The Mass Media

Setting the Agenda
The Mass Media and Public Opinion

Maxwell McCombs

polity

Copyright O Maxwell McCombs 2004 T h e right of Maxwell McCombs to be identified as Author of this Work has been asserted in accordance with the UIC Copyright, Designs and Patents Act 1988. First published in 2004 by Polity Press Polity Press 65 Bridge Street Cambridge CB2 lUR, UIC. Published in the United States and Canada by Blackwell Publishing Inc. 350 Main Street Malden, MA 02148, USA All rights reserved. Except for the quotation of short passages for the purpose of criticism and review, no part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the publisher. ISBN: 0-7456-23 12-3 ISBN: 0-7456-23 13-1 (paperback) A catalogue record for this book is available from the British Library and has been applied for from the Library of Congress. Typeset in 10.5 on 12 pt Plantin by Kolam Information Services Pvt. Ltd, Pondicheny, India Printed and bound in Great Britain by MPG Books, Bodmin, Cornwall For further information on Polity, visit our website: www.polity.co.uk

How Agenda-Setting Works

The agenda-setting effects of mass communication are widespread. Observations have found agenda-setting effects all across the United States in a variety of small towns and large cities. These effects also have been found abroad in cities as diverse as Tolyo, Japan, and Pamplona, Spain, and in countries as different as Argentina and Germany. Altogether, there are now more than 400 empirical studies of agenda-setting, many following the original Chapel Hill example and conducted during political campaigns, others monitoring public opinion in non-election periods. There is considerable diversity i n the public issues that have been examined over the past thirty-five years, a diversity encompassing the economy, civil rights, drugs, the environment, crime, a wide variety of foreign policy questions, and dozens of other public issues. Agenda-setting is a robust and widespread effect of mass communication, an effect that results from specific content in the mass media. For many persons, one of the most surprising aspects of these wide-ranging effects is the tremendous variability of the geographical and cultural settings in which agenda-setting by the media occurs. The culture and politics of the United States are exceedingly different from the cultural and political setting of Pamplona and the province of Navarra in northern Spain, where numerous agenda-setting effects have been measured in recent years. There is even more of a cultural and political contrast when we shift from Western countries t o rhe young democracies of East Asia, where agenda-setting effects also have been observed. A few years ago, a seminar in Taipei discussed this widespread international replication of media effects originally found in t h e United States and came to the conclusion that agenda-setting effects

How Agenda-Setting Works

37

- the successful transfer of salience from the media agenda to the public agenda - occur wherever there is a reasonably open political system and a reasonably open media system. Arguably, there is no perfectly open political system in any country in the world today, n o system where the principle of one person, one vote, fully applies to every adult in the population. But the political systems of the United States, Spain and Taiwan - to cite some countries previously mentioned - are reasonably open in that elections really matter and actually determine the course of political history. Moreover, the vast majority of adults are eligible to participate in these elections. The media systems of these countries - or at least significant portions of them - are also open in that they are independent sources of news and political expression free from the domination of the government and major political parties. Where both of these conditions of openness exist, the public accepts considerable portions of the issue agenda put forward by the news media. Observations made during the 1994 Taipei mayoral election in ~ a i w a nunderscore the validity of this axiom that explains the wide' spread occurrence of agenda-setting. At the time of that election, there were three television stations serving Taipei, and all three in one way or another were controlled by the government and the longdominant KMT political party. At the extreme, the Department of the Navy held 40 per cent of the shares in one of these television stations. Not surprisingly, no agenda-setting effects were found for television news. To echo a signature expression of American political scientist V. 0. Key, albeit in a vastly different cultural setting, 'The voters are not fools!' In contrast, significant agenda-setting effects were found for the two dominant daily newspapers in Taipei. Although these newspapers, like most newspapers around the world, favour a particular political perspective, they are independent businesses free of any direct control by the Taiwan government or the KMT. This Taipei example is a useful comparison of the influence of open and closed media systems where all the other political and cultural factors are essentially held constant.

Evolution of issue agendas
An ongoing stream of public opinion evolves in these civic arenas around the world that are defined by open political and mass media systems. Over time, the salience of individual issues rises and falls as the attention of the mass media and the public shifts, Here we will outline major aspects of this public opinion process as issues appear

38

How Agenda-Setting Works

on the media agenda and then move to the public agenda. We consider the capacity of the public agenda and competition issues for a place on this agenda, the time-span that is involved in tfie evolution of the public agenda, and the comparative roles of newspapers and television news in the agenda-setting Process. The intense competition among issues for a place on the agenda is the most important aspect of this process. At any moment there are dozens of issues contending for public attention. But no society a n d its institutions can attend to more than a few issues at a tinle. T h e resource of attention in the news media, among the public, a n d in o u r various public instiltions is a very scarce one. One of the earliest insights about agenda-setting was the limited size of the public agenda. For many years, a statement that the public agenda typically included no more than five to seven issues at any moment was and regarded as another accepted as an empirical instance of what the psychologist George Miller called the 'magic number seven plus or minus two', a sweeping empirical generalization that describes the limits of a wide variety of sensory ~ r o c e s s e s . ~ The accumulation of evidence over subsequent years suggests a n even smaller limit. Only a few problems demonstrate any sizeable constituency among the public when the Gallup Poll asks a national sample of Americans, 'What is the most important problem facing this country today?'. Across the ten Gallup polls asking this m o s t important problem (MIP) question from 1997 to 2000, only half found a public agenda on which as many as five issues had a constituency of 10 per cent or more. Ten per cent is the level of concern among the public that has been identified as the threshold for signifi.~ cant public a t t e n t i ~ nFive issues, of course, is the bottom of the range for Miller's axiom. Across all ten polls, the public agenda ranged from two to six issues. This tight constraint on the size of the public agenda is explained by the limits of the public's resources, limits that include both h e and psychological capacity. Limits on the size of most media agendas are even more obvious, a limited amount of space in the newspaper and a limited amount of time for broadcast news. Even in the case of internet web sites, with their seemingly unlimited capacity to a d d Pages, the public's attention span and available time impose severe constraints. All these constraints on the agendas of public issues wimn a society at any moment are summed up in the idea of the agendasetting Process as a zero-sum game, a perspective that underscores the intense com~etition among issues for attention by the medin the public. Explicit evidence that the public agenda is a zero-sum

How Agenda-Setting Works

39

game was f ~ ~ in an analysis of three issues that dominated the US n d public agenda in the early 1990s." The salience on the public agenda for each of the dominant issues at that time - the Gulf War, economic recession and federal budget deficit - resulted from two factors. One factor, the one emphasized by agenda-setting theory, of course, is the pattern of news coverage. The second factor is the relative salience of the two competing issues on the agenda. Historically, one result of this limited agenda capacity and intense competition among issues is that a few perennial concerns have held centre stage in US public opinion. In the years immediately following World War 11, foreign affairs and economics occupied centre stage, with foreign affairs nearly always in the leading role. Although other issues were able to garner a constituency from time to time, this pair of issues dominated the US public agenda.5 Updating this portrait, another analysis of trends from 1954 to 1994 - based on responses to the 'most important problem' question in 140 Gallup polls - found no change in the capacity of the Arnerican public agenda despite major increases in the level of formal education among the American public. During those years the proportion of the population with a high school education increased from 34 per cent to 78 per cent, and the proportion with a college degree rose from 6 per cent to over 21 per cent. Despite the lack of change in the capacity of the public agenda during this time, there is evidence that rising levels of education impacted another aspect of these trends in American public opinion, the diversity of issues on the public agenda.6 In The Reasoning Voter, Samuel Popkin observes: Education affects politics not by 'deepening' but by broadening the elect'orate - by increasing the number of issues that citizens see as politically relevant, and by increasing the number of connections they make between their own lives and national and international event^.^ His observation acknowledges the widely documented situation that most people, even highly educated persons, rarely possess detailed, in-depth knowledge of public issues. Persons with higher levels of education do read newspapers and discuss the news more frequently with their family, friends and co-workers. The principal outcome of this activity, notes Popkin, is that educated persons 'will have limited information about a wider range of subjects, including national and international events, that are further from daily-life e ~ ~ e r i e n c e . ' ~ This broadening effect of education on the public agenda is readily apparent in the growing diversity of issues found on the public

40

How Agenda-Setting Works

agendaVg previously noted, at the time of World War I1 and in the As Post-war years up to 1960, a single categoIy, inteImati0nal affairs, largely dominated the public agenda. But in the next two decades, the 1960s and 1 9 7 0 ~ ~ a larger array of issues were prominent. International issues were still on the agenda, principally Vietnam and the Cold War, but there also were iarge ~0nstitUenciesfor economic issues as well as for civil rights. In the final two decades of the twentieth century, the public agenda continued to broaden and diversify. Four major issues each claimed more than 1 0 Per cent - jobs, personal economic issues, law and order, and international affairs. Another four minor issues concerned with other aspects of the economy and domestic issues each claimed a 5 to 10 per cent share. T h e American public agenda reflected a growing sensitivity to a broader array of issues. H o w can this increasing diversity of the public agenda be reconciled with the evidence that its capacity remained constant? T h e answer is that some issues now move on and off the public agenda faster than in previous decades. I n other words, the explanation that reconciles these aspects of the agenda-setting process is that a collision between the expansive influence of education and the restrictive influence of limited agenda capacity has resulted in a more volatile public agenda. At mid-century, one category, international affairs, dominated centre stage. But the cast of issues began to grow in the 1960s, and this trend for major issues to share the stage with minor issues, at least for brief intervals, has continued. As we see in box 3.1, the long-reigning divas of public affairs continue to get starring roles on the public agenda, and the duration of their time on stage is lengthy, often exceeding two years or more. But now they share the limelight from time to time with an array of minor issues. These minor issues, such as the environment, education and health, do not appear nearly as often, nor is their duration o n stage nearly as lengthy. But they do appear despite the limited capacity of the public agenda, the result of education's broadening influence o n public perspectives about the issues of the day. Further insight into the role of formal education in the agendasetting process comes from a comparison of five demographic characteristics that appear time and again in public opinion polls: age, education, income, sex and race. Using a mix of issues that received either very high or very low coverage in the local newspapers, their salience was examined among nearly a thousand Americans in three communities stretching diagonally across the United Stares from Florida to the Pacific Northwest.'' Only a single demographic characteristic was related t~ the pattern of salience for these issues.

How Agenda-Setting Works
BOX 3.1

41

Duration of tTIaj0r issues on the public agenda Average duration per cycle (in months)* Number o cycles, f 1954-94+
7 8 4 13

Personal economic issues Politics and government Asia General foreign policy issues Government spending Russia and Eastern Europe Jobs General economic issues Law and order Technology

47.4 40,8 27.8
25.2 21.8 19.3 15.1 14.0 10.3
8.7

5
4 14 5

12 3

*A cycle is the period of time beginning when 10% or more of the responses to the MIP question first name this issue and continuing until this issue I named in less than 10% o s f

the responses. Source: Maxwell McCombs and Jian-HuaZhu, 'Capacity, diversity, and volatility of the public agenda: trends from 1954 to 1994', Public Opirrion Quarterly, 59 (1995), pp. 495-525. Details of the specific issues mentioned over time that fall in these ten f categories are in Appendix A o the article.

Citizens with more years of formal education more closely mirrored t h e media agenda. This primacy of t h e educational experience is striking throughout the realm of politics and public affairs. Whether one is dealing with cognitive matters such as level of factual information about politics or conceptual sophistication in assessment; or such motivational matters as degree of attention paid to politics and emotional involvement in political affairs; or questions of actual behavior, such as engagement in any of a variety of political activities from party work to vote turnout itself; education is everywhere the universal solvent. ' I Education has the conjoint effect of increasing individuals' attention to t h e news media and sensitizing them t o a wider range of issues appearing in the news. O n the other hand, higher levels of education do n o t appear to increase individuals' defensive responses to the pattern of emphasis in the news. Well-educated persons d o not show any greater tendency than less-educated persons to argue against o r erect psychological barriers t o acceptance of t h e media agenda.''

42

How Agenda-Setting Works

However, one must be careful not to overstate the role of education and individual differences in the agenda-setting process. TO define further the role of education vis-i-vis the messages of the media i , determining the public agenda, the salience of four issues among the American public between 1977 and 1986 - inflation, unemployment, international problems and government spending - was cornpared to the pattern of coverage on national television during the same tenyear period.13 Shifts in the salience for each of these four issues was examined separately for population subgroups that were simultaneously defined by both education and family income. T h e salience of all four issues was expected to be higher among the more educated, Family income was also used as a measure of issue sensitivity t o these four issues because inflation and unemployment were assumed to be less relevant and international problems and government spending more relevant to higher income families. There were massive shifts in the salience of these issues between 1977 and 1986. Each issue displays a pattern of peaks and valleys, rising and falling sharply both on the media agenda a n d among all the income and education subgroups. In contrast, the differences among the demographic subgroups themselves are minimal. Specifically in terms of the fit between the salience of each issue on the media agenda and its salience among the public, for three of the issues inflation, unemployment and international problems - all the demographic subgroups followed a similar trajectory over time that paralleled the number of TV news stories. There are significant demographic differences, but, in statistical terms, individual differences defined by education and family income accounted for only 2 per cent of the variance in salience while the wide swings from year to year accounted for 37 per cent. 'In other words, media agendasetting effects are not manifested in creating different levels of salience among individuals, but are evident at driving t h e salience of all individuals up and down over time.'14 Finally, there is an important footnote on the lack of agenda-setting effects for the fourth issue, government spending. In the final three years of the decade that was examined, the salience of government spending rose sharply and remained at high levels among most sectors of the public despite a low level of attention in television news. Pan of the explanation for the high salience of this issue among the public may come from what we know about the limited capacity of the public agenda and about the recurring appearance of some issues. During those fmal three years, 1984 through 1986, the salience of two other aspects of the economy, unemployment and inflation, were low o n both the media and the public agenda. Recall

How Agenda-Setting Works

43

that unemployment was one of the reigning divas on the public agenda during the last half of the twentieth century and that inflation also made frequent appearances. Government spending is best described as one of the minor issues that make an occasional appearance on the stage of public opinion. Its move to centre stage during 1984-6 may well have occurred because both unemployment and inflation were offstage during much of this period. This again calls attention to the powerful constraints on the size of the public agenda.

Tirneframe for effects
The old hypodermic theory viewed media effects as essentially immediate. I n that view, media messages were injected into the audience much as medical injections are administered to patients and typically achieve rather quick effects. Support for that view essentially disappeared with the accumulation of empirical evidence in the 1940s and 1950s, a body of evidence summarized in Klapper's The E8ects ofMass Communication as the Law of Minimal Consequences. In response, scholars such as Wilbur Schramm asserted that the truly significant effects of mass communication were likely to be very long term, much as awesome formations of stalactites and stalagmites in caves are created drop by drop over eons of time. Against that background, how long does it take for media attention to a n issue to translate into high salience for that issue on the public agenda? Does it really take the psychological equivalent of eons? Or could it be that the shifc from attitude and opinion change to earlier points in the communication process, such as focus of attention and perceived importance, yields evidence of relatively short-term media effects? Recall that the rise and fall of public concern about civil rights in the United States across a 23-year span of time reflected primarily the pattern of media attention to this issue in the preceding month.15 Further examination of that evidence adds a few details, but yields the same conclusion: agenda-setting effects are far from instantaneous, but they are relatively short-term. Comparing the salience of civil rights on the public agenda across twenty-three years with the number of civil rights stories on the front pages of the New York Times in each of the six months preceding those Gallup polls reveals a monotonic pattern of declining correlations across those prior six months. The strongest correlation was +0.71 for the month of news coverage immediately preceding the polls. For the second month prior to the polls, the correlation was +0.70. Continuing back through the third, fourth, fifth and sixth

44

How Agenda-Setting Works

~ months, the values continued to decline, reaching a low of t 0 . 3 for the sixth month. This pattern might suggest some significant mulation of influence and contribution by the news coverage in each of the preceding months. But when partial correlations Were calculated in which the influence of each month on the public salience of civil rights was determined with the influence of the other five months eliminated, the results clearly indicated the primacy of the first month of news coverage preceding the public opinion polls- The correlation of +0.7 1 for the first month preceding the polls remained unchanged while all the other correlations substantially declined. T h e highest value for any of the other partial correlations was +0.48. Taking the analysis one step further, use of all six rmnths together to predict the salience of civil rights on the public agenda produced a correlation of +0.89, but the first two months alone produced a correlation of +0.84. The difference is not statistically significant, The conclusion from this long statistical journey is that, although the strongest relationship between the public agenda and t h e media coverage of civil rights was found for the single month immediately preceding the measurement of public opinion, addition of t h e second month preceding the polls did result in a modest increase in the correlation. Adding the other months did not increase the strength of the relationship at all. In short, the accumulation of news coverage over a two-month period accounted for the trend over these twentythree years in the salience of the civil rights issue among the American public. Within that two-month period, the month of news coverage immediately prior to the polls exerted the most influence. There is, of course, the question of how generalizable this picture is of the agenda-setting process. We already know that the strength of agenda-setting effects can vary from issue to issue. However, two other investigations of the timeframe for agenda-setting effects also suggest that the span of time involved in the transfer of issue salience from the media agenda to the public agenda is generally in the range of four to eight weeks. A longitudinal analysis of the public opinion trends for each of three major issues during the 1960s a n d 1970s pollution, drug abuse and energy - found a median correlation of +0.66 between the public agenda and the national television news agenda of the preceding month.16 A three-wave panel study found a median correlation of +0.77 between the salience of the environment among the public and the agenda of three local newspapers during the preceding two months. l7 Our confidence that the public agenda typically reflects the media agenda of the preceding one to two months is enhanced by both the strength and the high degree of convergence among the correlations in all three investigations,

How Agenda-Setting Works

45

which included both newspapers and television news and a variety of issues. Under conditions of high personal involvement in the news, the timeframefor agenda-setting effects may be even shorter.'' A particularly lively aspect of the internet, electronic bulletin boards where individuals discuss public issues, was monitored during the 1996 US presidential election. The frequency of discussion from September until a week after the November election for four issues -immigration, health care, taxes and abortion - was compared with the pattern of news coverage on these issues in the New York Times, Reuters, Associated Press, CNN and Time magazine. Discussion of immigration responded immediately to news coverage. Discussions of health care and taxes had longer timeframes, but still the effects were evident within a week. Among the four issues examined, only the discussions of abortion were not linked to the pattern of news coverage, an outcome most likely linked to the highly controversial and emotional nature of this issue. For the three issues where the pattern of media coverage did influence the salience, the timeframe was much shorter than for the agenda-setting effects of traditional news media. This outcome is not surprising because electronic bulletin boards are an outcropping of the public agenda where persons with high interest in an issue respond behaviourally. All this evidence about the timeframe for agenda-setting effects is based on analyses tracing the salience of individual issues on the public agenda across time, analyses that are designated in the Acapulco typology as the natural history perspective. Obviously, there are other perspectives to consider, notably the competition perspective that takes into account the full array of issues competing for positions on the agenda. While it is useful analytically to examine a single issue in order to understand the process underlying its natural history, the competition perspective provides a portrait of the real world where there is always a mClange of issues in flux. From this perspective, what timeframe links the media and public agendas? The comprehensive body of evidence summarized i box 3.2 is n based on an agenda of eleven public issues and a range of news media from local T V news and the local newspaper to national T V news and the weekly news magazines.lg Although there are variations across these five news media in which weeks of news coverage show the best fit with the public agenda, the variation is relatively small and falls in essentially the same range observed for individual issues. The range of time-spans producing the optimum match between the media and public agendas is one to eight weeks, with a median span of three n weeks. I every case, the agenda-setting effects are sizeable.

46
BOX

How Agenda-Setting Works
3.2

Time-spans for appearance and disappearance of agenda-setting effects Time lag (in weeks)* Maximum correlation -1-0.92 $0.91 $0.88 $0.60 -t0.58 Decay of effect (in weeks)** 8

News medium National TV news Local TV news Regional newspaper Local newspaper News magazine

1
2 3

12
26 26 26

4 8

"Number of accumulated weeks producingthe maximum correlation between the media agenda and the public agenda. "Number of weeks before the disappearance of a significant correlation between the media agenda and the public agenda. Source: Wayne Wanta and Y. Hu, 'Time-lagdifferences in the agenda setting process: an examination of five news media', International Journal of Public Opinion Research, 6 (1994), pp. 22540. Reproduced by permission of the World Association for Public Opinion Research and Oxford University Press.

If our benchmark is the four to eight weeks typically found for the natural history of issues, the distribution in box 3.2 based on a competition peEspective falls towards the shorter end of that range. However, another look at issue competition under two very different sets of circumstances - the agenda of issues during a US national election that included the highly salient issue of Watergate and another agenda that included the Egypt-Israeli War among college students during a non-election period - found time-spans falling i t the other end of the distribution, eight weeks or more." Nevertheless, considering the complexity of the situation here - the full array of issues on the media and public agendas - the time-span for the appearance of significant agenda-setting effects is still rather short. Over the course of a relatively few weeks, the salience of topics featured in the news media is absorbed by significant numbers of the public. This continuous and virtually invisible learning process is an important civic instance of a larger phenomenon, incidental learning from the mass media. Long ago, Paul Lazarsfeld described mass communication as an informal classroom where the students continuously come and go and, much like some students in more formal classrooms, do not always pay full attention even when they are present. But people do learn from the mass media. They learn a panoply of facts, many of which they incorporate into their images and attitudes about a variety of objects. They also learn about the most important issues of the moment, incorporating the agenda of the mass media into their own agenda of the key issues facing society.

How Agenda-Setting Works

47

The circumstances of this incidental learning differ from the learning that commonly takes place in school, but the outcomes can be just as powerful and influential. After all, if incidental learning did not yield significant outcomes, the vast advertising industry would not exist. The mass media are teachers whose principal strategy of communication is redundancy. Over and over again, our mass media teachers repeat topics, at times with great emphasis, at other times just in passing. It is primarily the accumulation of these lessons over the period of one to eight weeks that is reflected in the responses of citizen students when we inquire about the most important issues facing the nation. Of course, in most cases the lessons did not begin abruptly eight weeks previously, but it is the pattern of coverage in the most recent weeks that has by far the greatest impact on the public. There is also empirical evidence about the reverse side of the learning coin, the decay of information and the forgetting that takes place for any pattern of learning. Without delving into this aspect of learning in as much detail as we have spent on the acquisition of information, we see in box 3.2 that the time-span for the decay of learning lacks the tight focus reflected in the evidence about acquisition of current public concern^.'^ The decay of agenda-setting effects, which is defined in box 3.2 as the point in time where significant correlations between the media agenda and the public agenda disappear completely, ranges from eight to twenty-six weeks. Not entirely without surprise, each lesson learned in the mass communication classroom is visible over a considerable expanse of time. These conclusions about the duration of issues on the public agenda, both the learning process involved in the rise of issues in public attention and the decay of learning as issues disappear from public attention, are essentially empirical generalizations. We know about these timeframes because the logs of various social scientists' exploratory voyages into this realm yield rather consistent data, especially in regard to the rise of issues on the public agenda. But empirical generalizations are less sound than empirical findings that support carefully reasoned hypotheses grounded in an explicit theoretical context. In this regard, the status of agenda-setting theory differs little from the larger literature on media effects. Examination of the indices of two comprehensive and widely used texts on communication theory reveals scant attention in the field at large to the question of timeframes for various media effects.22 This is simultaneously a theoretical deficit and an opportunity for advancement.

48

How Agenda-Setting Works

There are the beginnings of a theoretical framework for agendasetting effects in an early discussion of time-related concepts.23These concepts include the time lag between the appearance of an issue o n the media agenda and its appearance on the public agenda as well as the optimal effect span, the length of time yielding the peak association between the two agendas. There is also a larger theoretical framework in the idea of the agenda-setting process as a zero-sum game.24However, considerable work remains to be done.

Newspapers versus television news
Longstanding concern with the effects of the mass media has frequently been accompanied by a certain fascination with the relative power of the various mass media to achieve those effects. With the advent and widespread diffusion of television in the latter half of the twentieth century, attention has been directed particularly at comparisons of newspapers and television. Probably the best-known example of this topic on the intellectual agenda is Marshall McLuhan's The Medium is the Message. Examinations of the agenda-setting influence of the mass media have been no exception. Once people understand the basic idea of agenda-setting, they are usually quick to ask which medium is more powerful in setting the public agenda, television or newspapers. The best answer is, 'It depends.' Whether there is any difference at all in the influence of the two media or whether one clearly surpasses the other in impact varies considerably from one situation to another. A moment's reflection will reveal why this is hardly surprising. After all, the evidence on the agenda-setting influence of the mass media covers a wide variety of situations in terms of geography, historical periods, political settings, the array of issues, and news organizations. Furthermore, the evidence that we do possess - despite its considerable breadth - in no way represents any systematic sampling of these varied situations. To complicate matters further, not every agenda-setting study even affords a comparison of newspapers and television. Many investigations use a single medium as a surrogate for the news agenda, relying upon the well-established assumption of a high degree of redundancy across the news agendas of individual media.25 In the United Stares, the New York Times has frequently been assigned this surrogate role. At other times, the national television networks have been assigned this role because of the Vanderbilt University Archive's exhaustive abstracts of network news programme content. Yet other invesdga-

How Agenda-Setting Works

49

tions merge the agendas of print and electronic media to create a single media agenda. Our measures of the media agenda reflect many variations. Under these circumstances, what can we say about the comparative agenda-setting influence of newspapers and television beyond 'It depends'? We can begin with a broad generalization grounded in decades of accumulated evidence on agenda-setting effects. Then a variety of evidence will be reviewed that illustrates this empirical generalization. Finally, some theoretical suggestions for solving this intellectual puzzle about the comparative roles of the various mass media will be advanced. First, as a broad empirical generalization, about half the time there is no discernible difference in the agenda-setting roles of newspapers and television news. The other half of the time, it appears that newspapers have the edge by a ratio of about 2 to 1. This latter pattern is a surprise to many because the conventional wisdom is that television is the dramatic, powerful medium of our time. But if we briefly consider the natural history of most issues, the larger capacity of daily newspapers relative to television news means that audiences often have a longer period of time to learn the newspaper agenda. Television news is more like the front section of the newspaper. Newspaper readers may have considerable exposure t o an issue long before it ever reaches the top of the newspaper agenda or appears on the television agenda at all. In the international setting, there is an additional factor that fiequently places newspapers at an advantage over television. In many countries, all or a considerable portion of the television service is controlled to some degree by the government whereas almost universally, newspapers are in private hands. This is a situation that can diminish the credibility of television news. As we turn to the empirical evidence, another complicating factor must be recognized. In many situations, especially in the United States, which is the setting for much of our evidence on agendasetting, comparisons of newspapers and television involve a mix of levels. T h e original Chapel Hill study involved two national television networks, four local newspapers and the New York Times, the closest approximation to a national newspaper in the United States until the appearance in the final decades of the twentieth century of USA Today. Local television news was not included in that seminal 1968 examination of agenda-setting, but other investigations do take in news agendas based on local television news. As the examples that follow illustrate, just about every possible combination of local newspapers, regional newspapers, the New York Times,local T V news, and

50

How Agenda-Setting Works

some mix of the national television networks' evening news can be found. Beginning with the original Chapel Hill study, the general Pattern found is that the agenda-setting influence of the New York T h e s was greater than that of the local newspaper, which, in turn, was greater than that of the national television news.26 But there are overlaps in the correlations that summarize this media influence on the public agenda. The differences between the various media are not totally distinct. Moving on to the 1972 US presidential campaign, an exten. sive set of comparisons again favoured the local newspaper over national television news by a wide margin.27 But in the next presidential election, comparisons of local newspapers in three cornmu. nities with national television reversed the pattern: national television news was more influential in 1976 than local newspapers.28 Th;, pattern was also replicated in the 1980 US presidential election.29 Outside an election setting, the evidence is equally mixed. Cornparisons of local newspapers and local television news in Minneapolis30 and ~ouisville~l favoured newspapers. In Toledo, for the agenda of national issues, both national and local television news had greater influence than the local newspapera3' For an agenda of local issues, the local newspaper had greater influence than local television news. Finally, revisiting box 3.2's summary of the findings from three American communities, the local newspaper trailed national television news, local television news and a regional newspaper.33 Among the latter three, there was no significant difference. As we said, 'It depends.' A major portion of the answer to the subsequent question, 'Depends on what?', is discussed in the next chapter, which presents a psychological explanation for agenda-setting effects. By way of preface to that discussion, it must be noted that both the frequency and the quality of attention to the news media differ considerably from individual to individual. Some people are exposed to television news and a newspaper nearly every day. Many others are exposed far less frequently. Looking at the public's motivation to attend to the news media, about a fourth of the American population strongly believe it is their civic duty to keep up with the news.34 The vast majority of these dutiful citizens read a newspaper daily and nearly half view national television news daily. As the strength of this civic belief declines, so does regular exposure to any news medium. Whatever the pattern of exposure, there are also individual differences in the level of attention. While few students in the media classroom prepare for an election, much less everyday conversarion, in the way that graduate students prepare for their comprehensive

I

How Agenda-Setting Works

51

some of these citizen students do pay considerable to election news and news about the issues of the day. One analysis found that serious newspaper readers, defined as persons who regularly read national and international news, local government news and political news, account for about a fourth of the total newspaper audience. Albeit a minority, it is a sizeable and well-informed minority.35 Closer examination of individual differences in why people come to the news may explain differences in frequency of exposure, degree of attention to public affairs news in the media, and the variable pattern of outcomes found in the comparisons of newspapers and television. A theoretical explanation, perhaps grounded in the ideas about audience psychology that are presented in chapter 4, is needed to bring greater meaning to what is at best a vague empirical generalization about the agenda-setting influence of various news channels. At the societal level, even more important than the relative impact of newspapers and television is the sheer diversity of news sources available to the public. Diversity in the public agenda - measured by the number of different problems mentioned by people when asked to name the most important problem facing the local community or the nation - is significantly related to the number of newspaper, radio and television voices in the ~ o r n r n u n i t y . ~ ~

Summing up
Citizens are involved in a continuous learning process about public affairs. Their responses to the pollster's quiz about what are the most important issues typically reflects the media's lessons of the past four to eight weeks. Sometimes it is the newspapers' lessons that dominate these responses, but often both newspapers and television news share equally the role of civic teacher. T h e content of these media lessons also reflects the intense competition among issues for a position on the media and public agendas,37 a situation in which only a small handful of perennial issues frequently have the spotlight. The agenda-setting effects that are frequently the outcome of this complex process are shaped to a considerable degree by characteristics of the media's messages and to a far lesser degree by the characteristics of the recipients of those messages. Mass communication is a social process in which the same message, in either printed, audio or audiolvisual form, is disseminated to a vast population. Numerous characteristics of these messages influence how many persons pay any attention to the message and apprehend at least some portion of its

52

How Agenda-Setting Works

content. Front-page stories in the newspaper have about twice the readership of stories that appear inside the newspaper. Stories with attractive graphics and large headlines attract more readers. Many other characteristics of the newspaper - and analogous characteristics of television and other mass media - influence the extenr t o which mass communication is successful in finding an audience. Ultimately, mass communication is a transaction between a single member of the audience and the media message, a transaction in which individual differences might seem paramount. In a sense, mass communication effects are a large set of overlapping, b u t nor idenrical, personal experiences. Although, arguably, no two of these experiences are identical, it is fortunate for our goal of a parsimonious theory of mass communication effects that persons with vastly different characteristics frequently have highly similar experiences. Among all these individual differences, the most important ones a r e the basis of the psychological explanation for the occurrence of agenda-setting that is presented in the next chapter.