Learning Center
Plans & pricing Sign in
Sign Out

mass media as sentinel


									                           The mass media as sentinel:
               Why bad news about issues is good news for participation

                                         Paul S. Martin


        This article argues that negative news coverage of politically relevant social issues
stimulates political participation by shaping citizen awareness of collective problems and interest
in politics. By drawing citizen attention to social problems that government may attend to, the
press acts as a sentinel for the mass public, cuing them to periods when participation is more
important. Drawing on an analysis of the 1974 National Election Study in combination with the
Center for Political Studies’ content analysis of newspapers, I find evidence that bad news about
issues is good news for participation.

       Published by Political Communication, 25:180-193, 2008
This article makes the case that political participation is partially conditioned on citizens’

perceptions of the state of the country, which are largely informed through mediated channels,

specifically, newspapers and television news.1 In doing so, I try to explain the role of mass

media as a force of political mobilization. Mass media serve as a conduit to social context by

relaying key information about social conditions to citizens. The media continuously inform

citizens about the world outside. Citizens, in turn, use this information in their decisions about

whether to participate in politics or not. Consequently, the mass media serve as a sentinel to the

public by cueing the public as to when it is important to participate in politics.

        Interestingly, recent scholarship has both upheld the modern media as a vital institution

that might aid in cueing citizens about when to pay attention to government (Schudson 1998) and

accused the media of turning citizens off from government altogether (Cappella and Jamieson

1997). This argument takes a position closely connected to Schudson’s theory of the “Monitorial

Citizen”, which suggests that modern citizens generally pay limited attention to government, but

are alerted by the press as to when they should pay closer attention to government. I extend

Schudson’s contention that the media serve as a cue to tell citizens when it is more important to

pay attention to government by connecting this attentiveness directly to citizen participation

through a social threat mechanism.

         The argument is that (1) citizens become more and less active in politics depending upon

their perceptions of conditions in the country and that (2) these perceptions are largely informed

through media channels. In other words, citizens pay low levels of direct attention to politics;

they survey the nation through the information provided by the mass media and become more

active in politics when that information cues them to increasing national problems. In this way,

  I thank Diana Mutz, James Baughman, Charles Franklin, Richard Merelman, Virginia Saprio, and the anonymous
reviewers for their valuable suggestions that improved this article. The Miller Center of Public Affairs provided a
scholarly environment complete with a porch and rocking chairs conducive to completing this article.

the media contribute to citizens’ decisions governing political participation by relaying important

information about the social and political context. When the media tell citizens that all is

relatively well, the incentive to act in politics is weaker; yet, if the media tell citizens that the

country is in turmoil, then the incentive to act becomes stronger. In short, bad news about issues

is good news for participation.

A theory of media influence on participation

        The idea that bad news could encourage voting may seem counter-intuitive because

media negativity is often thought of as demobilizing citizens and making them cynical and

disillusioned about politics (Ansolabehere & Iyengar 1995; Cappella & Jamieson 1997;

Patterson 1993). Studies that emphasize a detrimental effect from media negativity examine

mostly personal negativity associated with negative campaigns. Recent studies have challenged

even these conclusions and have pointed to potentially positive influences of negative campaigns

on voting (Freedman & Goldstein 1999; Kahn & Kenney 1999; Lau & Pomper 1998; Freedman,

Franz, & Goldstein 2004; Martin 2004).

        Issue negativity, on the other hand, points to the existence of problems in society, thereby

identifying problems that may require government attention such as increasing crime, economic

troubles, or inadequate health care. The effects of issue negativity have received considerably

less attention than the effects of negativity targeted at campaign actors and political institutions,

but the effects of issue negativity should not be overlooked. Indeed, they deserve emphasis.

People learn of collective problems via issue negativity. Furthermore, compared to negative

coverage of campaign actors, issue negativity is ubiquitous. Issue negativity exists within and

outside of campaigns and continuously informs citizens about social conditions in the nation as a

whole. Consequently, people become more aware of social problems and engaged in politics

when exposed to higher amounts of issue negativity.

       While media influence on politics is most often thought of in terms of direct persuasion

(Page, Shapiro, & Dempsey 1987), agenda-setting and other cognitive effects (Iyengar & Kinder

1987), and reinforcing pre-existing beliefs (Lazarsfeld, Berelson, & Gaudet 1944), media

influence is not limited to persuasion, agenda setting, and reinforcement. Media also influence

politics in more subtle, but powerful, ways by informing beliefs about social reality that in turn

shape political attitudes and behavior (e.g. Mutz 1998). Most often citizens either support or

oppose policies and programs based on their media driven perceptions of social reality. Two

examples illustrate this idea. First, citizens alter their support for incumbent politicians based on

their perceptions of economic reality. When people perceive the economy to be doing well they

reward incumbents and when they perceive the economy to be faring poorly, they punish

incumbents (Fiorina 1981). Second, citizens alter their support for welfare state provisions based

on their perceptions of the race and status of target groups being aided. Support for welfare

provisions is stronger when the target groups are perceived to be either children or elderly and

white, and support is weaker when the target groups are perceived to be of working age and

black (Gilens 1996). If media alter or simply inform perceptions of social reality – how the

economy is doing, how much crime we have, and who is poor – then the power of the press to

influence politics is unquestionable.

Bad news and republican citizenship

       To stimulate political participation, issue negativity must alter the motivation for voting.

I argue that issue negativity simultaneously makes people more aware of problems and

stimulates their interest in the campaign. Problem awareness motivates political participation via

a social threat mechanism. Threat is widely seen to motivate the behavior and attention of

organized2 and unorganized3 groups, and threat has become an increasingly important

explanation for individual behavior. When organized interests are under threat, interest groups

are able to increase membership and contributions (Hansen 1985). More importantly, when

individuals are confronted with the threat of a policy change on specific issues that they consider

to be important – such as abortion or the environment – they respond by contributing money to

interest groups and volunteering for campaigns that will try to protect the existing policy (Miller

& Krosnick 1999).

         Threat, however, need not be in the form of grave immediate danger. In earlier research,

scholars have found that candidate-based threat, measured as anxiety in response to candidates,

encouraged people to learn more about the candidates, pay greater attention to the campaign, and

alter their habitual patterns of voting behavior (Marcus & MacKuen 1993; Marcus, Neuman, &

MacKuen 2000). During the 1980 Presidential election, citizens who perceived more threat in

the political environment in response to Reagan and Carter were more likely to correctly place

Reagan and Carter with respect to liberal/conservative policy positions (Marcus & MacKuen

1993:679). Moreover, citizens who demonstrated greater anxiety in response to George Bush

and Michael Dukakis in the 1988 presidential campaign were more likely to pay attention to the

campaign and less likely to rely on partisan predispositions to make their vote choice (Marcus &

MacKuen 1993).

  Interest groups seem to proliferate in the face of perceived threats to their interest (Lowery & Gray 1993), and
memberships in interest groups receive a boost when interests are threatened (Hansen 1985).
  The contextual threat of concentrations of black Americans is thought to condition the political behavior of white
southerners in opposition to blacks (Giles & Buckner 1993; Giles & Hertz 1994; Wright 1977).

         The threat-mobilization hypothesis suggests that in the face of social threats – bad news

about high unemployment or problems with health care, for example – citizens mobilize to

counteract the perceived social threat in two ways. First, the awareness of social threats in the

form of perceptions of collective problems provides an estimate of the importance of a given

election as a symbolic event encouraging civic action. The more things that are going wrong, the

more things government is likely to act on. Elections that are perceived to be about a broad array

of social problems are likely to be thought of as more important than elections that are perceived

to be covering a narrow array of social problems. At the individual level, a person who thinks

that both the economy and health care are important issues is likely to see the election as being

more important than a person who only sees only the economy or health care as an important

issue. Citizens respond to perceived social threats as if they were external threats to the country

– they rally and vote at higher rates.

         Under this scenario, voting, and other affirming acts of civic engagement in response to

perceived social threats can be explained, in part, as a symbolic act.4 As Sears and Funk suggest,

“people may be socialized to respond to public issues in a principled and public-regarding

manner” (1990:169). This notion is entirely consistent with republican models of citizenship

starting with George Washington as “the modern Cincinnatus, forming the new nation, ruling

without excess, and returning to ordinary life” (Bellah, Madsen, Sullivan, Swidler, & Tipton

1985:254). Citizens mobilize to counteract national problems and withdraw from participation

when they perceive fewer problems.

  While citizens may become more likely to participate when they see more problems for symbolic reasons,
collective problems can also become political opportunities. Hence, citizens could rationally choose to participate to
try to further their party’s interests when election stakes are higher. Given weak parties and American lack of
interest in politics, it seems more likely that citizens will participate out of disinterested duty rather than partisan
interest. Moreover, modern civic expectations placed on citizens fit better with notions of obligations to society
rather than to political parties (Schudson 1998). In short, participation in reaction to awareness of collective
problems is most likely symbolic rather than instrumental.

          Secondly, perception of threatening social conditions in the form of multiple social

problems produces anxiety and increased attention to politics. These cognitive triggers – anxiety

and interest -- translate into political participation. Many studies have found that when groups or

individuals feel threatened by political actors or other groups they respond proactively. Marcus

and MacKuen argue that threat is an attention-getting device and find that when citizens feel

stronger threat or anxiety during political campaigns they are more likely to pay attention to

campaigns and learn more about the candidates (1993; also Marcus et al. 2000). Citizens

“abandon complacency and start to pay attention when the world signals that something is not

right” (Marcus & MacKuen 1993:673). Threat is often translated into increased arousal and

attention that would enable greater participation.

         In sum, social threat incorporates an idea of republican citizenship that channels potential

anxiety toward rather than away from participation. People participate in politics more when

they believe that more is at stake in a given election based on their evaluation of collective

problems, and they believe more is at stake when issue media is more negative. In this sense,

they are rational actors, limited by imperfect information (e.g. Simon 1957; Simon 1985), who

try to accommodate a flexible, republican sense of civic duty by participating more when they

believe there are greater challenges in the social and political system. Hence, as Berelson and

colleagues argued long ago, nonvoting may well be, in part, an expression of contentment rather

than discontentment (Berelson, Lazarsfeld, & McPhee 1954). Forgoing the opportunity to vote

may be a passive expression of consent, especially for those who possess resources necessary to


  This type of political motivation need not be cast in this positive light. People may also decline to participate in
politics because they fail to make the connections between the problems they see and politics. Many things that
theorists of politics would consider political (i.e. where values are being authoritatively allocated) are not conceived
of as being political by citizens (Gaventa 1980), or at least not connected to the political institutions that could

Hypothesis testing

         The previous sections described a two-step model of media influence on voter turnout

where negative issue coverage makes people more aware of collective problems and more

interested in politics, which in turn stimulates voting. More formally, the two hypotheses are:

    1) Higher levels of issue negativity will heighten problem awareness and political interest.
    2) Individuals who are aware of more collective problems and more interested will be more
       likely to vote.

         A combination of secondary analysis of survey data with content analysis is used to test

the hypotheses in the first step of the model – that issue negativity motivates problem awareness

and political interest. Secondary analysis of National Election Studies (NES) survey data is used

to test the second step of the model.

Media influence on problem awareness and political interest

         How exactly does issue-media influence problem awareness? The process is

exceptionally simple and very similar to Zaller’s “receive-accept-sample” model of

communication effects (1992). Zaller’s model, building off Converse (1962) and McGuire

(1968), assumes that citizens vary in their ability to resist messages. Zaller’s model has two

basic components – reception and acceptance – where acceptance is a function of political

awareness or sophistication. The implication for the effect of issue media on problem awareness

is that the least sophisticated will accept news on problems more readily than those who are more


address the problems (Merelman, Streich, & Martin 1998). If the problems that people face most often are defined
as being non-political (e.g. private issues such as business downsizing or gender relations) then it is unlikely that
they will act on them politically (Kinder & Kiewiet 1981; Sniderman & Brody 1977). When more things are
brought into the perceived realm of political problems, either by the development of political consciousness or by
politicizing increases in the perceived prevalence of problems, people become more engaged in politics.

                     Pr (attitude change) = Pr(Reception) X Pr (Acceptance|Reception)

        Reception, however, is more complicated. Based on the psychological influence of

positive and negative information, citizens are less likely to receive positive information than

they are to receive negative information (Fiske 1980; Kahneman & Tversky 1984; Pratto & John

1991). They are also less likely to pay attention to or remember positive information (Lang,

Newhagen, & Reeves 1996; Reeves, Lang, Thorson, & Rothchild 1989). Conversely, negative

information is most likely to be paid attention to, remembered, and then used for judgment (e.g.

Shields & Goidel 1998:108). Consequently, bad news is more likely than good news to

influence problem awareness.6

        News coverage is not the sole source of problem awareness. First, partisanship may bias

estimates of national conditions so that people see more problems as a reflection of their partisan

opposition to government (Fiorina 1981). For example, Republicans thought the economy was

doing better than Democrats did in 1992 (Hetherington 1996). I measure partisanship using the

direction and strength of the respondent’s partisanship using a seven-point scale. Second,

demographic differences between people such as age, education, income, race, and gender are

likely to influence citizens’ problem awareness for both individual and structural reasons.

Education, for example, may influence problem awareness because of individual differences in

public attention or awareness. But education may also influence problem awareness because of

structural differences in the life experiences of the higher and lower educated. The more

educated segment of society is likely to notice problems that affect higher status people, like the

drops in the stock market, whereas lower educated people may see problems that influence lower

  Stories of good news should not counterbalance stories about bad news in how they affect problem awareness.
Exposure to 2 bad news stories and 2 good news stories should have a similar effect as 2 bad news and no good
news stories.

status people more frequently such as unemployment. Likewise, lower income citizens are more

likely to personally experience the kinds of problems that comprise politically relevant collective

experience – crime, unemployment, criminal victimization, poor health care – but they are also

less attentive to public affairs which may hinder them from connecting their private experiences

to the larger collective experience (Mutz 1992).

        In addition to these alternative causes on problem awareness, and as previously

mentioned, political knowledge may significantly influence the way that people process news

information (e.g. Zaller 1992). It is also necessary to control for political knowledge to

distinguish the potential media effect of content from the characteristics of the consumer.

Political knowledge is included in the model both to control for its direct effect and also as an

interaction with negative media. As a surrogate for political information, the interviewer

assessment of the respondents' political information is used (Bartels 1996; Zaller 1985). Zaller

(1985) found that this measure is highly reliable.

        The same psychological advantages of negative information should also drive the effects

of bad news more directly to stimulate political interest, albeit in a simpler fashion. As news

gets worse, it becomes more pressing and more engaging. In theory, interest rises because

people become more aware of problems through exposure to bad news. The observable

implications are that problem awareness and interest should increase when exposure to bad news

is higher.

Estimating the influence of issue negativity on problem awareness and political interest

        To measure the influence of issue negativity on problem awareness and political interest,

I use the 1974 National Election Study (Miller and the National Election Studies 1999),

supplemented with an extensive content analysis conducted by the Center for Political Studies

(Miller, Miller, and Kline 1978)7 of the newspapers read by the respondents.8 The CPS analyzed

the content and tone of ninety-six newspapers read by respondents in the 1974 NES survey. If at

least seven respondents in the survey read a newspaper, the CPS selected the paper for inclusion

in the study. Ten randomly selected days – October 16, 17, 24, 25, 27, 29, 30; November 1, 3

and 4 – were included in the content analysis. For each day, all articles appearing on the front

page and editorial sections of the newspaper were analyzed. In all, the CPS analyzed 8768 front-

page stories and 9458 editorial stories. Every article was coded originally for up to two subjects

and up to two assessments of tone (one for each subject). So each article is represented by one or

two stories depending on whether a second subject was coded. Hence, the unit of analysis is the

story. Stories were originally categorized into subjects based on the primary and secondary

(where applicable) emphases of the headline and story text. I re-categorized stories as pertaining

to either the candidates or to social issues. The original study coded stories as campaign stories

only if they did not fit into an issue category (e.g. public announcements of future speeches,

personal stories, campaign statistics, candidate comparisons on style, or reporting on polls).

Hence a story about one candidate’s criticism of another’s economic policy would be coded as

an issue story and a candidate’s criticism of another’s personal life would be coded as a

campaign story. I categorized the stories originally coded as campaign events, polls, and debates

as candidate cases. I categorized stories originally coded to discuss economics, crime, social

problems, and general news as issue cases.

  The data used in this manuscript were made available by the Inter-University Consortium for Political and Social
Research. The data for the CPS media content analysis study, 1974, were originally collected by the Center for
Political Studies of the Institute for Social Research, the University of Michigan, under a grant from the John and
Mary R. Markle Foundation. Neither the original collectors of the data nor the consortium bear any responsibility for
the analyses or interpretations presented here.
  See Miller, Goldenberg, & Erbring (1979) or Erbring, Goldenberg, & Miller (1980) for a fuller description of
content analysis procedures.

         The tone of the stories – negative, positive, or mixed – was originally assessed by

determining whether the actors or editorialist in the story “criticized or praised the actions,

positions, persons, or policies, etc., mentioned in the content” (Miller, Miller, and Kline 1978:

27).9 Issue negativity is measured as the proportion of national issue news stories that were

assesses as being negative in tone. In addition to the proportion of stories that are negative, a

count of the total number of stories is included as a control.

         To measure problem awareness, I take a modified count of the number of problems a

respondent answers to the open-ended question: “What do you think are the most important

problems facing the country?” which is followed up with prompts to give up to two additional

problems. If a respondent named two problems within the same category,10 the second mention

was counted as one-half, so the measure is coded 0, 1, 1.5, 2, 2.5, and 3. To alleviate concerns

about the intensity of perceived problems, a second study conducted at the University of

Oklahoma tested a follow-up question after the most-important problems measure where

respondents were asked for each problem they listed, how much of a problem they believed the

item to be. On a scale from 1 to 100, all three problems were rated nearly the same at 75, which

provides some confidence that respondents do not mention problems they consider to be trivial.

         The first dependent variable, problem awareness, is a borderline case for using either

OLS or ordered probit. As measured, problem awareness is discrete and ordinal ranging from

zero to three with six categories. The model was tested using OLS and retested using ordered

probit analysis. The findings were essentially identical; the OLS results are presented.

   This citation refers to the codebook that accompanies the computerized data set.
   The 16 categories are: economy, crime, government, national defense, health care, education, agriculture, public
morality, natural resources, poverty, civil rights, business/labor issues, immigration, infrastructure, and consumer
protection. However, the top five categories were consistently: economy, crime, national defense, poverty, and
government for the period 1974 through 1980 and economy, crime, national defense, poverty, and health care for
1982 through 1996. The top five categories make up between 80 and 90% of all mentions for every year.

The model is:

                              y(problem awareness) = b0(constant) +
                               b1(issue negativity) +
                              b2(total number of stories) +
                              b3(political knowledge) +
                              b4(political knowledge X issue negativity) +
                              b5(partisanship) +
                              b6-10(demographics) +

       The second dependent variable, political interest, is discrete and ordered, so an ordered

probit analysis is used. The model is similar to that of problem awareness with one exception.

Where the model of problem awareness model controls for the interaction of negative news and

political knowledge, the model of political interest controls directly for problem awareness.

Effects of issue negativity on problem awareness and political interest

       The first column of Table 1 displays the tests of the effects of negative issue coverage on

problem awareness. The results of the hypothesis testing are consistent with expectations: the

proportion of negative news stories increases problem awareness (p<.05). The second row of

Table 1 shows that the total number of stories does not influence problem awareness – it appears

that tone matters more. These results hold when controlling for political knowledge and the

interaction between negative stories and political knowledge.

                                       [Table 1 about here]

       As expected, those higher in political knowledge are more aware of collective problems.

More importantly, the interaction of negative issue news and political knowledge is negative

(p<.05). This result is highly consistent with the expectations of Zaller's "receive-accept-

sample” model of media effects (1992). Those who have higher levels of political knowledge

were less influenced by the negative coverage than those with lower levels of political

knowledge. In other words, the least politically sophisticated are more likely to believe what

they read.

       The results of these tests suggest that when the newspapers read by the survey

respondents in 1974 had a higher proportion of negative issue stories, the respondents became

aware of more problems. And conversely, readers of newspapers with very little issue negativity

were aware of fewer problems. In sum, the results support idea that citizens – especially those

with lower political knowledge – develop their problem awareness in part as a consequence of

the level of issue negativity in the newspapers they read.

       The second column from Table 1 presents the effect of issue negativity on political

interest. Just as the proportion of issue negativity influences problem awareness, so does it

influence political interest. This finding alone is sufficient to suggest that issue negativity may

indirectly influence political participation as political interest is well known to drive participation

(see for example Marcus and MacKuen (1993). Even without considering the effects of issue

negativity on political participation, these two findings from Table 1 ring consistent with the idea

that the mass media acts as a sentinel when they feature bad news about issues: people become

more aware of problems and their interest in politics increases.

From problem awareness to political engagement

       Table 1 shows how issue negativity stimulated problem awareness and political interest.

This section continues with the 1974 National Election Study and examines the second

hypothesis: that problem awareness and political interest stimulate voting.

         A long literature has examined why Americans vote or refrain from doing so. Most

importantly, we know that voting is highly habitual (Plutzer 2002), suggesting that any media

related effects would be relatively modest. Resources such as time, money, and education

provide important gateways for individuals to access the political system (Verba, Schlozman, &

Brady 1995), and prior research has both noted the importance of these factors in determining

political participation and connected these resources to socio-economic status (age, income, race,

education, and gender as the representation of the resource model). Standing motivation

stemming from ideological or partisan extremity also plays a roll as those more ideologically

invested in the current system are more likely to continue to participate (Campbell, Converse,

Miller, & Stokes 1960). Party contact is also well known to be an important influence on voting

(Rosenstone & Hansen 1993). Resources, standing motivation, and party-based mobilization are

fairly consistent forces acting on citizen participation and provide a baseline with which to test

effects that are more variable, such as problem awareness and political interest11 that is partly

stimulated by bad news.

         A second empirical consideration is disentangling the effects of problem awareness on

voting from those of political knowledge on voting. A skeptical reader may be concerned that

problem awareness is simply another measure of political knowledge.12 To seriously consider

this point, the empirical model must examine the influence of political knowledge on this

process. As problem awareness has the potential of being subsumed by political knowledge,

observing the effect of problem awareness while controlling for political knowledge is important.

   The inclusion and operationalization of political interest is admittedly problematic because it likely taps long
standing political interest as well as short term interest that would vary. The finding from Table 1 indicating that
bad news affected political interest gives greater credence to the presence of a short term effect, but it most likely
has a sizable long-standing component.
   Further investigation of the effects of negative news showed no effects of negativity on political information in
1974; however, the total number of stories, irrespective of tone, predicts political information (available from the
author on request).

Of equal concern, the influence of problem awareness on voter turnout may not be equivalent

across citizens with varying levels of political knowledge. Specifically, those with medium

levels of political knowledge are expected to see the greatest impact from problem awareness on

voter turnout. Keeping in mind that the process of problem awareness influencing voting is akin

to a persuasion model – persuading ones self that voting is important – it is precisely the medium

knowledge folks who would be most influenced to change their voting behavior along the lines

of their awareness of problems. The highly knowledgeable are well known to be most likely to

vote and less in need of reasons to vote. The less knowledgeable are equally known to be least

likely to vote. But more importantly, those with the lowest levels of political knowledge may be

less adept at connecting perceptions of collective problems with their own political participation.

It is those with medium levels of political knowledge whose participation is most likely to waver

from election to election with problem awareness.

        The model examining how problem awareness and political interest influence vote

turnout is:

                              vote turnout = b0(constant) +
                              b1(problem awareness)+
                              b2(interaction of problem awareness and medium knowledge)+
                              b3(interaction of problem awareness and low knowledge)+
                              b4(medium knowledge dummy)+
                              b5(low knowledge dummy)+
                              b6-11(Socio-economic Resources) +
                              b12-13(political interest & partisan extremity) +
                              b14(party contact) +

The model interacts separately problem awareness and medium knowledge, and then problem

awareness and low knowledge, while controlling for medium and low levels of knowledge. High

knowledge is the comparison group.

       The findings in Table 2 show that the effect of problem awareness is strongly mediated

by political knowledge in ways that are consistent with expectations. We see no direct effect

from problem awareness, while the interaction between medium knowledge and problem

awareness is positive and statistically significant (p<.05) and the interaction between problem

awareness and low political knowledge fails to achieve statistical significance. These three

findings collectively mean that the direct effect of problem awareness is channeled exclusively

through those with a medium level of knowledge.

                                       [Table 2 about here]

       The other factors in the model are as one would expect. Those with lower and medium

levels of knowledge are less likely to vote than are high knowledge individuals as indicated by

the negative and statistically significant coefficients on lower and medium knowledge. Political

interest positively predicts voting as does strength of partisanship and party contact. The

representation of the resource model also performs as expected with positive effects on

education, income, and age.

       Keeping all other factors at their means, problem awareness strongly influences voting

amongst those with medium knowledge. Those mentioning but a single public problem that they

thought was important had a probability of voting at about 57%, those at the other end of the

scale mentioning three distinct problems had a probability of voting at 71% (a difference of

14%). The average number of mentions for those with medium levels of knowledge was 2 with

a standard deviation of .7. Moving one standard deviation from the mean changes the likelihood

of voting by about 5%.

Discussion and Conclusion

       The data used to test both hypotheses come from a unique source and a unique time. The

1974 CPS study remains unrivaled in the detail and extensiveness of its content coding of

positive and negative stories (other projects have done content analysis in combination with

surveys, but none to the author’s knowledge could adequately test the first hypothesis). The

1974 election also followed a particularly rough year. Nixon resigned on August 9th, Ford

pardoned him on September 8th, and the Republican Party lost 48 seats in the House and 4 in the

Senate in November. The American involvement in Vietnam was winding down, but would not

be completed until after the election. Unemployment was at 6.6% and rising, the consumer price

index neared 11%, and world oil prices jumped 252% in 1974 from $3.59 to $11.58 per barrel.

       In spite of what appears to be a depressing year by any standard, the amount of bad news

in the newspapers read by the survey respondents varies substantially. One vivid example comes

from differences in the coverage of three Chicago-based newspapers included in the study. The

Tribune ran 84 stories coded as negative, the Sun-Times ran 40, and the Daily News ran 27

stories assessed as negative. Readers of these three papers would get very different impressions

of the levels of social, political, and economic problems facing the country. Cast in this light,

1974 offers a strong set of conditions to test the paper’s argument. It has a significant array of

problems that people could be cued to, yet the range of bad news in the press is wide enough to

have variation in the independent variable. In other words, the effects shown in this paper may

be present in most election years, but there may not be enough variation in exposure to bad news

to document them.

       Still, the effects appear to be most pronounced in those with medium levels of

knowledge. Taking into consideration the degree of problems facing the nation in 1974, these

results suggest that while the press may play a sentinel roll along the lines implied by Schudson’s

theory of the Monitorial Citizen, it may only do so with those who are moderately informed to

begin with.

       This research suggests is that citizens may become more active in politics if politicians

and the press focused more on tackling social problems rather than tackling one another or

putting candidates under the microscope. Contrary to conventional wisdom, media negativity

need not be detrimental for democratic citizenship. Indeed, the media may serve as a sentinel to

the people arousing them to participate when conditions seem bleak. Rather than decrying the

press for being too negative, we may wish to encourage newspapers and media outlets that paint

the world in rosy colors to put the thorns back into the picture.

       While the central claim of this paper is that bad news may stimulate a sense of republican

duty in the American citizenry, the findings also bear on the debate over the effect of negative

advertising on voter turnout. One of the more troubling aspects of that debate has been the lack

of specificity of clear mechanisms that could translate negative campaigns into political

participation (see Martin 2004). One consequence of this lack of specificity has been

inconsistency across research results (see Lau, Sigelman, Heldman, & Babbitt 1999). As

Sigelman and Kugler (2003) neatly point out, citizens do not equally read campaigns as negative

or positive, and therefore do not have consistent reactions to campaigns that scholars consider

“negative”. While there could be many mechanisms that could effectively translate negative ads

to political participation, the degree to which bad news or negative campaigns encourage people

to see greater problems facing the country may help to further explain why some negative

campaigns stimulate turnout and others do not.


Ansolabehere, S., & Iyengar, S. (1995). Going Negative: How Attack Ads Shrink and Polarize
        the Electorate. New York: Free Press.
Bartels, L. M. (1996). Uninformed Votes: Information Effects in Presidential Elections.
        American Journal of Political Science, 40(1), 194-230.
Bellah, R. N., Madsen, R., Sullivan, W. M., Swidler, A., & Tipton, S. M. (1985). Habits of the
        Heart: Individualism and Commitment in American Life. Berkeley: University Of
        California Press.
Berelson, B., Lazarsfeld, P. F., & McPhee, W. N. (1954). Voting: A Study of Opinion Formation
        in a Presidential Campaign. Chicago: University Of Chicago Press.
Campbell, A., Converse, P., Miller, W., & Stokes, D. (1960). The American Voter. New York:
Cappella, J. N., & Jamieson, K. H. (1997). Spiral of Cynicism: The Press and the Public Good.
        New York: Oxford University Press.
Converse, P. E. (1962). Information Flow and the Stability of Partisan Attitudes. Public Opinion
        Quarterly, 26, 578-99.
Erbring, L., Goldenberg, E. N., & Miller, A. H. (1980). Front Page News and Real World Cues:
        A New Look at Agenda Setting by the Media. American Journal of Political Science, 24,
Fiorina, M. P. (1981). Retrospective Voting in American National Elections. New Haven: Yale
        University Press.
Fiske, S. T. (1980). Attention and Weight in Person Perception: The Impact of Negative and
        Extreme Behavior. Journal of Personality and Social Psychology, 38, 889-906.
Freedman, P., Franz, M., & Goldstein, K. (2004). Campaign Advertising and Democratic
        Citizenship. American Journal of Political Science, 48(4), 723-741.
Freedman, P., & Goldstein, K. M. (1999). Measuring Media Exposure and the Effects of
        Negative Campaign Ads. American Journal of Political Science, 43, 1189-208.
Gaventa, J. (1980). Power and Powerlessness: Quiescence and Rebellion in an Appalachian
        Valley. Urbana: University Of Illinois Press.
Gilens, M. (1996). Race and Poverty in America: Public Misperceptions and the American
        News Media. Public Opinion Quarterly, 60, 515-541.
Giles, M., & Buckner, M. (1993). David Duke and Black Threat: An Old Hypothesis Revisited.
        Journal of Politics, 55, 702-13.
Giles, M. W., & Hertz, K. (1994). Racial Threat and Partisan Identification. American Political
        Science Review, 88, 317-326.
Hansen, J. M. (1985). The Political Economy of Group Membership. American Political Science
        Review, 79, 79-96.
Hetherington, M. J. (1996). The Media's Role in Forming Voters' National Economic
        Evaluations in 1992. American Journal of Political Science, 40, 372-95.
Iyengar, S., & Kinder, D. R. (1987). News That Matters: Television and American Opinion.
        Chicago: University Of Chicago Press.
Kahn, K. F., & Kenney, P. J. (1999). Do Negative Campaigns Mobilize or Suppress Turnout?
        Clarifying the Relationship between Negativity and Participation. American Political
        Science Review, 93, 887-890.
Kahneman, D., & Tversky, A. (1984). Choices, Values, and Frames. American Psychologist, 39,

Kinder, D. R., & Kiewiet, D. R. (1981). Sociotropic Politics: The American Case. British
        Journal of Political Science, 11, 129-161.
King, G. (1989). Unifying Political Methodology: The Likelihood Theory of Statistical Inference.
        New York: Cambridge University Press.
Lang, A., Newhagen, J., & Reeves, B. (1996). Negative Video as Structure: Emotion, Attention,
        Capacity, and Memory. Journal of Broadcasting & Electronic Media, 40, 460-77.
Lau, R. R., & Pomper, G. M. (1998). Accentuating the Negative? Effects of Negative
        Campaigning in US Senate Elections. Paper Presented at the American Political Science
        Association, Boston.
Lau, R. R., Sigelman, L., Heldman, C., & Babbitt, P. (1999). The Effects of Negative Political
        Advertisements: A Meta-Analytic Assessment. American Political Science Review, 93,
Lazarsfeld, P. F., Berelson, B., & Gaudet, H. (1944). The People's Choice: How the Voter Makes
        Up His Mind in a Presidential Campaign. New York: Duell Sloan and Pearce.
Long, J. S. (1997). Regression Models for Categorical and Limited Dependent Variables.
        Beverly Hills: Sage Publications.
Lowery, D., & Gray, V. (1993). The Density of State Interest Group Systems. Journal of
        Politics, 55, 191-206.
Marcus, G. E., & MacKuen, M. B. (1993). Anxiety, Enthusiasm, and the Vote: The Emotional
        Underpinnings of Learning and Involvement during Presidential Campaigns. American
        Political Science Review, 87, 672-685.
Marcus, G. E., Neuman, W. R., & MacKuen, M. (2000). Affective Intelligence and Political
        Judgment. Chicago: University Of Chicago Press.
Martin, P. S. (2004). Inside the Black Box of Negative Campaign Effects: Three Reasons Why
        Negative Campaigns Mobilize. Political Psychology, 25(4), 545-562.
McGuire, W. J. (1968). The Nature of Attitudes and Attitude Change. In G. Lindzey & E.
        Arnonson (Eds.), Handbook Of Social Psychology (2nd Ed., Pp. 136-314). Reading, MA:
Merelman, R. M., Streich, G., & Martin, P. S. (1998). Unity and Diversity in American Political
        Culture: An Exploratory Study of the National Conversation on American Pluralism and
        Identity. Political Psychology, 19, 781-807.
Miller, A. H., Goldenberg, E. N., & Erbring, L. (1979). Type-Set Politics: Impact of
        Newspapers on Public Confidence. American Political Science Review, 73, 67-84.
Miller, J. M., & Krosnick, J. A. (1999). The Impact of Policy Change Threat on Grassroots
        Activism. Paper Presented At The Midwest Political Science Association, Chicago,
Miller, W.E., Miller, A.H., and Kline, F. G. (1978). Media Content Analysis Study, 1974
        [Computer file]. Conducted by University of Michigan, Center for Political Studies.
        ICPSR ed. Ann Arbor, MI: Inter-university Consortium for Political and Social Research
        [producer and distributor].
Miller, W.E. and the National Election Studies. (1999). National Election Studies, 1974: Post-
        Election Study [dataset]. Ann Arbor, MI: University of Michigan, Center for Political
        Studies [producer and distributor].
Mutz, D. C. (1992). Mass Media and the Depoliticization of Personal Experience. American
        Journal of Political Science, 36, 483-508.

Mutz, D. C. (1998). Impersonal Influence: How Perceptions of Mass Collectives Affect Political
        Attitudes. New York: Cambridge University Press.
Page, B. I., Shapiro, R. Y., & Dempsey, G. R. (1987). What Moves Public Opinion? American
        Political Science Review, 81, 23-44.
Patterson, T. E. (1993). Out of Order. New York: A. Knopf.
Plutzer, E. (2002). Becoming a Habitual Voter: Inertia, Resources, and Growth in Young
        Adulthood. American Political Science Review, 96(1), 41-56.
Pratto, F., & John, O. P. (1991). Automatic Vigilance: The Attention-Grabbing Power of
        Negative Social Information. Journal of Personality And Social Psychology, 61, 380-91.
Reeves, B. R., Lang, A., Thorson, E., & Rothchild, M. (1989). Emotional Television and
        Hemispheric Specialization. Human Communication Research, 15, 493-508.
Rosenstone, S. J., & Hansen, J. M. (1993). Mobilization, Participation, and Democracy in
        America. New York: Macmillan.
Schudson, M. (1998). The Good Citizen: A History of American Civic Life. New York: Martin
        Kessler Books.
Sears, D. O., & Funk, C. L. (1990). Self-Interest and Americans' Public Opinions. In J. J.
        Mansbridge (Ed.), Beyond Self-Interest (Pp. 147-170). Chicago: University Of Chicago
Shields, T. G., & Goidel, R. K. (1998). Taking Credit and Avoiding Blame: Good News, Spin
        Control, and Democratic Accountability. Political Communication, 15, 99-115.
Sigelman, L., & Kugler, M. (2003). Why Is Research on The Effects of Negative Campaigning
        So Inconclusive? Understanding Citizens' Perceptions of Negativity. Journal of Politics,
        65(1), 142-160.
Simon, H. A. (1957). Models of Man: Social and Rational; Mathematical Essays on Rational
        Human Behavior in Society Setting. New York: Wiley.
Simon, H. A. (1985). Human Nature in Politics: The Dialogue of Psychology with Political
        Science. American Political Science Review, 79, 293-304.
Sniderman, P., & Brody, R. (1977). Coping: The Ethic of Self-Reliance. American Journal of
        Political Science, 21, 501-21.
Verba, S., Schlozman, K. L., & Brady, H. E. (1995). Voice and Equality: Civic Voluntarism in
        American Politics. Cambridge: Harvard University Press.
Wright, G. C. (1977). Contextual Models of Electoral Behavior: The Southern Wallace Vote.
        American Political Science Review, 71, 497-508.
Zaller, J. (1992). The Nature and Origins of Mass Opinion. New York: Cambridge University
Zaller, J. R. (1985). Pre-Testing Information Items on the 1985 NES Pilot Survey. Paper
        Presented At The Report To The National Election Studies Board Of Overseers.

    Table 1: Effects of negative national issue stories on problem awareness and
                     political interest during the 1974 campaign
                                          Problem Awareness                Political interest
    Proportion of negative stories              2.13**                          1.22**
                                                 (.707)                          (.380)
       Total number of stories                    .001                           -.001
                                                 (.002)                          (.003)
         Political knowledge                    .607**                          .459**
                                                 (.172)                          (.042)
Interaction of political knowledge and
                                                 -.558**                           --
     proportion of negative stories
         Problem awareness                          --                           .055
            Partisanship                          -.002                        .106**
                                                  (.011)                        (.038)
             Education                           .038**                          .025
                                                  (.009)                        (.017)
               Income                             -.006                          .013
                                                  (.006)                        (.009)
               Female                             -.003                        -.184*
                                                  (.042)                        (.074)
               White                              -.132                          .008
                                                  (.074)                        (.127)
                Age                               -.001                        .013**
                                                  (.001)                        (.002)
                 n                                 1045                          1041
             R-square                              .105                            --
     Ordered Probit Threshold 1                      --                          1.86
             Threshold 2                            --                           2.79
             Threshold 3                            --                           3.99
         model significance                     p<.0001                            --
          model chi-square                         --                         296.19**
        initial log likelihood                     --                         -1181.39
         final log likelihood                      --                         -1033.29
Note: * indicates significance at p < .05; ** indicates significance at p < .01; standard errors in
parentheses. Column 1 estimates are regression coefficients; column 2 estimates are ordered probit
Source: National Election Studies 1974 with content analysis from CPS.

    Table 2: How political knowledge influences the effects of problem awareness on voter
                              turnout during the 1974 campaign

                     Problem awareness                                                  -.095
   Interaction of problem awareness and medium political
                         knowledge                                                     .402*
     Interaction of problem awareness and low political
                         knowledge                                                      .037

                Medium political knowledge                                             -.930*
                  Low political knowledge                                             -1.11**
                       Political interest                                             .436**
                   Strength of partisanship                                           .282**
                         Party contact                                                .944**
                          Education                                                    .054*
                           Income                                                     .071**
                            Female                                                      .111
                            White                                                       .225
                             Age                                                      .029**
                           Constant                                                   -4.46**
                                n                                                       1921
                       Model chi-square                                              505.91**
                     initial log likelihood                                          -1304.92
                      final log likelihood                                           -1051.97

Note: * indicates significance at p < .05; ** indicates significance at p < .01; standard errors in parentheses.
Estimates are logit coefficients.
Source: National Election Studies, 1974.


To top