voters; we believed that if there would be some kind of media influence, they were likely to be seen among them, because they would be more open to media messages. We decided to look at the agenda of issues in the elections presented in the media and at the undecided voters' agenda, and see of we could find some evidence of media influence over the perception among members of the public. The kind of influence we were interested about was not so much attitudes and opinions, but rather perceptions of importance; the key connecting concept here is: transfer of salience or prominence. So the media agenda is simply a description of the patterns of coverage - in the case of Chapel Hill study - of the issues in the campaign. These effects may take some time to take place. We could look at the pattern of coverage of over a week, or over a month, six weeks, two months, over some period of time, using content analysis. So, we did content analysis of nine media that undecided voters used in Chapel Hill to find out about the elections, and we could determine what the media's agenda was, using rank order, that is, what was the pattern of coverage for the issues of the campaign, and more specifically, which issue is most prominent, which is the second, third and so on. The public agenda generally means that you have to do some kind of survey or poll. At the time we were doing this research, we did our own survey, making a lot of phone calls and knocking on doors to find undecided voters. So, we found a small group of undecided voters and fortunately we had some guidance from earlier public opinion research about how to get the agenda of issues among the public. Back in the 1930s, the Gallup Poll began asking the open-ended question that goes something like this: "What do you think is the most important problem facing this country today?" The jargon of this question is MIP. It's a well-established question among public opinion research and we used a variation of Gallup question. Then, we had a set of issues that have been ranked according to how they have appeared in the media, and the same set of issues - how important the public perceives them to be - and we can construct a correlation between the two. In the original Chapel Hill study, the correlation was almost perfect. That suggested that, indeed, there is a great correspondence between the media's agenda and the public agenda. Obviously, it didn't establish causality, but it showed that it was a promising area of research. That was in 1968. We tested one other thing in Chapel Hill study. We found this striking correlation between the pattern of coverage of issues in the media, and the prominence of those issues among the public, which suggested a significant influence of the media on the public. Conventional wisdom would say that selective perception blunts these effects, and we thought that maybe they would only occur among people who were very undecided. So we went back and looked at our data, and included all types of undecided voters - the ones who had no idea about who they were going to vote, and those who were leaning towards a candidate, but were not totally committed - and wondered if, among those people, we could find influence of selective perception, and to what extent selective perception is a better explanation, or to what extent is agenda-setting a better explanation for what issues they think are important. The selective perception argument for agenda-setting effects is that if you were leaning to the Republican candidate, you would mostly read articles or watch TV news stories, and tend to pick up the issue agenda about the Republican candidate. The same thing would happen to the Democrat leaning voters. Our hypothesis was that if a person were leaning to vote for the Republican candidate, the agenda they would report as important would reflect the total pattern of media coverage, not the Republican's, and the same for the Democrat's. That would be a competing hypothesis: the agenda setting hypothesis against the selective perception hypothesis, and, overwhelmingly, we found the evidence supporting agenda-setting effects. This study is called Stage 1 of agenda-setting research and it is still a very active area. There are five stages at present and they are not stages in the sense of historical stages, because all of them remain active arenas of research. In 1972, the next presidential election, we set out to do two important things in expanding this research. First, to test the idea of causality, measuring both the public agenda at several points overtime and the media agenda at several points overtime, and also to see if we could determine in which direction was the influence. That had been done in a larger city in North Carolina, Charlotte. Another important aspect of that study was that we took a random sample with all voters: decided and undecided. In the 1972 study, the evidence was very clear-cut, that it was the influence of the media on the public and not the other way around. This was replicated again in the next presidential election, in 1976, in three different communities in the United States, and at this point the research began to expand very very rapid. We also introduced the Stage 2 with this research. Our argument was that the media had unlimited power in focusing public attention on certain issues, and this was not a return to the earlier theories, such as the hypodermic theory. There were obviously some constraints of media influence and in 1972, in Charlotte, we began to define what those constraints are, specifying the processes that links the media agenda to the public. The key aspect of the process, which is the second stage, is the psychological concept of need for orientation, that explains the limits of media influence on the public agenda. The idea is that people need to map or understand their surroundings.
McCombs 2
Figure 1: Maxwell McCombs
David Weaver, a graduate student of Donald Shaw in Chapel Hill, worked on this idea of how people approach the media. The first aspect of need for orientation is that no one reads every item on a newspaper everyday, because they don't see all of them as relevant. And the second aspect is uncertainty. If you already know everything you want to know about a certain topic, you may or may not need to know more about it. This concept cannot be measured, because there's no absolute level of uncertainty. Some people read just two facts about a topic and are quite happy with that. Other people read 12 books on the topic and they feel that they don't know enough. So, uncertainty is defined by each individual. Regarding the public issues, if the need for orientation is very low, you are not going to find very strong agenda-setting effects. If the need for orientation is moderate, you are going to find moderate effects, and if the need for orientation is very strong, you are going to find very strong effects. In Chapel Hill study, we were very lucky in one sense; we studied only undecided voters. This concept had not been developed at that time, and that group we studies had a high need for orientation. The third stage stands from the idea of what is on the agenda. I talked about an agenda of issues, but we would talk about other items on the agenda, such as corporations, political candidates or brand names. The Chapel Hill and Charlotte studies have looked at the agenda of objects, that is, the thing you have an opinion about, and if the agenda of objects become prominent on the public agenda. But the objects have some characteristics, attributes that define them. When the media talk about some object, they don't just name the object, they describe it in some fashion. And this is the third stage, the attribute or second level agenda setting, which began in the 1970s. In this sense, agenda-setting effects can occur in terms of objects or in terms of attributes. The basic research model is the same, with rank order and content analysis to see if there are correlations between the public and media agenda. To measure the attribute agenda among the public, the open-ended question we used in 1976, with three different communities in the US, was: "Suppose you had a friend that have been away for a long time and knew nothing about George Bush (or another political figure), what would you tell them?" It doesn't push you to any particular attribute, and we find the same kind of agenda setting effect for attributes, as we have found for objects. As soon as we begin to know the image in people's heads, we already begin to move into the area of the consequences of agenda-setting effects, which lead to attitudes and opinions, meaning that your behavior is a response to the pictures you have about an object. If we get back to the chapter of Walter Lippmann's book - The world outside and the pictures in our heads - first level of agenda setting says: What are the pictures about? Second level of agenda setting literally says: what are the pictures? And then, we begin to say: What are the consequences of having the pictures? On the fourth stage, we went back to considerations of attitudes and opinions, because if we think back 40 years, when we began doing agenda setting research, people were convinced that there were no media effects on attitudes and opinions, and we went off in a different direction. So, we came back to that question, but now looking in a much more nuanced way. We don't expect all of the content of the media to have an impact on people's attitudes and opinions, only those aspects of the content that people consider relevant, and for what they see conditional information in the media. So we kind of understand the conditions under which we may find these attitudes and opinions. The fifth and last stage is concerned with the sources of the media agenda. So, the media agenda, which has been an independent variable, becomes a dependent variable. Historically, this occurs in 1980s and asks: if the public is set by the public agenda, who sets the media agenda? Where does this agenda come from? I have a metaphor to answer that question, and the metaphor is peeling an onion. An onion has many layers and you can make this theoretical onion as more simple or complex as you want to. I usually do a very simple onion and it has about three layers: the outer layer is composed of humongous events, like typhoons and plain crashes, and things that obviously get into the news. But if we think about more deliberate building of an agenda, then, the other layer consists of those efforts by political campaigns, public relations agencies, public information officers etc. to feed information to the media, organized information that comes from the traditional news sources of journalism. So, one source would be external, another sources are other media. The professional values of Journalism, the traditions of Journalism influence the shape of the media agenda. Some kinds of things are considered important, some are not considered important, some aspects of a news event are considered important, some are not so important, these help shape not only what gets covered, but also how it gets covered. And this changes somewhat overtime, not radically, but they change somewhat overtime. So, the third area of influence is often referred to as Intermedia Agenda-Setting, that is, the influence of news organizations on each other. In most content, there is at least one, sometimes more than one, elite news media, and it tends to influence - more than these other aspects - the shape of media agenda overtime. In the United States, The New York Times plays that role of agenda-setter of American press. Such an institutional role has simply developed informally overtime, but it is sufficiently established now that every afternoon, The Associated Press sends its U.S. clients a memo, not for publication, it is an internal memo that says:  `The news stories on the front page of The NYT tomorrow morning will be:' and, then, the list. In Austin, Texas, where I live, the Austin American-Statesman can even take one step further, by printing the South Western addition of The NYT, which is transmitted by satellite, even before printing the local paper. The design of The NYT and the design of the Austin American Statesman are totally different; they don't look anything alike. But if you look at the agenda of the stories, they will be very similar. Obviously, a big local story is going to get the top of the agenda in Austin. But when you begin to look at the national stories of the agenda, they are going to be closely modeled after The New York Times. So, there is an entire area of research within Journalism - Intermedia agenda-setting - the influence of the news media on each other to the study of news media themselves. The earlier work refers to this area as Sociology of News that typically goes back to the traditions of journalism and the influence of sources on Journalism, and that begins to link agenda-setting within an established academic tradition. Those are the five stages of agenda-setting research, and they represent the settings within the domain of public affairs. The domains and the settings are nothing more than operational definitions, and they open up the possibilities to all sorts of new applications, where you simply look at the relationship between two agendas at the extent to which the prominence of the item on one agenda can influence the prominence of that item on another agenda, and it can be any agenda that you are interested in, any set of objects, and any set of the characteristics of the attributes of those objects that you are interested in.
J: We know you are a journalist, but how was the link between you journalistic career and your academic career? How did it start? How did that happen?
MM: My short answer is that I was a victim of a benevolent conspiracy (laughs). When I studied Journalism as an undergraduate, at Tulane University in New Orleans, I had a professor there, Walter Wilcox, who, when I was finishing the course, said: `Before you begin to work as a journalist, you should go to Stanford, and get a Masters'. So, I applied to the Masters program at Stanford and went to the very next fall to California to study there. The head of the program was the advisor to all graduate students and this was his last year before retiring. I appeared in his office in California in the fall and referred to my professor in Tulane and he said: `Oh yes, Walt told me about you'. He suddenly took a pad and a paper and said: `what do you want to take? It's this theory course', and wrote that down, `and my Content Analysis course', and wrote that down. "Go upper to the Psychology department and take Statistics course, and, while you are there, take a Learning Theory course". And then, as I sometimes put it, with some lots of democracy, he says: `Go home, think about it, and come back tomorrow'. I had no criteria! He said I should take these things, and I did. Essentially, as a Master student, I took the first year of the Ph.D. program and, then, he said: `go be a Journalist for a few years, then come back and we'll support you Ph.D. program'. So I wrote a number of newspapers, and the Times-Picayune in New Orleans had an opening, so I went back to New Orleans, because I liked that city very much, and was a journalist for three years. But by then, I had discovered this whole world of research, and sometimes went to the library in Tulane and read. So, as what I've said, I was a kind of a victim of a benevolent conspiracy that pull me into the academic.
J: And Stanford at that time was doing research on media effects...
MM: Everything at Stanford was geared. They had a few undergraduate Journalism courses, but their program was really a Ph.D. program in Communication Research.
J: And your thesis? What was it about? We read your articles, but we don't know much about your thesis...
MM: It was published, but I don't think anyone has cited it in years; it was an experiment on how children learn language from television. It had nothing to do with later research.
J: What year was it?
MM: That would be in 1966, and an article based on that was published in Journal Communication about four years later.
J: Many scholars complain that media research is not centered in Communication itself, but it is rather the result of interest from many distinct fields. Do you believe that media effects is the research tradition which comes closer to the Communication field itself?
MM: Yes, it is certainly. If you look at the literature of the years, it has been the dominant focus. And it draws mainly on the studies from other fields in the 1930s and in the 1940s, and from Lazarsfeld's group; of course they were sociologists who studied elections, Lasswell, a political scientist, Berelson... they began to define media effects in the 1930s and 1940s and then, when the Journalism schools began to offer their own Ph.D. programs, they began to follow that same tradition very strongly. But of course, as I remarked the other day, by roughly 1960, the evidence was: well, there aren't a lot of effects on attitudes and opinions, and people began to redirect their attention to other areas, where agenda-setting came from. But the effects are still there, I mean, agenda-setting is still very much a theory of effects, and it covers a lot of other things, but it is interested in effects, its evolution tends to follow that hierarchy of effects model. Framing is essentially going down on the same path, as it is concerned with effects. There are some other parts of the field in our college in the University of Texas; there is a Department of Communication Studies, and many of those people come out of more of a speech tradition of rhetoric, but a lot of them also are into the political communication area, and they are looking at...effects! The only time it appeared it might move in a clearly different direction was in the...I have to look up the citations...but a big swing interest in uses and gratifications, I guess in the 1970s. I mean, again, that tradition goes way back, there were studies particularly by the Columbia group in the 1940s on uses and gratifications, studies of audiences for soap operas, I think Berelson did a research on what it means to miss the newspapers, when the newspapers strike, the routes of uses and gratifications also go way back, it had never been a big segment of the field until I guess it was Blumler and Katz did a book, some time in the 1970s, and, for a period of maybe roughly ten years, it was a big area of interest, the uses and gratifications, and many people at the time said: `uses and gratifications is going to push aside the dominants of media effects and that is going to become, if not the main focus of the field, equally big with the effects'. It really didn't happen because somehow - I mean, we have the residual impact of that, now we routinely talk about uses and gratifications - there are a few studies of uses and gratifications. It never became a theory, it has always been a kind of remained perspective, and, after roughly eight or ten years, the perspective was reasonably well developed, and then, it got into what I call the dark side of academic world, where you do a study and you say: 'For Internet users, when you look at the uses and gratifications, there are these five categories'. And they say: 'no, no, no, there are seven'. I mean, those kinds of arguments can go on forever and that just turns in on itself of these kind of minor disputes over, you know? 'Is it the boundary line here or is it over there?' And my concern is the same thing may happen to framing because there are already so many competing definitions and perspectives on framing that one person says 'oh, framing is this', and other says, 'no, no, it's this'. And I would say: `well, at least some aspects of these definitions overlap with agenda-setting', and the other would respond: 'no'.
J: I've read some articles of the book Framing Public Life (1991) and some of them say that attribute agenda setting is not framing research. I have to confess that I couldn't see quite clearly the difference between them.
MM: Most people can't.
J: When Steve (Reese) came over here, I believe that the first question I asked him was about the difference between attribute agenda-setting and framing.
MM: It depends on what definition of framing is. In very general terms, both, agenda-setting theory and the various approaches to framing are concerned with - using some of Lippmann's terms - how the media structure representations of the world and in terms of how those representations influence how people see the world. So, the general interest of the two theories is very much the same: representations of reality and the influence of that on individuals. The comparisons of the research in the two areas have become somewhat problematic because there are many definitions of frames. My approach is not to look at the literature about he definition of frames, because there are probably a hundred, but rather, review the literature on the studies that have been made, and where framing and attribute agenda-setting overlap. Other kinds of research will say they're not equivalent, and some others say they're related, that is, they overlap, but they are not quite the same thing. Taking that approach where attributes and frames are equivalent: I put in my briefcase an issue of Communication and Society because there are two articles on framing, according to the titles, and one article on agenda-setting, according to the title. They are three out of the five articles of this issue of the journal. If you look at the lead article: `Media framing on capital punishment and its impact on individuals cognitive responses'. The article has two parts. The first one is content analysis of media framing on capital punishment, and the second part is an experiment. If you look at the data tables of this particular article, you see that you could have done the same study calling it a second level of agenda-setting study, where the attribute of the study is capital punishment that have appeared in the media overtime. Both approaches tend to start at the very traditional content analysis, and you develop a list of categories, where different perspectives on this issue have been talked about, and then, you go through to count how many articles talked about this and how many talked about that. There are a very large numbers of these articles that appear, and that is probably the most common that has been for perhaps ten years now. Some people have argued that this is really not framing, that is just traditional content analysis with a fancy new title. Steve Reese has made that argument that this isn't really framing research. But it appears under that name in the journals. Twenty years ago it would have been media coverage of, or if you are doing an agenda-setting study, you would do something other than just measure the attributes' agenda, because the agenda-setting deals with the relationship between that media agenda and usually the public agenda. The recent issue of Journal of Communication talks mostly about framing. Two of them talk about the knowledge activation model, and this is a theoretical paper originally done about ten years ago that uses that model, and says agenda setting is this and framing is that. But the empirical evidence is that distinction doesn't hold up. As soon as I got that issue of the journal, I scanned it and looked at the footnotes and I said: 'what evidence do they support for this distinction?' They don't. About three weeks ago, Joan Mellor, who is a psychologist, but her appointment is in the Political Science Department at the University of Minnesota, launched a very interesting set of experiments which directly tested that model, and two of the experiments directly tested the notion where they say 'agenda-setting is this, framing is that', and neither experiments support that position, they found no supporting evidence. I reviewed a proposal for a book recently by Franklin Angelo, who proposes to get chapters representing different points of view about framing, to try to create a dialogue, and get some convergence to sort some of this out. So, when I wrote my review back to the publisher, an editor with whom I work for about 20 years, I said: 'This is a very exciting book, an important needed book, and I certainly endorse it for publication, and I wish Frank good luck to get that among those framing people.
McCombs 3
Figure 2: Maxwell McCombs
He talks about a dialogue and I hope he can create that, rather than everyone saying: 'my view is...' There are many definitions of frames that, in fact, are a series of relationships, and, in some instances, I think they are essentially the same. We have already talked some about the large body of primary descriptive research, where the word framing appears in the title, but in many ways, they are simply the same kind of traditional content analysis of news coverage that exists within seventy years. They would just have a nice name, where they call framing, rather than just press coverage. But the operational definitions in most of those studies of a frame would function as the operational definition of attribute agenda-setting. If I clip the data out of those I would say: Is this a framing study or an attribute agenda-setting study? There is no way in the world to make a distinction! So there's a large literature there, it's primarily descriptive, a few of those studies move on to look at effects and that's perhaps the main distinction between that body of work on attribute agenda setting. We've looked at the example of campaign story with the descriptions of the candidates as framing stories simply presented as frames. But my seminar analyzed that it's an attribute agenda setting study, and found that, in fact, it was evident it was significantly influenced in the campaign year of the campaign of the media agenda. Moving over toward the proposition that perhaps these concepts overlap, they operate in essentially the same domain, and perhaps what it distinguished slightly different is the idea of Salma Ghanem (1996) of compelling arguments. That is: particular attributes seem to play a distinct role in the process of influence. One possible use of the term frame is to restrict to more fundamental critical trades that functions as what Salma would call compelling arguments. If you have evidence of that, the simple definition would be the dominant attribute in some set of attributes. The compelling argument idea has a lot of potential. Here and there in the literature you find evidence, and not always using that term, that certain aspects of the situation exert most of the influence of the media coverage on the public response to that coverage. I don't think there is a huge number on that study. You may be able to count them on one hand, and almost certainly on two, but there might be an area where people could ask: `are there, for various issues compelling arguments, certain attributes that play a particularly strong role in influence in the public response to that?' More interesting would be - and you could find some generalizations across issues - what it would get into what we might call cultural definitions of frames. Are there certain ways of talking about various issues that a particular public in a particular country find specially compelling? That begins to get into a more fundamental cognitive structure, that is how people form their images and organize their thoughts about particular public issues. To my knowledge, no one has truly done anything with that, but I think it's a very nice extension of Salma's idea of compelling arguments. This brings to an extent to what Toshio Takeshita is doing in Japan; his distinctions about attributes and frames have a hierarchical relationship to each other. That is where they are connected, where they overlap. He has frames as more abstract categories of dimensions, and you have attributes as more or less what we are accustomed to dealing with, through the accumulated tradition of content analysis and agenda-setting research. The studies he has done to date, look at one issue - the economic situation in Japan - and he began the research with focus groups, getting pictures of people talking about the issue and the purpose of the focus groups was to identify the appropriate attributes of the economic situation that describe the way people at that time were thinking about the problem in that country. Out of those focus groups, he asked people: `how much of these things matter? Why we are having economic problems?' for: performing banks, overprotective policies, rising levels of public expenses, unemployment, changes in the unemployment system, a pretty traditional list of, I think, 13 attributes, and the survey he did to measure the public attribute agenda. So, the first step of the research is simply to measure the attribute agenda in the public. Once he has that data, he goes to the content analysis and look at the same attributes agenda in the media. At that point, it looks very much to the traditional study of second level of agenda setting. The addition is that he fits this data into a higher level or the abstract theoretical framework. The general theoretical framework he is working with is the one of problematic situations, a kind of way to look at attributes in a more abstract fashion. A book titled something like Communication and Culture is a study done not just in Japan, but also in Germany and the United States. I recall also some data from Hong Kong and a number of other studies from other countries. It's a very interesting comprehensive international study looking at this notion of problematic situations. In the original study by Ito Kepplinger and Eldelstein, they came up with seven basic problematic situations: loss of value, need for value, institutional breakdown, social conflict, steps towards solutions, blocking the obstacles and undetermined situations. The attribute agenda-setting answer is using a time lag of 13 weeks. If you compare the way the media talk about the economic situation and the way the public in their response on the survey talk about them, there will be the usual kind of outcome attribute agenda-setting analysis, the way public is thinking about the economic situation is very much the way that the media was talking about that.
J: What makes you to establish partnership with other scholars to work on a project?
MM: I've always been very open to working with other people. I just find it much more interesting to exchange. I did not know Don Shaw until I moved to the University in North Carolina, but we quickly became...well, I'm not quite sure whether we became friends first or research partners first, it kind of happened simultaneously, we were roughly the same age, we were interested in many of the same things, and we can still work together. It's much the same over the years with people like Toshio Takeshita, and particularly lots of graduate students. If people have interesting ideas, and I think I have something to contribute...but sometimes students come and say 'would you work with me on this?' and I like the idea, and I say 'I think you've got a great idea, I'm not sure you need me, I say 'I'll be happy to kind of listen to its developments, but go do it yourself, you don't need a partner'. I've always enjoyed collaborating with people.
J: I made a list of some criticisms on agenda-setting research and I would like you to make comments about them.
MM: Ok.
J: That agenda-setting it's not a theory, but rather a model or a hypothesis, because it is always subject to be tested.
MM: People have very strange definitions of a theory. As you saw, it states very clear propositions about a whole series of relationships, and how agendas are related in the psychology of that relationship. I think it's well developed as any theory in our field. You could say, maybe the Chapel Hill study tested a hypothesis, but we are a long way from Chapel Hill study.
J: How about the sterilization of public issues when traditional agenda setting research follows the legacy of public opinion polling?
MM: To some extent, I think that is true, but if you think about that the key measure, the Most Important Problem question is an open-ended question, and people can say whatever they feel like saying. There's always the temptation in analyzing the data to fit in the categories they'd been using in the past, but if you have some sensitivity to your data, you'll become aware of that. What it used to be one category maybe now needs to be two, or I need to address that not just one. I think that's probably particularly the strength of attribute agenda-setting, you don't just deal with the general category of topics, I think you begin to dig into what are the components. A study that no one has never really done (I'm not sure if it has been done): there are studies that look at the history of issues overtime; there are a few like this, which have looked at one issue over long periods of time. It would be interesting to do more longitudinal attribute agenda-setting studies, and particular try to link those across issues and here is another possible junction point with framing, to the extent that you take a definition of a frame from a more cultural perspective. This is the kind of part of the culture of the way people look at these things. That would suggest that you ought to find some more patterns of attributes across certain kinds of issues, and even, perhaps overtime, they all systematically change. If they didn't all change together, there is some interesting question, if at this point, for a long period you say, I'll make up an example, not for issues, but there's a very clear pattern of attributes of political figures people emphasize. So, it's very easy when you do that kind of attribute agenda setting analysis on what your categories are. I'm not sure if someone has ever looked at those over time or across countries, and that would be interesting. Then again, I would argue that is kind of framing is more sociological, in a sense of finding more cultural patterns of perception. No one has really done that. But certainly, if you have some sensitivity to you data to say to the MIP question, then you should be able to avoid just locking things into the categories that there have always been into the public opinion. But, there is certainly a major influence of public opinion research on agenda setting, because it is from where we started. I say that's the domain we started in, it's been moving to another areas now, but that's the domain, so, obviously, it shares some characteristics with that.
J: The general assumption of agenda setting is concerned to what people talk about on their daily lives. But measuring the public agenda through MIP question is one thing, and talk about media issues is another...
MM: There's actually some evidence that indeed they do. A study had been done many years ago when I was at Syracuse. What we were interested in was: In Syracuse University, freshmen, unless they were living in town, had to live in the university housing and they had a common dining room. And the question was this: what people talk about related to what they tell us in the survey. So, we did a survey and asked what was the most important problem facing the country, and we had our traditional data. Then, my graduate students get to spend an exciting week going to this dining hall at lunch and dinner, and they put their books there and were writing, and people were probably thinking they were doing their homework, or a paper. They were actually listening to what freshmen were talking about and making notes. It's a technique we've used a few times to study public opinion. There are some interesting studies done during World War II in England because English people conveniently line up at bus stops and you could listen to what they were talking about while they were waiting in the queue. Of course that 98% of what they were talking had absolutely nothing to do with public opinion and public affairs, but what they did talk about very much corresponded to what we found on the surveys, and you could, of course, study chat rooms online, in order to see what people talk about; there's some research on that, and probably the best evidence on this are the studies I mentioned the other day on consumer behavior, that people perceive "oh, the economy is the big problem", and they have pessimistic attitudes about it, they find, within a very short time, that their behavior corresponds to that, and the purchase of certain types of products begins to decrease. There's another piece of behavior with a lot of research: voting. People go vote and which way do they vote? So, I think there's good linkage between what people tell us in the survey and what they do.
J: That traditional Mass Communication Research is concerned more about measuring things and not thinking and reflecting about the findings...
MM: I guess that there are two responses, and one of them is that some people just don't like numbers, and I think that part of it comes out of that. The other, of course - and I think both some agenda-setting and some framing articles which are dealing with public issues - are much too narrowly focused. I think you need particularly in the introduction to the article to briefly make a case of what "why would anyone care about this? What do you need to know?" And it seems to me you can justify a piece of research in two ways, either if the topic itself is of considerable interest or in some ways it advances the theory. I'm not really sure if a lot of articles do that. I'm not sure this article on the media framing on capital punishment...it's kind of interesting, but why should I read this? I mean, you know why I would read it, but would I ask an undergraduate Journalism student to read this article? Probably not. So, I think sometimes I think people become so narrowly focused on their particular interest or sometimes a particular methodology. I read an article many years ago and the title was something likeTwo many hacked problems the other than explicating them, and, ok, it's yet another study of because the topic itself has very important conditions of change. I think they change in some significant way, in about every presidential election in the United States, there will be dozens of studies of the elections. From that point of view, I think it's a fair criticism. How interesting is it going to be to look at? How many articles The New York Times ran on the presidential primaries? It needs to make some more important point. You need a certain amount of that as you develop a point of view, you need some replication, but after a certain point, that's enough, let's move on to something new.
J: As the literature in agenda setting is so vast, to what extent should a person read until start his or her empirical research?
MM: Going back to the previous question: when you say to yourself: I've seen other examples of this, this is more of the same, theoretically, you are ready to stop. It's a serious problem people have. Sometimes, the students never want to finish their dissertations, because they think: `in the new issue of this, I want to read these three articles'. You may never finish! There is always a new article. At some point, you just have to cut off and, unless someone says that there's this super important new article that you have to use, read and take into account. At some point, you just need to say `I think basically I've got it'. And then start doing the research. You may want to browse some of the newer material, and you may include new literature review, but once you feel that you have the basic ideas, you should start the research. You know, I've started collecting new material for a new additional of the book and I think we've looked into four journals from 2000 through last year, and there were about 30 new articles that we have abstracts and I would guess maybe four or five of those are really important and new. The others are interesting and they update and apply to maybe slightly different areas, but if you look at the list of 30 new and think `do I need to read the 30?' That would be nice if you did, but read the abstracts, and then go read these two or three if the abstracts are interesting, but most of them we rely on the abstract. From there, I know what basically is all about.
J: Steve Reese explained the hierarchy of effects model and I asked him: `how about the audience? Does the audience influence journalists on the selection of news?' And he told me to forget about the audience.  I wonder if the Internet would have the potential to change that.
MM: I think it does and it needs to. My theory about why newspaper readership and TV news viewing are declining in the US is that journalists are too detached to the audience. If we did a simple model, most journalists see it is as producers-message-audience. For journalists, it is producers' message. That's it. I'm a message producer. Particularly elite journalists are probably out of touch. One example, let's go back to Clinton administration. They called a news conference to announce a program of expanded scholarship for graduate students, a fairly ambitious program they were trying to create, funded by the Federal Government, and it would be US$ 10.000. For a Texas student, it's a big amount of money. And the person related to the story said that they were leaving the news conference, the person heard one saying to another journalist - they were both television networks Journalists with very big salaries: "That is not an important program". US$ 10.000 is nothing. That is just a chance to gain into political attention. Particularly for the elite journalists, they live in a very unreal world.
J: To what extent has the Internet changed the way we consume information and shaped the public agenda?
MM: An interesting question to pursue is: If people have access to the Internet, what is the behavior pattern? What do they actually use? Where do they go on the Web? If agenda setting has become a more fragmented process, channels people use reflect their diversity of agendas? Here, we can go back and look at some background from the past, what we learned since the Chapel Hill study, and basically what we've learned is that across many different media agendas2 , the medium correlation between those agendas is +81, which is very high. If we tighten the comparison either by just looking at the newspapers, the correlation would go up a little bit more or if we look just at the TV or to only national newspapers, the correlations would go up some more, or to the correlation of a local newspapers would be higher. But, even when we mix, national, local news media and different formats on newspapers and television, which presents the news in slightly different ways, there's an extraordinary high level up agreement on what is the issue agenda and that's been replicated many times since. And that is really not a surprising finding because if you think about the power of professional norms and traditions in journalism as a window through the world, when they look at that window, they have a certain set of news values in mind. For those who are involved in research, I sometimes say you can think of these correlations as reliability coefficients for the applications of news values. If many different Journalists look out at the same window on the same day in the world, and apply basically the same set of news values and traditions, it's not surprising they come to highly similar answers to what we should tell our audience about today. There's a medium +.68 correlation also in a study done by Shaw's student, who content analyzed traditional and Internet news media outlets. And they agreed even more on what the top 3 stories are, with a correlation of +.78. So, in terms of diversity of agendas, it appears that to a considerable degree, we've been finding the same pattern on the Internet that we've found among traditional media that the media tend to agree with each other. The additional question, which goes back to what channels do people use, how many people make substantial use of specialized sites, which might turn their agenda to a particular direction. But the surveys, particularly in the United States, indicate that, for most people, using the Internet as news source is a very casual activity. People by large do not read the online versions of newspapers on the way they actually read the printed newspaper. What we find is that people in fact complain about reading things on a computer screen for any linked period of time more than five minutes, people by large don't find comfortable to sit there for 15, 20, 25 minutes to reading online.
St: How about blogs? They are news sources for a lot of people...
MM: They are sources for Journalists, not actually a lot of people. There's an interesting survey, I think about a year and a half ago in the United States, and they asked a very straightforward question: "How often do you read a blog on the Internet?" The most common answer given by slightly more than two thirds of the respondents was: "What is a blog?" There are some important blogs out there, but they're mostly important in the world of Journalism and political operatives. They really haven't achieved the state of a mass media. They seem to be more a creature of - the term sometimes used in the USA is their audience consists of political junkies - people who are intensely interested in politics, which, for example, I think it explains why they have been successful - particularly in this presidential primaries season- as fund raising resources. Because the people who go to the blog, who go to the candidate websites, are intensely interested in politics, and of course, they are also the ones who are more likely to contribute with money and the advantage of the bloggers or the campaign websites is that it makes easier to do that because in most instances you can pay with your credit card or bank account number and give an amount of money, and you pay your contribution, you don't have to write a check or put it in an envelop and address it and go mail it, you can do all that online. There have been a number of agenda-setting studies looking at the relationship between blogs and the media and the bulk of the evidence suggests they're very much creatures of the media, that is, their agendas are set by the mainstream media. There is a dissertation completed in Texas at the end of 2007 looking at this link between political blogs - conservative, liberal and moderate - and the agendas of The NYT and The Washington Post. And by chance, the period that Sharon Maraz did her dissertation, the period that she happened to pick - it turned out that was a month with bad news for the current administration - was the period where the attorney general, Alberto Gonzalo, was called before congress and he was a very absent minded professor, he didn't remember much of anything. What was interesting was that the conservative blogs were very quiet during that period, the liberal blogs of course were over the top, and the moderates were kind of in between. There is also an interesting study from two years ago, based on a very large Pew survey looking at the agenda setting influence of blogs and the media. And it again suggested that, by large, the influence is from media to blogs. Occasionally, you'll see spectacular kind of case studies where purely the influence went the other way, but those seem to be the exception rather than the day-to-day rule of what's going on out there on the Internet.
J: Going back to the centripetal perspective on agenda setting3, how could one define an agenda of a blog, which is maintained by a person - which can be either a professional or an expert or just an amateur - but, at the same time, it's a medium, and, depending on the kind of interaction and audience, it can be part of an environment of other mediums? I would not be able to define what kind of agenda is that.
MM: There's a strict agreement on that. There are a lot of problems on trying to content analyze blogs because many of them change during the course of the day. Maybe the best way of sampling what the blog is talking about is to sample its contents three times a day to look at it very early in the morning, then look at it again in the early afternoon, and then look at it again in late evening. You may not capture all changes because some things may come in and go out, but you tend to get the main things. That's one aspect of measuring the kind of agenda. The other aspect is that most studies includes is what hyperlinks are there, and sometimes, they include the other direction, what other blogs or media sites are hyperlinking to this blog? So, you have the content of the blog itself, its links out to other blogs, websites or newspapers. Interestingly, although you could guess what comments are, nearly all the conservative blogs extensively link to the New York Times, which they regard it as awfully liberal, but they link to it. So it's a more complex content analysis. On the other hand, it's fast, most of them are text, it's actually, in some ways, simpler than doing television, because when you get into the television, of the visual images, which you can spend hundreds of hours doing it. It becomes very difficult. For television, people ignore most of the visuals, they just analyze essentially the words and that's beginning to change. If you do visuals, you have to make sure you have very clear hypothesis in advance, if it's presented in this way, by that person, then you'll know exactly which aspects of the visuals to get. If you try to do everything that you can imagine in terms of the visual, you'll never get the project finished.
St: I have the perception that people don't read blogs when they want to get information by the media. My perception is that people go to the versions of newspapers or to Internet news media to get information, rather than a specialized media. Maybe people use personal blogs or social networks like Facebook, to be part of communities where you can interact, but if they want to get information, they just go to traditional news websites.
MM: That's generally what the evidence shows. The question becomes, then, for things like Facebook and social networks: "What's the content of those? Are they personal?" It sometimes may replace telephone conversations. If you think of older media, in the case of blogs, the specialized magazines and economic journals come, people read them very intensely, but the general public doesn't. If you think about some of the long-standing tradition of political journals in the United States, things that have been around a hundred years or more, their circulations are very tiny.
J: Floyd Allport (1937) says that the public opinion is different of the published opinion. I wonder how to distinguish public opinion on the Internet and if published opinion produces effects as well on Internet environments.
MM: That's actually very difficult. It reminds me an idea that I call "civic osmosis". The short answer to your question is `I'm not sure if it's all possible'. This is the larger answer. In a paper I started writing on this, I used a cliché metaphor, that everyone knows, that says: You can't see the forest for the trees. As researchers, we get very interested in the trees and we spend a lot of time describing the trees, measuring the trees, and sometimes forget that we have to back off and think about the forest. Some individuals sit on some trees, others individuals sit on other trees, most individuals sit on a lot of different trees. We sometimes try to focus our research too narrowly. If we go all the way back to one of the early Lazarsfeld's studies in the 1940s, one of the important findings there was that use of news media for most people overlaps. What they did was a series of cross tabulation, looking at the readership of frequency with which people read the newspapers, listened to news on the radio, to a lesser extent, read news in the magazine, but they concentrated primarily on newspapers and radio, and they found that they tremendously overlap. Most people either consume a fair amount of both or not very much of both, there weren't many specialist people saying that listened to the radio a lot, but paying not much attention to newspapers, or spent a lot of time reading the newspapers, but listening to the news on the radio. I did a paper to bring that to television, using the national election studies and replacing radio, which declined as a news medium, by that time with television, and found exactly the same pattern, that people's behavior tends to be pretty much distributive across all the news media. Another example that comes to mind about how this mixes. For the most part, if you ask people where they've found out about something, they don't really know. If you only read one newspaper or listened to news on a certain radio station and I say: `where did you find out about?'
McCombs 1
Figure 3: Maxwell McCombs
It would be a very easy question to answer. Some years ago, researchers were very aware of how this mix was and people had no idea about where they really found out about things. It was a diffusion study that Ray Funkhouser and I did. He had developed a very interesting mathematical model of how the news of an event should spread through a population, and I was going to do a study in North Carolina to test this. The idea was that we needed some anticipated news event that would appear in the news by the middle of the day. So, we could begin interviewing late afternoon, and go until late evening to see where people found out about this and how many people know about it. The news event that popped up was - this is back in the Cold War - an American submarine simply disappeared, it is presumed that it sank, and six months later, they found the wreckage of the submarine. This was announced at the news conference by the navy in Washington at noon, and we began interviewing at about three o'clock in the afternoon, and continued until nine at night. And we asked a big question: "Have you heard any news about an American navy vessel?" to see if the people knew about it, and when they heard about this. The Vietnam War was going on at that time, and ships were having all sorts of maintenance problems. Some people would have guessed it would be something about Vietnam War, but not. We came up with answers like: "We read in the news this morning, or I saw it on TV last night". Among people who knew a submarine was found, 15% of the answers were clearly invalid to where and when, as they could not have read on the newspaper in the morning, because it wasn't announced until noon. People don't keep track of where they learn things; they jumble. Sometimes people find out about an event because people tell you. So, the metaphor the osmosis is that the information acquires creeps into people's consciousness from many different sources, and they don't keep close tabs, and, in addition to those, these media agendas overlap.
J: Maybe on the Internet, the role of media is less important than the role of the environment. I've read a Reese's article saying that on the Internet, the boundary lines of professional journalism are blurring. Wouldn't it be more appropriate to study agenda-setting from an information environment perspective, rather than the media perspective?
MM: You certainly could. I could answer that in two directions. The paper I presented in Brazil in November was focused on news organizations. Are you familiar with Chris Andersen's idea of the Long Tail (2006)? I used that, and I said: news media can draw on these experts. Well, they don't have to be experts, but people who are knowledgeable on some areas particularly to expand their local news franchise. Two examples, one where you don't have to be quoted as an expert: A big dilemma always for decades for newspapers has been the sports section. There are too many sports, played at too many different levels, and very few people are interested in all of them. Most people are interested in certain ones. So, even if you have a large sport section, 12, 14, 16 pages, still many things you can't publish. Then, - it's been an Austin example - there's a very large park near where we live. If you drive through there on Saturday, there are about six or eight soccer games going on. These are organized leagues of children of different ages, starting at about eight o'clock in the morning, and going on until about 4 o'clock in the afternoon. This is just one part, and there are several other places, so there is an immense amount of soccer going on. Obviously, not all of that are on the newspapers, they don't have space for that, but if you take the concept of the long tail, and you put your sports online - you can put all that online, there is plenty of room for it - you could find plenty of parents and grandparents who are quite willing to write up the results of those games, you need an editor to browse, to make sure weird things don't get online. So, you can take the whole area of amateur sports, you could tremendously expand that coverage, and these people have to be in the events, but they don't have to be experts sports journalists to do this. Moving to a more quote expert area, perhaps an example of what Steve had in mind: there is a river that goes to the middle of Austin, and then downtown, and the university going north of the river. South of the river used to be totally residential, there are a lot of musicians and composers and most of them used to live there. As the city has grown larger and larger and more people want to be closer to downtown, when we first bought our house - we've lived just across of the south of the river - people would ask my wife: `Where do you live?' `We live in south Austin' and it had a distinct field. About ten or twelve years ago, the newspapers reported on things in our area, and then we lived in Central Austin. In the last two or three years now they're all building these big condominium developments, and my line is now one more economic boom and people ask: `Where do you live?' `We live downtown'. Well, downtown is encroaching on what was a very nice middle class residential area, right adjacent to this park, and there's a lot of concern about, particularly, I mean, some of the condos are only five or six stories high, but there a number of post of 15 or 20 story buildings in what have always been residential areas. So, on the list service of our neighborhood, for the last two or three years, there has been a lot of material about the zoning of our neighborhoods, for what kind of buildings could be there, the history of areas projects, of various developers who are pushing these projects. In most instances, this is not their profession, I think for one or two of them, they are essentially urban planners, others simply have just an intense interest in this area, and they have become experts. So, yes, that kind of material is on the Internet, and they feel really concerned with a proposal for a certain kind of imminent project in our area, you're more likely to turn to the neighborhood list service than the local newspaper. Part of my argument in this article is the newspapers need to tap into this expertise. My guess would be, across Austin, there are probably 15, 20, 25 of these neighborhood list services, there may be even more, and a lot of what is on the list service are like: `could you recommend a good painter for my house?' `Where should I take my car to be repaired?' But a certain amount of that material is serious discussion. This isn't really Journalism, but it's useful. Whenever any significant crime occurs in the neighborhood, it's there. There were a whole series of articles that were talking about a number of burglaries and attempted burglaries in the next neighborhood over, and then, a long report from someone who obviously was knowledgeable about this. I mean, they gave the address and the location that had been ruined and there seem to be a lot of people coming and going. In effect they said the police had this house under surveillance if they thought it was a site for drug dealing. Part of the related robberies in the area is from people who needed money to buy drugs. This clearly is evolving the Internet, and I think it has the potential for news organizations to organize that. Let's say if you and I are the editors for neighborhood news in Austin, which means mainly we sit by the computer all day browsing each list service and of course, once you hit the pattern, you are able to choose what is important or what it's new, and then, you move that out, you may need a little bit of editing, you may need to call that person and say "I have some questions", and get the reaction of the city government about that. So I think it's an open question: how much does that get integrated into journalism? It's not necessarily Citizen Journalism, as that sometimes has been talked about, but it certainly has the potential to become that. And, in this case, I think people are genuinely experts.
J: I wonder if public agenda could be measured through network analysis...
MM: That's clearly the direction to get. Network analysis in terms of sociological components that fit together. It can also be brought with other names, in terms of content analysis. Most of the content analysis that you find in both framing studies and agenda-setting studies is pretty traditional, they use computerized content analysis, it just plucks keywords out, it's just about using frequencies. It is really using the capability of the more sophisticated content analysis to look at the Internet connections among for instance these attributes, looking at the ranking others, but I think we can go beyond that and kind of begin to say `how these things link to each other and how do they link back to objects?'
J: Agenda-setting began as a reaction to selective exposure, but I feel that, through network analysis on the Internet, it goes back to the perspective to which was initially opposed to, because people go to search what they want, they have more control about what information to choose.
MM: We may find a substantial amount of that. There is this kind of interesting ironies, as it was a reaction against attitudes and opinions, we've come right back to that, it was a reaction against selective perception, I think we're finding some of that certainly on websites, and that becomes an interesting question to be asked to what extent do people lock themselves into a pretty close circle. Sometimes, we refer to the expression The Daily Me, the newspaper that has only what I want to be in. Most people at least want some exposure to a broader wave of news because of certain things; you can't say `I want news about' because you don't know what it's going to happen. So, most people leave the door open about it, but I think it's an interesting phenomena.


1The interviewer is a Ph.D. researcher from the Post-Graduate Program in Communication and Contemporary Culture at the School of Communication of Universidade Federal da Bahia, as well as a professor of Theories of Communication, Theories of Journalism and Methodology of Scientific Research. Sponsored by Capes, her Ph.D. investigation is concerned with the development of research designs to investigate agenda setting on the World Wide Web. At the time of the interview, she had been doing a doctoral stage at Universidad de Navarra (in Pamplona, Spain).
2McCombs talks about the matrix correlation and the mediums they have analyzed.
3McCombs makes a distinction between two opposite trends in agenda setting research: the first one is the centripetal inwards perspective, in which agenda-setting basic concepts continue to be explicated; and the second is the centrifugal outwards perspective, in which agenda-setting theory tends to be expanded.