![Sample our Humanities journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5939/TOC-Subject-Sample-2021-Humanities.jpg"})
Abstract
New media dramatically increase citizens' access to information and decrease governments' ability to control the flow of communication. Although human rights nongovernmental organizations have advocated that access to independent news media will improve government respect for human rights, recent empirical studies have shown this is not always the case. We posit that media independence and the presence or absence of democratic characteristics, in particular political competition, have substantial effects on government repression because these factors determine the degree to which the government is vulnerable to public pressures. The model developed here includes three equations that encompass the impact of interaction between and among the news media, citizens, and government. The first equation specifies the influences on the news media's decision whether or not to perform a “watchdog” role regarding government repression. The second equation represents public reaction to the news media's coverage of government repression (i.e., protest). Here access to news media via traditional and new media is an important factor. The third equation represents government repression. Solutions to the system of equations are derived for four scenarios (a) Democracy and media independence are both present, (b) democracy is present but media independence is absent, (c) democracy is absent (autocracy) and media independence is present, and (d) democracy is absent (autocracy) and media independence is absent. We then consider interesting properties of the anticipated behavior from the government, media, and general public through case illustrations for the Netherlands and Myanmar/Burma.
[Supplementary material is available for this article. Go to the publisher's online edition of Political Communication for the following free supplemental resource: two additional case illustrations (Tanzania and Brazil).]
Acknowledgments
The authors wish to thank Babak Bahador, Shawna Sweeney, Nicholas Weller, and three anonymous reviewers and Shanto Iyengar at Political Communication for their valuable feedback on this project. Any remaining errors are the authors' own.
Notes
An appendix to this article is featured as an online supplement at the publisher's Web site.
1. Our conceptualization of news media incorporates any medium that is used to gather, produce, and transmit news. Thus, we use the term “media” to apply to print, broadcast, and online media.
2. By institutionalized political competition, we mean free and fair elections and competitive recruitment of executives. We are not referring to political competition in the form of threatening behavior at the domestic level, such as protests, general strikes, and armed rebellion. As we will discuss in some detail later in this article, we conceptualize media independence as the interaction between media freedom from government control and media freedom from commercialization.
3. Research on repression and protest is more extensive than what is covered by the following review. Studies of the repression-protest nexus per se produce support for investigations based on rational choice (CitationLichbach, 1987; CitationShellman, 2006). Case studies portray states as “purposive actors” with regard to responding to protests (CitationMoore, 2000, p. 120). These diverse cases include the Iranian Revolution (CitationRasler, 1996), state repression and accommodation of protest in Peru and Sri Lanka (CitationMoore, 2000), and government cooperation and conflict with dissidents in Chile and Venezuela (CitationShellman, 2006).
4. Interestingly, when it comes to human rights, findings from quantitative studies are sometimes at odds with those of qualitative studies. This is especially the case with research on the effectiveness of international human rights instruments, where quantitative studies indicate that these instruments do not have the desired effect and qualitative studies suggest that they do (see CitationHafner-Burton & Ron, 2009). CitationHafner-Burton and Ron (2009) recommend that one way to address these discrepancies is to employ mixed methods, which we do in this study.
5. The media also can perform several functions in conflict resolution. See CitationGilboa (2008, pp. 463–468) for an account of signaling and communication, mediation, confidence-building, and promoting agreements as such functions.
6. Ownership does not always translate to editorial control. For example, the BBC is owned by the state but is editorially independent. Indeed, publicly owned media in democracies generally are more free from commercial pressures than those that are privately owned. Thus, in democracies we might expect publicly owned media to play more of a watchdog role. In comparison, state-owned media in non-democracies are unlikely to criticize government.
7. In Mexico, for example, the reporting of activist media actually influenced the mainstream media. CitationLawson (2002) found that in the 1990s fringe media were the first to cover scandals of government corruption, but when it became apparent that Mexican audiences had an appetite for scandals, the mainstream media began to publish exposes of government abuse. CitationLawson (2002) argued that providing this sort of coverage benefited journalists because uncovering scandals helped to launch careers. Therefore, while the presence of activist media would likely promote watchdog reporting and minimize the effects of commercialization, the absence of activist media would leave only media that are closely tied to and dependent on government sources, thereby intensifying the effects of commercialization wherein media owners are more concerned about securing the next story and maximizing profits. A more general point of caution comes from CitationDavenport and Ball (2002, p. 428) on the subject of data gathering; their study of repression in Guatemala reveals variation among newspapers, human rights organizations, and interviews in reporting “statistical patterns of state violence.”
8. The use of new media to mobilize protests is an emerging area of research. For example, CitationBennett, Breunig, and Givens (2008) found that the use of digital communication (e-mail, lists, Web sites, texting) facilitated transnational protests against the war in Iraq. Yet, as the government response to the protests in Iran suggests, new media can be silenced and users of new media remain vulnerable to intimidation.
9. The effect of media freedom, however, varies by regime type. CitationWhitten-Woodring (2009) finds that media freedom only improves government respect for human rights in the most consolidated democracies and that it actually is associated with increased repression in autocracies. The reason for this variance, as posited by CitationWhitten-Woodring (2009), is that autocratic governments are not vulnerable to public opinion. Citizens, therefore, have no institutional outlets other than protest with which to hold the government accountable and push it to be responsive.
10. CitationGehlbach and Sonin (2008, pp. 1–2) conceptualized media freedom in two dimensions, media ownership (state or private ownership) and media bias (meaning “the extent to which the media misreport the news in favor of government interests”), and developed a formal model of government control of media. They found that “large advertising markets reduce media bias in both state and private media but increase the incentive for the government to nationalize private media.” Our conceptualization of media independence is more nuanced in that it includes both government control (which incorporates ownership and the type of influence implied by Gehlbach and Sonin's concept of media bias) and commercialization.
11. This figure is a kernel density plot that shows the frequency of cases of free and controlled media across a range of regime types. Controlled and free media are measured using the Global Press Freedom data set (CitationVan Belle, 2000). Regime characteristics are measured using the Polity Index, which ranges from −10 (most autocratic) to 10 (most democratic) (CitationWhitten-Woodring, 2009).
12. Prior studies that focus on media openness, democracies, and interstate conflict processes include CitationVan Belle (2000) and CitationChoi and James (2005, 2006).
13. One other qualification concerns the simultaneous nature of cause and effect in the equations that follow; no lags or leads appear. While it is understood that such temporal elements occur in the real world (e.g., ), the model abstracts away from them at this point for two reasons. One is the tradeoff between the desire to include a reasonably comprehensive list of variables in each equation and the need for tractability. With the inclusion of time lags, it would become necessary to reduce the number of variables in order to permit solutions that remain comprehensible. The other reason is that no empirical version of our equations as yet exists (i.e., a three-equation simultaneous model). A desire to facilitate empirical testing of just that kind reinforces the need to start with simplicity and build in complexity as solutions are derived.
14. Observe that R is not bounded. Therefore, as it reaches values further below zero, the impact of the squared term in the equation is increasingly negative. This differs from the role of D in the third equation; the squared component there covers the interval from 0 to 1 inclusive because its values span that same range.
15. We acknowledge that there is much debate about the impact of repression on dissent. CitationFrancisco (1995, p. 277) reports that the “standard inverted-U hypothesis is supported weakly in Czechoslovakia, marginally negated in the GDR and invalid in the Intifada.” Likewise, there is debate about the impact of information about repression on dissent. For example, in looking at leftist protests in Japan and the United States, CitationZwerman and Steinhoff (2005, p. 87) found that “draconian responses on the part of the government served as a stimulant” and facilitated mobilization. What if, however, the level of repression within a given cultural setting is not yet at the level required to produce inversion? This query is not meant in a normative sense, that is, the Israeli and GDR governments have been “nice” to their protesters. Instead, within a given time frame, it may be that we are looking at just the range of data for which the association between repression and protest remains positive. Further observations that include higher levels of repression may be needed to bring out the curvilinear relationship in some contexts, while acknowledging that refutation remains a possibility even there.
16. State of the art empirical research on repression and democracy produces a trichotomy. Along a 10-point scale for democracy, categories 0 to 7 show no connection, categories 8 and 9 reveal “some negative impact” on repression, and category 10 exhibits “a strong negative effect on state repression” (CitationDavenport & Armstrong, 2004, p. 584). The relatively simple squared term is consistent with these results, but subsequent formal modeling could take the precise intervals just noted and probe for differences in the results obtained for W, R, and P.
17. Empirical studies have found in particular that increased dissent leads to increased state coercion (CitationCarey, 2006; CitationGurr & Lichbach, 1986).
18. We say approaching rather than equal to infinity because the corner solution is undefined due to division by zero (i.e., when D equals exactly 1). This is appropriate because even a perfect score for an observed state would not equate with perfection. In principle, it makes sense for the corner solution to sit at infinity because only a utopia would warrant such classification for either protest (public voice) or watchdogging (media conduct) or other positively valued aspects of society. In other words, there is no perfect democracy, so perfection cannot be reached for dimensions associated with that overarching concept.
19. We had hoped to use the Freedom House score for the economic environment as a proxy for commercialization, but found that for our purposes this measure focuses too much on economic constraints imposed by the state and not enough on the pressures of commercialization. For example, the news media in the United States have been widely criticized for failing to report independently because of intense commercialization and a focus on profit maximizing rather than public service (CitationBennett, 1990; CitationBennett et al., 2007; CitationHamilton, 2004). Yet, the Freedom House subscore for the U.S. media's economic environment is consistently low, meaning that it is relatively free from economic constraints. While the U.S. media do not typically depend on state advertising or subsidies, they are subject to increasing concentration of ownership and pressures to turn a profit.
20. Although CitationVan Belle's (1997, Citation2000) Global Media Freedom data set goes back to 1948, we needed more qualitative information about our cases. Freedom House began providing qualitative descriptions of media environments with its 2002 report covering the year 2001, and we had access to qualitative information for the year 2000 from a forthcoming update of the Global Press Freedom data set.
21. This appendix is available online or upon request from the authors.
22. We could have used the Political Terror Scale (CitationGibney, Cornett, & Wood, 2011), which like the CIRI data set is coded from U.S. State Department and Amnesty International reports. We chose to use CIRI because it provides a breakdown of indicators for different kinds of physical integrity rights, though in the end we used the State Department, Amnesty International, and Human Rights Watch reports to get more detailed information.
23. For protest data, we could have used CitationBanks's (2008) Cross-National Time-Series Data Archive, which includes several protest variables that are largely based on New York Times coverage, but since we needed more qualitative information, we decided to go directly to the source. Initially we searched Lexis Nexis for news stories in “major world publications,” but this produced thousands of results, many of them duplicate accounts of the same events, so we narrowed our search to articles from the New York Times. We acknowledge that this limits the variation of perspectives on the protest stories (CitationDavenport, 2010), but since our aim here is to compare the relative levels of protest across cases, we decided this limitation was acceptable. Our searches yielded numbers that roughly support our model's predictions for protest (maximum protest in democracies, some protest in non-democracies with controlled media, and minimum protest in non-democracies with independent media), with 506 stories for the Netherlands, 715 for Brazil, 315 for Myanmar (Burma), and 108 for Tanzania; however, many of the articles were duplicates or false positives (articles that contained both terms but were not actually about a protest within the given country). Also, we suspect that the New York Times is more likely to focus on events in the Netherlands and Brazil, given the close ties between the United States and those countries, than on events in Myanmar and Tanzania. Consequently we do not focus on the quantity of stories, but rather the qualitative description of events, to get an overall view of the type and level of protest activity.
24. The Netherlands typically ties with a couple of other European countries for first place on the Reporters Without Borders Press Freedom Index. The exceptions to this are 2007 and 2008, when the Netherlands tumbled to 12th and 16th place (respectively) in the rankings after the government held two journalists in custody for 2 days for failing to disclose their sources. The Netherlands moved up to 7th place in 2009 and was back at the top in 2010 (Reporters Without Borders, 2010).
25. For example, in 2005, Dutch public broadcasting received more than 66% of its funding from public income and about 22% from advertising revenue; however, in 2006 it received about 55% of all funding from public income and 29% from advertising income (CitationTrappel et al., 2011). Whether this change signifies a shift in reliance on advertising funds remains to be determined.
26. As a consolidated democracy with relatively independent media, the Netherlands approaches but does not quite reach the scenario where both democracy and independent media are at their highest level of 1.
27. The CIRI Physical Integrity Rights Index (Cingranelli & Richards, 2010) placed the Netherlands at the level of 8 (indicating that the government has complete respect for these rights) for the years 2000 to 2007 and at 7 for 2008 and 2009.
28. Satellite television does provide access to international news networks, but these broadcasts are for the most part in English and are focused on international rather than Burmese news.
29. The one exception to this was 2002, when Myanmar was scored a 3, which is still a relatively low score (Cingranelli & Richards, 2010).
30. The Democratic Voice of Burma does make use of undercover reporters from within Myanmar, 14 of whom have been incarcerated for their work (CitationCommittee to Protect Journalists, 2010).
31. This is certainly due in part to the difficulty in getting information about anti-government activities within Myanmar, but even so, the relative lack of news reports of protests in Myanmar does indicate a comparatively low occurrence of protest.