Show Summary Details

Additional citations and updates to text.

Updated on 11 January 2018. The previous version of this content can be found here.
Page of

PRINTED FROM OXFORD HANDBOOKS ONLINE (www.oxfordhandbooks.com). © Oxford University Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

date: 23 May 2019

Online News Consumption in the United States and Ideological Extremism

Abstract and Keywords

In an earlier study, the authors found evidence that supported a framework predicting that consumers of Internet news sources held more extreme political views and were interested in more diverse political issues than those who solely consume mainstream television news using data covering the period April 2000 to June 2007. In this essay, they test whether the same patterns hold using data from the 2008 National Annenberg Election Survey online panel conducted during the 2008 presidential election cycle. The authors combine insights from theories of selective media exposure from political communication and social psychology with economic theories of differentiated products markets to develop a theoretical framework for understanding how the Internet continues to impact the U.S. political news market. The driving force behind this framework is the dramatically lower cost of production for Internet news sources relative to traditional television news.

Keywords: political news, polarization, partisanship, ideology, Internet, selective exposure, online news

Introduction

The news media landscape is evolving, expanding, and becoming increasingly fragmented. Over the past two decades cable news has replaced broadcast network news as the source for national and international news. Also during this time, the ideologically conservative Fox News Channel has become the leading news cable channel (Morris, 2007; Pew, 2008). This shift coincided with an unprecedented expansion in Internet access. It is increasingly common for consumers to use the Internet as their primary source of news and information (Bimber and Davis, 2003; Davis and Owen, 1998; Gaskins and Jerit, 2012; Pew, 2011, 2015; Tewksbury, 2006).

In this essay, we combine insights from theories of selective media exposure from social psychology and political communication with economic theories of differentiated products markets to develop a theoretical framework for understanding how the Internet continues to affect the U.S. political news market. The driving force behind this framework is the dramatically lower cost of production for Internet news sources relative to traditional television news. Lower cost of production allows Internet news providers to profitably provide content to consumers with more diverse and less centrist political views. Combining this insight with the concept of selective media exposure—the idea that consumers tend to seek out media that conform to their own political tastes—leads to the testable predictions that consumers of Internet news sources should, on average, hold political views that are farther away from the center, and be (p. 810) interested in more diverse political issues than those who solely consume mainstream television news.

In an earlier study, we found evidence that supported these predictions using data covering the period April 2000 to June 2007 (Nie et al., 2010). Here we test whether the same patterns hold using data from the 2008 National Annenberg Election Survey online panel (NAES) conducted during the 2008 presidential election cycle. It is important to determine whether the same pattern continues to hold as the Internet grows as a regular source of news for more and more citizens.

We believe this research is important for several reasons. First, our framework generates new predictions about the relationship between individuals’ news viewership and their political views. Because the Internet allows consumers to fit their news exposure to their own political preferences, these predictions tie our study to the large and emerging literature on political polarization. Second, these results highlight the importance of considering interaction effects when studying the impact of the Internet (Bimber, 2005). Third, and perhaps most important, our framework can give theoretical grounding for future work on the consequences of changes in the political news market.

Previous Empirical Research

Much of the previous empirical work on the relationship between consumers’ political attitudes and views and their news source(s) comes out of the literature on the political fragmentation of news consumers. Sunstein (2001) has argued that the Internet will lead to fragmentation and what he calls “balkanization.” Rather than operating in an open society of diverse ideas and discussion, citizens interact in an echo chamber, limiting their discussion and interaction to those whose opinions are similar to theirs, where there is little opportunity for their ideas to be challenged. In Sunstein’s view, the diversity of communication options and increased choice will lessen the opportunities for common public experiences, shared realities and effective public deliberation.

The growth of “new” media has greatly expanded the number of available channels of communication, leading to increased audience fragmentation. Davis and Owen (1998), for example, compare the attitudes of talk radio listeners, television news magazine viewers, and those who acquire news and political information on the Internet using data from the 1996 American National Election Study (ANES) and found significant differences among these groups. More recently, in a national survey of 1,506 adults, Mardenfeld et al. (2006) found that self-identifying liberals and moderates were less likely to choose the Fox News Channel than they were to choose ABC, CBS, NBC, and CNN for their news. Those who self-identify as conservatives were more likely to choose Fox News. Similarly, Morris (2007) found, using surveys from Pew conducted in 2004 and 2005, that Fox News Channel viewers had distinct attitudes toward both President Bush and his opposition. Stroud (2008), using data from the 2004 NAES, showed that (p. 811) people’s political beliefs are related to their media exposure, a pattern that persists across media types (newspapers, political talk radio, cable news, and Internet).

Literature on political fragmentation and the media news market primarily has focused on network, cable, and talk radio, with very few studies done about the Internet (see also, e.g., Dimmick et al., 2004; Jones, 2001; Mardenfeld et al., 2006; Morris, 2005). While Davis and Owen (1998) did include the Internet in their study, their research used data collected well before the Internet explosion and the rise of such cable news channels as Fox News (CNN has been around since the 1980s) so it is unclear whether their results still describe the situation today. Yet more recent research (Tewksbury and Riles, 2015) shows evidence of Internet news consumption associated with widening areas of disagreement between Democrats and Republicans.

Our prior work on which this essay is based (Nie et al., 2010) was among the first to study the relationship between a consumer’s source of news and his or her political attitudes while looking simultaneously at network, cable, and Internet information sources. Here, we follow the theoretical framework of the political news market we developed for that study and test predictions derived from that framework using more recent data.

Theoretical Framework

Premise I: The Demand Side—Selective Exposure

The first premise underlying our model is that consumers tend to expose themselves to news content that covers issues they care about and is in line with their own political views. This premise is supported by a long line of literature from social psychology and public opinion and communications research.

Several political studies, going back to the seminal work by Lazarsfeld et al. (1944) and Berelson et al. (1954), point to evidence of selective exposure when it comes to seeking political information. Selective exposure is “any systematic bias in audience composition.” (Sears and Freedman, 1967, 195) The psychological theory of cognitive dissonance, developed by Festinger (1957) is the foundation for selective exposure. Festinger posits that people seek out reinforcing messages to reduce and avoid dissonance.

In the field of communication, the concept of selective exposure was incorporated into the “limited effects” model of media exposure developed by Klapper (1960). In this model, which dominated the field for many years, selective exposure was used as a means for asserting that the media have little to no effect on social and political behavior. “Attitude predispositions largely determine the communications to which the individual is exposed” (Klapper, 1963, 67). Individuals seek information that supports rather than disputes their point of view (Sears and Freedman, 1967). Voters sought information that conformed to one’s existing values and predispositions (Berelson and Steiner, 1964). For example, Schramm and Carter (1959) found that Republicans were more likely than Democrats to watch a Republican-sponsored telecast.

Over time, the evidence proved to be more equivocal about whether people selectively seek out only information reinforcing their own view. Critics of selective exposure include (p. 812) Sears and Freedman (1967), who, in a review of prior research, argued that no empirical case had been made for its existence. Much of the research following their review has reached similar conclusions, citing other factors explaining media exposure patterns (e.g., Cotton, 1985; McCombs and Becker, 1979; O’Keefe and Atwood, 1981; Severin and Tankard, 1979). Chaffee and Miyo (1983) found that while selective exposure did occur, the underlying assumptions of this theory were questionable. Selective exposure was found mainly among those with the lowest political involvement and electoral experience.

Years later, D’Alessio and Allen (2002, 2007) conducted a meta-analysis of experimental studies testing selective exposure and provided evidence to support the notion that cognitive dissonance is weakly associated with it. Criticizing prior reviews, they argue, for example, that flaws in Sears and Freedman’s (1967) meta-analysis explain their inconsistent results. Despite criticisms of the studies, researchers who found other explanations never fully discounted the idea of selective exposure as a means of ideological reinforcement. (O’Keefe and Atwood, 1981; Severin and Tankard, 1979).

More recent studies point to evidence that selective exposure does contribute to media effects. Chaffee et al. (2001) found that people were somewhat more likely to pay attention to information about their preferred candidate. Taber and Lodge (2006) concluded that when faced with pro and con arguments, people uncritically accept the arguments they support, and counter argue the ones which they oppose. When they are given the option to self-select a source of information, they pick the ones which are most likely to confirm their arguments. In a political advertising study of the 1996 campaign, Kaid (1997) shows that Democrats became more positive toward Bill Clinton after seeing a Clinton advertisement and Republicans became more positive toward Dole after seeing one of his political spots. Stroud (2007) suggests that selective exposure may play a role in increasing polarization, finding that those who viewed Michael Moore’s film “Fahrenheit 9/11” had significantly higher negative feelings toward President Bush than those who intended to see it but had not.

Studies show that those who rely on the Internet for information have not abandoned mainstream sources altogether. Still, the Internet is distinctly different from network television and cable news because it is an interactive medium that promotes more than a one-way communication. “The Internet does not represent a singular mode of communication, but a flexible and adaptable set of opportunities for communication that can be exploited by individuals and groups in many ways” (Bimber, 2005, 16). As a result, Bimber argues, it is difficult for researchers to determine a single main effect.

In an early study of online use, Johnson and Kaye (1998) found that those online tended to be less trustful of government and less likely to vote. While the Internet had a positive and significant impact on political interest, there was a negative impact between reliance on the Internet and trust in government, efficacy, and voting behavior (1996 elections).

To be clear, we do not argue that selective exposure is all that drives media selection. There are other factors such as interest in news and current affairs that are also likely to affect the sources of news one seeks. Our claim is that, ceteris paribus, people will seek out news presented by sources with an ideological slant similar to their own (Iyengar and Hahn, 2009).

(p. 813) Premise 2: The Supply Side—Internet News Saturates the Taste Space

The second premise is that the news content available on the Internet covers a wider range of political opinions and issues than that available from more traditional news sources. This assertion is theoretically grounded in the large difference in cost of production for Internet news sources relative to television news sources (Baum, 2003; Hamilton, 2004; Prior, 2007) and the need for news producers to recover their production costs.1 When production costs are high, a large number of consumers are needed before profits can be made and so the market can only support a few producers. When only a few producers enter the market, they place themselves in a position designed to reach a lot of consumers, typically toward the center of the distribution. When costs are low, producers need fewer consumers to be viable, so more producers will enter the market and, among them, cover a greater range of locations in the market space. We argue that it is reasonable to view the Internet as a medium that effectively saturates the entire taste space for political news because of its low production costs.

The models of differentiated products markets from the field of economics provide theoretical support for why lower costs increase competition and cause programming to become more diverse. Originally developed by Hotelling (1929), these models describe behavior in markets where consumers have tastes that are distributed according to some distribution function and prefer products close to their ideal point in the distribution; producers in these markets select product positions so as to maximize profits. One of the key results in the product differentiation literature is that when people consume the product nearest to their ideal point (i.e., when selective exposure occurs), lower production costs leads to saturation (see Eaton and Lipsey, 1989, for a comprehensive, if somewhat outdated, survey of the literature).

Figure 55.1 provides a graphical illustration of our argument. The mainstream television news sources are all located relatively close to the center where they can attract a large enough audience to cover their high production costs. The Internet, however, has saturated the taste space by providing news from all different ideological perspectives.

Online News Consumption in the United States and Ideological ExtremismClick to view larger

Figure 55.1. Graphic representation of the model. Here, the distribution of political views within the population represents the local demand for political bias, and high cost mainstream media locate at discrete points close to the center of this distribution. Due to the very low cost of production and distribution of Internet news, there are minimal barriers to entry that lead us to model the Internet as filling in the entire continuum and saturating the taste space for political bias. Thus, when viewing online content, consumers can choose news that fits their political opinion exactly.

Together these premises predict that individuals whose ideological preferences are not completely served by mainstream television news sources (because their views are farther from the center) will tend to search the Internet to get news that is more complementary to their views. We test this prediction by looking at the news consumption of those getting news from television sources that appeal most to the viewers on the left (CNN and MSNBC in our dataset) and right (Fox News in our dataset).

For the purposes of our model, it is useful to divide those who are using noncentrist cable television news sources into three groups. Let us use Fox News viewers as an example. There are viewers who are more conservative than Fox News, viewers who have roughly the same ideological position as Fox News, and viewers who are more liberal than Fox News. Viewers who are more liberal than Fox and those with roughly the same (p. 814) ideological position should be able to satisfy their ideological preferences with the coverage they can find on television (either on Fox or more liberal sources). However those Fox News viewers who are more conservative than Fox News cannot supplement their news consumption by using an additional television news source that is more conservative than Fox. Because the most widely available option open to these consumers is the Internet, we should find that the most conservative Fox viewers supplement their Fox News viewership with Internet news content.

  1. H1: Of those who watch the Fox News Channel, individuals who also get news from the Internet will be more conservative than other Fox News Channel viewers.

Similarly for those using CNN and MSNBC (the most liberal television sources in our study):

  1. H2: Of those who watch CNN/MSNBC news, individuals who also get news from the Internet will be more liberal than other CNN/MSNBC news viewers.

As H1 and H2 suggest, one motive for seeking Internet news can be a search for greater ideological purity in issue coverage. Interest in issues not covered by mainstream broadcast news may provide another motive. Mainstream broadcast news is time-constrained and thus chooses to cover a subset of issues most viewers find important. In contrast, Internet news can focus on niche issues. Therefore, people who are interested in more diverse political issues are more likely to seek out news on the Internet. People interested in a broad range of political issues are more likely to identify a low salience issue (p. 815) as the single most important current issue than people with a narrow range of political interest. Combining these two observations about people with a broad range of political interests yields our final hypothesis:

  1. H3: Individuals who use the Internet for news should be more likely to identify low-salience political issues as important than those who rely solely upon mainstream television content.

Research Design and Results

To address these hypotheses, we analyze data from the 2008 NAES online panel. For our analyses looking at citizens’ ideology/partisanship we use a sample that includes about 4,000 regular Fox News viewers, 2,000 regular CNN viewers and 1,000 regular MSNBC viewers. Because the sampling scheme employed in administering surveys can potentially cause certain groups to be overrepresented or underrepresented in the sample, we use rim weights to adjust our sample composition to match that of the national population. We also present the results of these analyses when using unweighted sample means and running OLS regressions with all available demographic variables as controls.

Recall that our theoretical framework suggests that consumers whose ideological perspectives are not adequately covered by the mainstream media should be more likely to use the Internet for news. Using this insight, we predict that among those who watch news from the rightmost television news source, the Fox News Channel, those who also get news from the Internet will be more conservative than Fox News viewers who do not cite the Internet as a news source (see H1). To test this hypothesis, we limited the sample to those who responded that they watched a regularly scheduled Fox News program (“Fox News Sunday” and/or “The O’Reilly Factor”) frequently2 and compared the ideological position of consumers who searched the Internet for news on at least a weekly basis3 to those who searched the Internet for news less frequently.

We measured the ideological/partisan position of respondents in two different ways. First, we used their self-identified liberal-conservative position on a 7-point scale4 (with increasing values indicating increasing levels of conservatism). Second, we used respondents’ self-identified five-point party identification, where increasing values indicate increasing levels of attachment to the Republican party.5 For both measures, we calculated the difference between Fox viewers who consume Internet news and those who do not. The results are presented in Table 55.1 and show that Fox News viewers who also search for news online are more conservative and have a stronger attachment to the Republican Party than do Fox News viewers who do not use online news sources. The estimated differences are substantively significant. For example, the difference in (p. 816) conservatism between Fox News viewers who search the Internet for news and those who do not is nearly a full point on a 6-point scale. Given that the sample is limited to those who watch Fox News, and so are located right of center, this large of a difference is striking.

Table 55.1. The Ideological/Partisan Difference between Internet News Users and Non-Internet News Users Among Fox News Viewers

DV = Party Identification

Weighted Means

Unweighted Means

OLS

Difference

0.84**

0.86**

0.55**

Std. Error

0.11

0.05

0.06

N

2,233

3,875

3,546

DV = Liberal-Conservative Position

Weighted Means

Unweighted Means

OLS

Difference

0.99**

0.98**

0.77**

Std. Error

0.13

0.06

0.06

N

2,222

3,850

3,523

Notes:

(*) significant at 10% level

(**) significant at 5% level

((***)) significant at 1% level.

Control variables included in the OLS regression include controls for age and dummies for gender, marital status, race / ethnic groups, levels of education, levels of religious participation, and income levels.

We found similar results when we tested our hypothesis that frequent6 CNN viewers (i.e., those who watched “The Situation Room with Wolf Blitzer” and/or “Late Edition with Wolf Blitzer”) and MSNBC viewers (i.e., those who watched “Hardball with Chris Matthews”) who also get news from the Internet are more liberal than CNN/MSNBC news viewers who do not. As with the analysis of Fox News viewers, we measured respondents’ positions by using both their self-identified liberal-conservative position and their party identification. The results are presented in Table 55.2 and show that CNN and MSNBC viewers who also search for news online are more liberal and more attached to the Democratic Party than their counterparts who do not. The results are substantively significant but because of the smaller sample sizes, the differences do not always achieve statistical significance. Still, the difference is always in the predicted direction, is substantively significant, and in 75% of the cases, statistically significant. On balance the results provide evidence that the CNN/MSNBC news viewers who also use the Internet for news are more ideologically extreme than those who do not use the Internet to get news.

Table 55.2. The Ideological/Partisan Difference between Internet News Users and Non-Internet News Users Among CNN and MSNBC News Viewers

CNN Viewers

DV = Party Identification

Weighted Means

Unweighted Means

OLS

Difference

–0.34

–0.36**

–0.45**

Std. Error

0.24

0.12

0.15

N

1,095

1,977

1,729

DV = Liberal-Conservative Position

Weighted Means

Unweighted Means

OLS

Difference

–0.94**

–0.75**

–0.49**

Std. Error

0.36

0.15

0.17

N

1,092

1,963

1,716

MSNBC Viewers

DV = Party Identification

Weighted Means

Unweighted Means

OLS

Difference

–0.51

–0.41

–0.44

Std. Error

0.29

0.21

0.29

N

582

1,098

959

DV = Liberal-Conservative Position

Weighted Means

Unweighted Means

OLS

Difference

–0.95*

–0.82**

–0.45

Std. Error

0.24

0.21

0.28

N

580

1,093

955

Notes:

((*)) significant at 10% level

((**)) significant at 5% level

((***)) significant at 1% level.

Control variables included in the OLS regression include controls for age and dummies for gender, marital status, race / ethnic groups, levels of education, levels of religious participation, and income levels.

Our final hypothesis, H3, is that those who use the Internet for new, whether or not they also use television sources, will be more likely to identify low-salience issues as (p. 817) being important than those who rely solely on television sources for news content.7 For the dependent variable in the analysis we use whether the respondent choose the “Other” category when responding to the question “In your opinion, what is the most important issue facing the U.S. today?” Respondents could choose one of seven specified answers (Taxes, Education, War on terrorism, Situation in Iraq, Economy/Jobs, Moral issues, Healthcare) or choose “Other (please specify:____).” Because the available answers to the question represent the issues that are dealt with most frequently in the mainstream media, the likelihood of choosing “Other” represents interest in a wider variety of issues than those available in mainstream media. Table 55.3 reports the (p. 818) difference, in percentage points, of the weighted means between the two groups. In this case, there is no evidence for H3. The estimated differences are all small (sometimes as low as 0.3 percentage points) and statistically insignificant.

Table 55.3. Difference between Internet News Users and Non-Internet News Users in Terms of the Diversity of Issues Considered Important

Likelihood of Identifying “Other” as Most Important Issue

Weighted Means

Unweighted Means

OLS

Difference

0.3

0.6

0.4

Std. Error

1.6

0.7

0.8

N

2,544

5,544

4,909

Notes:

(*) significant at 10% level

(**) significant at 5% level

(***) significant at 1% level.

Control variables included in the OLS regression include controls for age and dummies for gender, marital status, race / ethnic groups, levels of education, levels of religious participation, and income levels.

Discussion

In this essay, we have followed up on an earlier study looking at how the Internet has changed the political news market (Nie et al., 2010). The key theoretical contribution from that earlier study, which we have reiterated here, is that the reduced costs of producing and distributing news, combined with the consumers’ tendency to selectively expose themselves to media with which they agree, has changed the U.S. political news market by giving consumers more control over their information environment. From this model we derived three observable implications about the difference between those who use the Internet for news and those who do not. In our initial report, we found evidence for all three hypotheses when using data collected between April 2000 and June 2007.

Since the Internet has grown and become a more regular source of news for more and more citizens, we wanted to see whether these patterns continued to hold using data from the 2008 presidential election. We found further support for the ideological extremity hypotheses (i.e., H1 and H2). Those who use the Internet to supplement their consumption of news from the most noncentrist television sources (CNN, MSNBC, and Fox) are farther from the ideological center than their counterparts who do not. However, we no longer find evidence that those who use the Internet for news are interested in a broader array of issues than those who do not (i.e., H3). We suspect that this reflects that fact that as the Internet becomes the primary source of news for more and more individuals, the topics covered by Internet and non-Internet sources are similar. (p. 819) Instead, the difference is the slant, with the Internet providing content that is more ideologically extreme.

References

Baum, M. A. 2003. Soft news goes to war: Public opinion and American foreign policy in the new media age. Princeton, NJ: Princeton University Press.Find this resource:

Berelson, B., Lazarsfeld, P. F., and McPhee, W. F. 1954. Voting: A study of opinion formation in a presidential campaign. Chicago, IL: University of Chicago Press.Find this resource:

Berelson, B., and Steiner, G. A. 1964. Human behavior: An inventory of scientific findings. Oxford, England: Harcourt, Brace & World.Find this resource:

Bimber, B. 2005. The Internet and political fragmentation. Paper prepared for presentation at the Democracy in the 21st Century Conference, University of Illinois Urbana-Champaign.Find this resource:

(p. 820) Bimber, B., and Davis, R. 2003. Campaigning online: The Internet in U.S. elections. New York, NY: Oxford University Press.Find this resource:

Chaffee, S. H., Saphir, M. N., Graf, J. Sandvig, C., and Hahn, K. S. 2001. Attention to counter-attitudinal messages in a state election campaign. Political Communication, 18(3), 247–272.Find this resource:

Chaffee, S., and Miyo, Y. 1983. Selective exposure and the reinforcement hypothesis: An intergenerational panel study of the 1980 presidential campaign. Communication Research, 10, 3–36.Find this resource:

Cotton, J. 1985. Cognitive dissonance in selective exposure. In D. Zillmann and J. Bryant (Eds.), Selective exposure to communication. Hillsdale, NJ: Lawrence Erlbaum Associates.Find this resource:

D’Alessio, D., and Allen, M. 2002. Selective exposure and dissonance after decisions. Psychological Reports, 91(2), 527–532.Find this resource:

D’Alessio, D., and Allen, M. 2007. The selective exposure hypothesis and media choice processes. In R. W. Preiss, B. M. Gayle, N. Burrell, M. Allen, and J. Bryant (Eds.), Mass media effects research: Advances through meta analysis (pp. 103–118). Hillsdale, NJ: Erlbaum.Find this resource:

Davis, R., and Owen, D. 1998. New media and American politics. New York, NY: Oxford University Press.Find this resource:

Dimmick, J., Chen, Y., and Zhan, L. 2004. Competition between the Internet and traditional news media: The gratification-opportunities niche dimension. Journal of Media Economics, 17(1), 19–33.Find this resource:

Eaton, B., and Lipsey, R. G. 1989. Product differentiation. In R. Schmalensee and R. Willig (Eds.), Handbook of industrial organizations. Amsterdam: North-Holland.Find this resource:

Festinger, L. 1957. A theory of cognitive dissonance. Stanford, CA: Stanford University Press.Find this resource:

Gaskins, B., and Jerit, J. 2012. Internet news: Is it a replacement for traditional media outlets? The International Journal of Press Politics, 17(2), 190–213.Find this resource:

Hamilton, J. T. 2004. All the news that’s fit to sell: How the market transforms information into news. Princeton, NJ: Princeton University Press.Find this resource:

Hotelling, H. 1929. Stability in competition. Economic Journal, 39(1), 41–57.Find this resource:

Iyengar, S., and Hahn, K. S. 2009. Red media, blue media: Evidence of ideological selectivity in media use. Journal of Communication, 59(1), 19–39.Find this resource:

Johnson, T., and Kaye, B. 1998. A vehicle for engagement or a haven for the disaffected? Internet use, political alienation, and voter participation. In T. J. Johnson, C. E. Hays, and S. P. Hays (Eds.), Engaging the public: How the government and media can reinvigorate democracy (pp. 123–135). Lanham, MD: Roman and Littlefield.Find this resource:

Jones, D. A. 2001. The polarizing effect of new media messages. International Journal of Public Opinion Research, 14(2), 158–174.Find this resource:

Kaid. L. L. 1997. Effects of the television spots on images of Dole and Clinton. American Behavioral Scientist, 40, 1085–1094.Find this resource:

Klapper, J. T. 1960. The effects of mass communication. New York, NY: Free Press.Find this resource:

Klapper, J. T. 1963. The social effects of mass communication. In W. Schramm (Ed.), The science of human communication. New York, NY: Free Press.Find this resource:

Lazarsfeld, P., Berelson, B., and Gaudet, H. 1944. The people’s choice: How the voter makes up his mind in a presidential campaign. New York, NY: Columbia University Press.Find this resource:

Mardenfeld, S., Taylor, C. E., Kirkyla, A., and Anderson, A. 2006. Choosing media based on political stance. Presentation at the 2006 Annual Conference of the New Jersey Communication Association.Find this resource:

McCombs, M. E., and Becker, L. 1979. Using mass communication theory. Englewood Cliffs, NJ: Prentice-Hall.Find this resource:

Morris, J. S. 2005. The Fox News factor. The Harvard International Journal of Press/Politics, 10(3), 56–79.Find this resource:

(p. 821) Morris, J. S. 2007. Slanted objectivity? Perceived media bias, cable news exposure and political attitudes. Social Science Quarterly, 88(3), 707–728.Find this resource:

Nie, N. H., et al. 2010. The World Wide Web and the U.S. political news market. American Journal of Political Science, 54(2), 428–439.Find this resource:

O’Keefe, G. J., and Atwood, L. E. 1981. Communication and election campaigns. In D. D. Nimmo and K. R. Sanders (Eds.), Handbook of political communication. Beverly Hills, CA: Sage.Find this resource:

Pew Research Center. 2008. Internet’s broader role in campaign 2008. Research report. Retrieved from www.pewInternet.org

Pew Research Center. 2011. Understanding the participatory news consumer: How Internet and cell phone users have turned news into a social experience. Research report. Retrieved from http://www.pewInternet.org/~/media//Files/Reports/2010/PIP_Understanding_the_Participatory_News_Consumer.pdf

Pew Research Center. 2015. Millenials and political news. Research report. Retrieved from http://www.journalism.org/2015/06/01/millennials-political-news/

Prior, M. 2007. Post-broadcast democracy: How media choice increases inequality in political involvement and polarizes elections. New York, NY: Cambridge University Press.Find this resource:

Schramm, W., and Carter, R. F. 1959. Effectiveness of a political telethon. Public Opinion Quarterly, 31, 194–213.Find this resource:

Sears, D. O., and Freedman, J. L. 1967. Selective exposure to information: A critical review. Public Opinion Quarterly, 31(2), 194–213.Find this resource:

Severin, W. J., and Tankard, J. 1979. Communication theories. New York, NY: Hastings House.Find this resource:

Stroud, N. J. 2007. Media effects, selective exposure, and Fahrenheit 9/11. Political Communication, 24(4), 415–432.Find this resource:

Stroud, N. J. 2008. Media use and political predispositions: Revisiting the concept of selective exposure. Political Behavior, 30(3), 341–366.Find this resource:

Sunstein, C. R. 2001. Republic.com. Princeton, NJ: Princeton University Press.Find this resource:

Taber, C. S., and Lodge, M. 2006. Motivated skepticism in the evaluation of political beliefs. American Journal of Political Science, 50(3), 755–769.Find this resource:

Tewksbury, D. 2006. Exposure to newer media in a presidential primary. Political Communication, 23(3), 313–332.Find this resource:

Tewksbury, D., and Riles, J. M. 2015. Polarization as function of citizen predispositions and exposure to news on the internet. Journal of Broadcast and Electronic Media, 59(3), 381–398. (p. 822) Find this resource:

Notes:

(1.) Hamilton’s analysis, which is the closest to ours, focuses on the effects of lowered costs on the provision of “soft” vs. “hard” news (assuming that “hard” news has the positive externality related to encouraging voters to vote and assisting them in making more-informed voting decisions). In contrast, we focus on differentiation along the dimension of political bias or slant, which has a more symmetric structure of both costs and preferences.

(2.) In the case of “Fox News Sunday,” this is defined as watching this show either “every week or almost every week” or “one to three times a month.” For the “O’Reilly Factor,” a frequent watcher is someone who indicated watching the show either “every night or almost every night” or “a few times a week.”

(3.) This included respondents who indicated that they searched wither “three times a week or more” or “every week or almost every week.” Respondents were not asked what type of news they were seeking online, which websites they went to or to what extent and in which direction were these online news sources biased.

(4.) The liberal-conservative scale in the Knowledge Networks data has the following values: 1 = extremely liberal, 2 = liberal, 3 = slightly liberal, 4 = moderate, middle of the road, 5 = slightly conservative, 6 = conservative, and 7 = extremely conservative.

(5.) The respondent’s party id were determined as part of a two-stage question—the first about which party they identify with, and the second question (asked only if the response to the first one was “Democrat” or “Republican”) about the strength of party association. The scale takes the following values: 1 = strong Democrat 2 = weak Democrat, 3 = Independent, 4 = weak Republican, 5 = strong Republican

(6.) For the show “Late Edition,” a frequent viewer is defined as someone who watches this show either “every week or almost every week” or “one to three times a month.” For the shows “Hardball” and the “Situation Room,” a frequent watcher is someone who indicated watching the show either “every night or almost every night” or “a few times a week.”

(7.) Specifically, the comparison group is those who did not search the Internet for news frequently, but did report frequently watching television news programs from at least one of the following sources: national network news, CNN, MSNBC, or Fox News.