Добавил:
Upload Опубликованный материал нарушает ваши авторские права? Сообщите нам.
Вуз: Предмет: Файл:

cohen_tsfati_and_sheafer_2008-libre

.pdf
Скачиваний:
21
Добавлен:
17.02.2016
Размер:
119.62 Кб
Скачать

Public Opinion Quarterly Advance Access published May 9, 2008

Public Opinion Quarterly, Vol. 1, No. 0 2008, pp. 1–14

THE INFLUENCE OF PRESUMED MEDIA INFLUENCE IN POLITICS

DO POLITICIANS’ PERCEPTIONS OF MEDIA POWER MATTER?

JONATHAN COHEN

YARIV TSFATI

TAMIR SHEAFER

ABSTRACT Much of what politicians do, we maintain in this paper, is driven by their belief in the power of media, which motivates their desire to be featured in news coverage. Our argument rests upon recent advances in communication theory, stressing “the influence on presumed media influence” (Gunther, Albert C., and J. Douglas Storey. 2003. “The Influence of Presumed Influence.” Journal of Communication 35(2):199–215) and contributes to our understanding of the mediatization of politics. Combined data from a survey of Israeli members of Knesset (MKs; n = 56), parliamentary reporters’ (n = 20) rankings of MKs’ media motivations, Knesset records of MKs’ political activity, and data on the frequency of MKs’ news appearances were used to test this argument. Structural equation modeling revealed that politicians’ belief in the power of media increases their motivation and effort to appear in media coverage, which in turn is related both to greater media prominence and to more parliamentary activity. These results are discussed in light of their implications for both our understanding of political actors and the role of journalists.

In 2005, a version of “The Apprentice” was broadcast on Israeli Channel 2 to great fanfare. The program’s goal was to choose a leader (as it was called), and in it contestants were to demonstrate leadership, for example, by mobilizing

JONATHAN COHEN AND YARIV TSFATI are with the Department of Communication, University of Haifa, Israel and TAMIR SHEAFER is with the Department of Communication and Journalism and the Department of Political Science, The Hebrew University, Jerusalem, Israel. The authors are grateful to Mr Many Avrahami of Ifat Media Information Center for the news coverage data, Roi Estline and Tsahi Hayat for their dedicated research assistance, Efrat Yaskil and the staff of the University of Haifa Survey Center, and Dr Mina Zemach of Dahaf Survey Research Institute. The authors would also like to thank the participants in the “Influence of Presumed Influence Workshop” for their encouragement and helpful feedback. Address correspondence to Yariv Tsfati; e-mail: ytsfati@com.haifa.ac.il.

doi:10.1093/poq/nfn014

C The Author 2008. Published by Oxford University Press on behalf of the American Association for Public Opinion Research. All rights reserved. For permissions, please e-mail: journals.permissions@oxfordjournals.org

2

Cohen, Tsfati, and Sheafer

communities to build a playground, recruiting as many volunteers as possible to a nonprofit organization and advancing legislation in the Knesset. To fulfill the last task, contestants had to lobby members of the Knesset (MKs) to help them, and a surprising number of Israel’s busy politicians indeed cooperated with what at first glance must have seemed to them a foolish request. Some MKs got so involved in the task that they delayed meetings and spent long hours lobbying their colleagues and going to the floor to vote on these laws, all within the two days in which this episode was filmed.

Why did MKs cooperate with this entertainment program to pass laws that were previously not at the top of their agenda? Political analysts argued that they did so because they wanted a share of the television exposure that this program offered (Verter 2005), an argument that is in line with much theorizing about the increasingly important role media play in politics. This trend-–the “mediatization” of modern politics—is taking place in Israeli politics as elsewhere (e.g., Blumler and Kavanagha 1999; Caspi 1996), as evident both in terms of its outcomes for the political world and in terms of the interaction of politicians with the demands of media routines (Altheide and Snow 1979; Shoemaker and Reese 1991).

Why are elected political actors so keen to be covered by media? Several scholars (e.g., Mutz 1989; Becker and Kosicki 1995) have suggested that the answer is a strong belief among political elites in media impact on the electorate, as well as on advancing their political agendas. The current investigation attempts to test this notion systematically, based on recent advances in media research that focus on perceptions of media influence and their role in indirect, yet sometimes powerful, effects.

The Influence of Presumed Media Influence and the Mediatization of Politics

More than two decades have passed since Davison (1983) suggested: “In some cases, a communication leads to action not because of its impact on those to whom it is ostensibly directed, but because others (third persons) think it will have an impact on the audience” (p. 1). This idea, now dubbed as “the influence of presumed influence” (Gunther and Storey 2003), has been extensively revisited in recent years in various aspects of social life including parental monitoring of harmful content (Hoffner and Buchanan 2002), and preparations for Y2K (Tewksbury, Moy, and Weis 2004). Taken together, this literature points out that people act upon their perceptions of media influence regardless of whether or not these perceptions are accurate. Put differently, “the myth of media impact is influential, too” (Katz 1986, p. 32).

Political communication scholars point to a process of “mediatization” of politics, which is characterized by an “increasing intrusion of the media in the political process” (Mazzoleni and Schulz 1999, p. 248). There is mounting

Politicians’ Perceptions of Media Power

3

evidence that, as part of this process, political actors adapt their behavior to the requirements of the news media, in order to achieve coverage (Altheide and Snow 1979; Ericson, Baranek, and Chan 1989).

But “mediatization” processes do not seem to affect all politicians equally. Studies reveal large differences in the extent of effort and resources spent in an attempt to receive media coverage on the part of different legislators in the same institution (e.g., Hess 1996). Why do the levels of media motivations among politicians vary so much? It is likely that variations in beliefs about media influence are an important determining factor in the degree to which politicians play along with “the media game.” The more a politician believes that media are powerful and are of key importance to achieving his or her goals—reelection or political survival (Mayhew 1974) and policymaking and legislation (Sellers 2000)—the more they are likely to desire news exposure and invest efforts to attain this goal.

Two main reasons for the general importance of media for politicians have been suggested by prior research. First, politicians recognize the significance of news and entertainment media for their reelection. In the age of “mediatization,” it is claimed that voters are unlikely to support candidates they do not know enough about, mostly from media (Just et al. 1996), and that politicians are well aware of this (Ericson et al. 1989). Hence, political actors are likely to compete for media coverage because they believe that media coverage influences the public and that it is a prerequisite for reelection.

Second, some scholars argue that legislators are increasingly recognizing that media coverage has become a part of the policymaking process within legislative institutions (Cook 1989). The news media “act as a communicator within a policy community” (Kingdon 1995, p. 59) and have become a tool legislators use to influence their colleagues in the policymaking process (Kedrowski 1996). In other words, a politician may desire news coverage because she perceives it to affect her fellow politicians and make them believe that she is seen positively by, and that the issues she is promoting are important to, the public. Based on their perception of the media’s influence on public opinion, she will be able to attract the attention of her fellow politicians to the issues and initiatives close to her heart. Thus, we hypothesize that (H1) legislators’ perceptions of media influence on the public will be associated with their media motivation and effort, and that (H2) legislators’ perceptions of media influence on other legislators will also be associated with such media motivation and effort.

Media Motivation and Effort and Media Visibility

Issuing press releases, keeping close relations with journalists, and playing by the media’s rules do not in and of themselves guarantee media visibility. Journalists’ news values and politicians’ talents and skills (see Sheafer

4

Cohen, Tsfati, and Sheafer

2001) ultimately determine whether or not politicians’ efforts will bear fruit. However, because the great majority of routine news stories involve the intentional promotion of purposive accomplishments by officials, and given that in such routine stories journalists do not often actively seek information that they are not “fed,” media motivation is necessary for receiving coverage, especially positive coverage. Indeed, previous research has demonstrated that politicians’ media motivation is positively related to the extent of their media coverage (Sheafer and Wolfsfeld 2004). Thus, (H3) legislators’ media motivation and effort will be positively associated with the amount of coverage they receive.

MEDIA MOTIVATION AND EFFORT AND PARLIAMENTARY ACTIVITY

In the past, observers of the political world categorized politicians into two groups, using the metaphor of work horses and show horses. Show horses are self-promoting legislators who direct their skills and ambition toward achieving media publicity at the expense of legislative activities, in contrast with the industrious work horses, who are more interested in legislative activity than in media coverage (Clapp 1963). However, by the end of the 1980s, the distinction between these two types of legislators no longer seemed to hold, perhaps because making news had become so central to the legislative process (Langbein and Sigelman 1989). Arguably, media motivation and effort do not come at the expense of, but rather dovetail with political activity. This is nicely demonstrated by Kedrowski (1996) whose analysis of the US Congress revealed a new breed of legislators she called “media entrepreneurs.” These legislators “are likely to have a greater sense of efficacy in . . .the media [ability] to influence policymakers and the public” (p. 46), and they “more strongly value using the media to influence the policy process” (p. 129). They attempt to use media coverage in order to facilitate their legislative initiatives. Thus, it is hypothesized that (H4) legislators’ media motivation and effort will be positively associated with the amount of their parliamentary activity.

Methods

To test these hypotheses, we conducted a survey of members of the Israeli 16th Knesset (MKs), elected in January 2003. The survey was administered by the University of Haifa Survey Research Center between March and May 2005. MKs’ offices were contacted by phone and asked to complete a one-page survey regarding their attitudes about media influence. For most of the part, the survey was faxed to the offices and returned after being completed by the MK. In 12 cases, the interview was conducted over the phone.1 Several calls

1. When surveys were faxed to offices, we asked specifically and repeatedly that the MK himself or herself would fill it out.

Politicians’ Perceptions of Media Power

 

5

Table 1. Distribution of Respondents Across Parties, and Response Rates

 

 

 

 

Party

Seats in

Number of

Response

 

Knesset

respondents

rate (%)

 

 

 

 

Likud

40

10

25.0

Labor-Meimad

21

6

28.5

Shas

11

8

72.7

Shinui

15

9

60.0

Smaller parties

33

23

69.7

 

 

 

 

(on average eight) were made to each MK in order to maximize cooperation. The survey data were supplemented by data on each MK’s parliamentary activity (obtained from the Knesset’s web site)2 and content analysis data on his or her appearances in electronic news media. In addition, to gauge media motivation and effort, we conducted a survey of a sample of 20 parliamentary reporters, who rated the motivation demonstrated by MKs in pursuit of coverage. A detailed description of the measures used can be found in the appendix.

SAMPLE

The Israeli Knesset has 120 members, of which 56 responded to our survey (47 percent), providing a relatively large cross section of this small population. In the 16th Knesset, 15 percent of members were women, and in our sample 12.5 percent. Our sample includes only seven respondents who were cabinet ministers at the time of the survey (four) or in the past (three). This is an underrepresentation, given that, of the 120 members, 45 had served in the cabinet at one point or another. However, given the seniority of ministers (and even former ministers) it is reasonable that they are less likely to respond to surveys. The distribution of respondents among parties represented in the 16th Knesset is presented in table 1. The average tenure of respondents in our survey as MKs was 8.7 years as compared with the mean tenure of 9.2 in general. In sum, while our sample represents the various sexes, ranks, and parties that make up the Israeli parliament, it does not exactly replicate the makeup of the 16th Knesset and thus may be subject to nonresponse bias. We explored this bias by examining the news media coverage and the parliamentary activity of all 120 MKs. This analysis revealed that when controlling for political centrality and seniority, coverage and activity were not significantly related to survey participation. We therefore attempted to alleviate the concern over nonresponse by controlling for these and other variables in the following analysis.

2. http://www.knesset.gov.il/stats/heb/16/template.asp.

6

Cohen, Tsfati, and Sheafer

Figure 1. Maximum Likelihood Estimates for the Model (Standardized Coefficients).

MODEL SPECIFICATION

The structure of the model is presented in figure 1. The endogenous variables in the model are media motivation and effort, media coverage, and parliamentary activity. To conduct a valid test of the hypotheses, several variables known to predict media motivation and effort were considered as potential alternative explanations. The equation predicting this construct thus controls for inefficacy toward media, sense of obligation to appear in media, and type of media motivation (national versus local or sectorial). Given past research and findings (e.g., Sheafer 2001), we theorized that media coverage and parliamentary activity are affected by sex, political standing, political centrality, tenure, and whether or not the MKs were part of the governing coalition. Paths corresponding to our hypotheses were added to the model.

Since our hypotheses imply mediational relationships between triads of variables, model building and trimming strategies (Kline 1999, p. 132) were used to examine the utility of including or excluding the mediational and direct paths in each triad. Results of this analysis supported the exclusion of the paths for direct effects of presumed media influence (on public or other MKs) on parliamentary activity and media coverage.

Despite the fact that according to our hypotheses media motivation and effort were expected to influence both media coverage and parliamentary activity, our model does not include paths for an association between coverage and activity, because most previous studies did not find such a correlation (Payne 1980;

Politicians’ Perceptions of Media Power

 

7

 

Table 2. Maximum Likelihood Estimates for the Model

 

 

 

 

 

 

 

 

Structural component

B (SE)

 

 

 

 

 

 

Dependent variable = media motivation and effort

 

 

 

 

Presumed media influence on public → media

0.28

(0.14)

 

motivation and effort

 

 

 

 

Presumed media influence on other MKs → media

−0.03

(0.12)

 

motivation and effort

 

 

 

 

More important to succeed in national media (= 1) → media

0.36

(0.19)#

 

motivation and effort

 

 

 

 

Media more powerful than politicians (= 1) → media

−0.36

(0.20)#

 

motivation and effort

 

 

 

 

Politicians have an obligation to appear in media → media

0.00

(11)

 

 

motivation and effort

 

 

 

 

Dependent variable = news media coverage

 

 

 

 

Media motivation and effort → coverage in news media

1.28

(0.20)

 

Political centrality → coverage in news media

−0.19

(0.15)

 

Coalition (= 1) → coverage in news media

−0.13

(0.18)

 

Tenure in the Knesset → coverage in news media

0.02

(0.02)

 

Political standing → coverage in news media

0.70

(0.21)

 

Sex (female = 1) → coverage in news media

−0.43

(0.38)

 

Dependent variable = parliamentary activity

 

 

 

 

Media motivation and effort → parliamentary activity

0.27

(0.10)

 

Political centrality → parliamentary activity

−0.12

(0.09)

 

Coalition (= 1) → parliamentary activity

−0.13

(0.18)

 

Tenure in the Knesset → parliamentary activity

−0.05

(0.01)

 

Political standing → parliamentary activity

−0.30

(0.13)

 

Sex (female = 1) → parliamentary activity

0.26

(0.24)

 

 

 

 

 

 

NOTE.—Table entries are unstandardized coefficients. Numbers in parentheses are standard errors.

p ≤ .001, p ≤ .01, p ≤ .05, # p < .10.

Robinson 1990; Squire 1988) and in our data, as well, after controlling for media motivations there was no association between these constructs.

RESULTS

Maximum likelihood estimates for the model were calculated using AMOS (Arbuckle 1989). Figure 1 presents standardized estimates for the paths relevant to our discussion, and table 2 presents unstandardized estimates and standard errors for these paths. The model explains 28.8 percent of the variance in media motivation and effort, 50.7 percent of the variance in parliamentary activity, and 63.9 percent of the variance in news coverage. Fit estimates for this model were satisfactory: the Normed Fit Index was .97, the Comparative Fit Index was .99, and 1-RMSEA was .93. Chi-square for the model was 49.95 (df = 39; p = .11), providing strong support for the model.

8

Cohen, Tsfati, and Sheafer

Results for the model showed that, consistent with H1, presumed media influence on the public positively affected media motivation and effort (β = .36; SE = 0.14). However, there was no evidence in the data supporting H2, which predicted an association between presumed media influence on other MKs and media motivation and effort. The media motivation and effort construct was also statistically unrelated to the control variables. H3 predicted that media motivation and effort would be positively associated with news coverage. All other factors in the model being equal, results provided support for this hypothesis (β = .71; SE = 0.20). H4 predicted an association between media motivation and effort and parliamentary activity. Ceteris paribus, the media motivation and effort construct was significantly associated with parliamentary activity, as predicted by this hypothesis (β = .28; SE = 0.10).

Tenure in the Knesset was negatively (β = −.48; SE = 0.01) related to parliamentary activity and not significantly related to coverage in news media. Political standing was positively related to news media coverage (β = .35; SE = 0.21), but negatively related to parliamentary activity (β = −.25; SE = 0.13). Membership in the coalition, sex, and political centrality were not related to either media coverage or parliamentary activity.

Finally, to examine the overall contribution of presumed influence and media motivation and effort to the model, a model that does not include paths representing our hypotheses was estimated. We evaluated the two competing models using the criteria of model fit, explained variances, and chi-square comparisons, following the strategy utilized by Cohen, Vigoda, and Samorly (2001). Results showed that according to all fit indices, our model performs better than the competing model. The explained variances in the dependent variables in our model were higher than those in the competing model (for media motivation and effort .28, compared to .12; for coverage .64, compared to .30; for parliamentary activity .51, compared to .39). Finally, the chi-square difference test (χ 2 = 55.54; df = 4) was significant (p < .001), supporting the inclusion of the hypothesized paths. In sum, our model performs better than a competing model that does not include paths representing our hypotheses.

Discussion

The influence of presumed media influence on politicians has been mostly demonstrated using anecdotes (Becker and Kosicki 1995; Mutz 1989). The problem with such anecdotal reasoning is that we cannot know for sure that it is perceptions of media influence, and not other factors which account for politicians’ actions. Our results provide systematic evidence on the importance of presumed media influence in politics. We were able to demonstrate that politicians’ coverage by electronic news media and their activities in parliament are both indirectly influenced by MK’s perceptions of media influence.

Though an indirect effect, the influence of presumed media influence was not at all trivial in magnitude (for coverage, the mediated effect was

Politicians’ Perceptions of Media Power

9

.28 1.28 = .36, for parliamentary activity: .28 .27 = .08). Presumed media influence on the public was by far the best predictor of media motivation and effort, and this construct was the best predictor of coverage.

Media motivation and effort not only affect politicians’ work in their relationships with journalists, but they also relate to their activity in parliament. One interpretation is that legislators propose laws, and submit queries in order to generate coverage, which in turn is used to promote their agendas. While this is not the only possible interpretation of the data, we believe it enjoys an advantage over others (e.g., that the correlation is explained by ambition), since it is in line with the recent literature that points out that parliamentary activity may be motivated by a quest for media attention (e.g., Mixon, Gibson, and Upadhyaya 2003). Whatever the explanation for the relationship, there was no support in our data for the show horse versus work horse distinction. That is, media motivation and effort were positively, and not negatively, associated with parliamentary activity.

Whereas our data provide strong support for an association between presumed media influence on the public and media motivation and effort, there was no evidence in our data supporting a relationship between media motivation and effort and perceptions of media influence on other politicians. One possible explanation for this null finding is that the operational definition of the presumed media influence on other MKs (“In your opinion, to what extent do Israeli media influence other Knesset Members?”) was not focused enough. Perhaps a question asking more specifically about the presumed influence of the media on the parliamentary activity of other MKs would have revealed the expected relationship (see Kedrowski 1996).

While our study utilizes a unique data set, collected from senior politicians and reporters, one limitation of this study is related to the possibility of reverse causation. However, in line with previous research on this topic, our study suggests that perceived media influence generates outcomes rather than serving as an excuse or an explanation for them. Our research design allows a rather strong test of the casual direction in that it is unlikely to claim that MKs’ assessments regarding media influence were merely excuses for their media success and motivations, as these motivations and success were measured by external data and for the most part from a point in time that follows the survey measuring the IVs.

Another limitation of the study is related to the fact that our sample underrepresents the more senior MKs—former and current cabinet ministers. Though the fact that participating MKs were not different from their nonparticipating counterparts in the major dependent variables in the current study suggests that this limitation may not be crucial, a more representative sample would have definitely increased our confidence in the generalizability of our findings.

Lastly, one additional limitation relates to the fact that 44 MKs completed the survey by fax, which raises the suspicion that it was not MKs, but rather their staffers, who completed the surveys. While we do not have any way to

10

Cohen, Tsfati, and Sheafer

ensure that this was not the case, the fact that despite our multiple calls our response rates were not higher suggests that staffers had to actually get the MK. Still, the sincerity of survey responses are always to be taken with a measure of doubt, and this is also the case in the current investigation.

Much research on the influence of presumed media influence assumes that the perceived effects of media are misperceptions or exaggerations. Is it possible that politicians are wrong in perceiving that mass media have an impact on their careers? It is difficult to answer this question decisively, given the current state of research on communication effects on voters. Some empirical research shows that media coverage of candidates indeed affects their electoral success (Bartels 1988; Zaller 1992), but additional research will be required before we can answer the question regarding the accuracy of politicians’ perceptions decisively.

Perceptions of public opinion, not only the actual distribution of opinion in the population, have long been considered an essential part of public opinion dynamics: political behavior is shaped not only by what people think, but also by what they perceive that others think. The current investigation shows that perceptions of media influences on public opinion, not only perceptions regarding current public opinion climate, are an important force in public opinion dynamics that motivate political actors to act both in the public eye and in the corridors of power. These findings extend a growing literature on the influence of presumed influence, apply it to the domain of policy making and political communication, and highlight the central role of perceived media influence in the functioning of modern democracies such as Israel. In essence, the findings demonstrate that at least some of the influence of media on political life is indirect and operates through (accurate or inaccurate) perceptions of their power.

Appendix A: Measures

PRESUMED MEDIA INFLUENCE ON THE PUBLIC

It was measured using two self-report items. The first was worded, “In your opinion, to what extent do Israeli media influence your voters?” The second item referred to presumed media influence on “citizens’ voting patterns.” Both items had answer categories varying between “1” for “no influence” and “5” for “very large influence.” The variable was constructed as a latent construct influencing the two indicators (M = 3.50; SD = 0.77; r = .49; p < .001).

PRESUMED MEDIA INFLUENCE ON POLITICIANS

It was measured using a single item, worded: “In your opinion, to what extent do Israeli media influence other Knesset Members?” Answer categories varied between “1” for “no influence” and “5” for “very large influence.” This variable

Соседние файлы в предмете [НЕСОРТИРОВАННОЕ]