How a High Choice Media Environment Leads to Greater Selectivity, Fragmentation and Polarization

By Rebecca.Chalif April 25, 2010

Posted in 2010 Journal

Abstract:

This paper analyzes how the proliferation of new media has created a high choice media environment, and how this environment has led to media fragmentation and ultimately increased political polarization. Before the advent of talk radio, cable news channels, and most recently, the Internet, citizens received a relatively consistent news package. This fostered a common collective intelligence among the citizenry where people were exposed to all types of news from varying perspectives. Today however, there are so many different places people can go to consume the news that a degree of personal selectivity is always present. People can filter their news intake to exclude topics they are not interested in, ideologies they do not agree with, or political news all together. This selectivity has led to media fragmentation- the idea that each person will likely consume different fragments of news content. This paper employs a statistical analysis to show the relationship between people’s political ideology and the media outlet they chose to rely on. This study analyzes the frequency of people’s television and Internet consumption of two partisan news outlets- Fox News and MSNBC, and compares it to their reported political ideology. The results support the hypothesis that people’s political ideology is related to the type of news they chose to consume. The data shows that people who are conservative consume news from Fox News more frequently than those who are liberal. The opposite trend is found with respect to MSNBC, as liberals gravitate to this outlet more frequently than conservatives. These results support the idea that people chose media that agrees with their pre-existing political ideology, and that ignoring news that supports the opposite ideology is likely to increase polarization and political partisanship.

The emergence of new media and the rise of novel media outlets have vastly changed the media environment, providing audiences with countless options for news consumption. This extremely high choice environment will undoubtedly have some major implications with respect to political news (Hindman, 2009). Before the advent of talk radio, cable news channels, and most recently, the Internet, local newspapers and evening news broadcasts served as the primary outlets for political news. As a result, there was a commonality to the news people consumed regardless of their geographic location, issue positions, or ideological stance.

Today, people can choose from thousands of news outlets, each producing a different news product. The Internet alone provides news consumers with an enormous array of websites that provide varying types of news (Hindman, 2009). Citizens also have the opportunity to cater their news intake to cover only issues that interest them, or perhaps, to only expose themselves to political information that supports their point of view (Sunstein, 2007). This notion of media selectivity may have harmful effects on a collective civic intelligence that is paramount for a nation governed “by the people.” If citizens only encounter political news that reinforces their preconceived notions, compromise will become increasingly difficult, and people may not be aware of the alternate perspectives of a political issue. If political news ceases to educate citizens by providing multiple perspectives, and instead, becomes an “echo chamber” for like-minded individuals to share their similar perspectives, it is likely that the result will be an increasingly polarized citizenry (Jamieson, & Cappella, 2008). Therefore, I hypothesize that an analysis of news media consumption will show that people are more likely to select news outlets that reflect their preconceived political preferences. This relationship will be analyzed by comparing people’s political ideology with how often they watch, listen, or visit the websites of the conservative leaning news outlet, Fox News, and the liberal leaning MSNBC.

The existing literature on media fragmentation and its polarizing effects, answers many of the important preliminary questions relevant to this study. Previous research sheds a great deal of light on the implications of the new media environment, and how it relates to increased polarization among American citizens. I will use the theoretical framework provided by the literature below to explain and support my hypothesis, stating that people will gravitate toward media that echoes their political ideologies, and thus show how a high choice media environment will lead to greater media selectivity and ultimately, greater political polarization.

High Choice Media Environment

Today, it is clear that media consumers have far more news choices. Markus Prior emphasizes this drastic shift when he points out that in 1970, television provided a mere seven channels to the average household and that the three broadcast networks captured 80% of all viewing. In comparison, by 2005, 85% of households had access to cable or satellite television providing the average viewer with about 100 channels to chose from (Prior, 2005). However, this data does not take into consideration the massive proliferation of new media outlets on the Internet. Online, people can access a plethora of information, some of which may be affiliated with a major news organization, but much of which is citizen-produced content in the form of blogs and message boards. New media outlets such as these provide consumers with the broadest possible sampling of people’s opinions on almost any topic. The old media environment of limited choice seemed to encourage moderation and conformity (Jones, 2001). It was reasonable for citizens to read their daily newspaper and watch a one-hour evening news report- creating an environment where most citizens were exposed to the same news.

Today, new media sources such as cable television, talk radio, and the Internet, have begun to lure people away from the traditional mainstream media. This shift has led to an increase in “niche media”, where news outlets no longer attempt to catch the broadest possible audience by offending nobody, but rather, focus on a smaller, more targeted audience with perspectives that are likely to offend many (Jones, 2001). Unencumbered by many of the constraints of objectivity and fairness, common to mainstream media organizations (Davis, & Owen, 1998), some media outlets have moved more toward sensationalism and bias as a means of maintaining a large market share. The new media environment, consisting of countless possible outlets, as well as a much larger range of opinions, forces consumers to chose which of these outlets they will use for information gathering (Baum, & Groeling, 2008). As it is impossible to expose one’s self to all possible media sources, people must have some basis for distinguishing and choosing their preferred media outlets. This notion leads to an important element of this study—media selectivity and fragmentation.

Media Selectivity & Fragmentation

Selectivity can occur at several distinct junctures during mass media consumption. Diana Mutz describes these junctures as exposure to a particular source of news, attention to what the source says, and biased interpretation when processing the content of the news (Mutz, 2006). For the purposes of this study, I will focus my analysis on selective exposure, as this step must happen before the other two types of selectivity can occur. Media selectivity can be understood by considering what Cass Sunstein calls, “The Daily Me” (Sunstein, 2007). His book describes the phenomenon correlated with high media choice, coupled with the extreme ease of access to information. Instead of relying on a mediating institution, such as a newspaper or a television channel, people now serve as their own news aggregators (Lee, 2009). They can pick and choose what content to read over an ever expanding “media sphere” that includes information on every topic, presented from almost every point of view. Considering that every citizen could be her own personal news aggregator, it becomes likely that no two people would receive the same “news package” on any given day.

There are two types of media selectivity that are relevant to this study. The first is the selection of news content and the second is the selection of entertainment and other non-news content. This distinction is important, as it shows how the new, high-choice media environment can facilitate either an inundation of news content, or a complete avoidance of it. A citizen interested in the news now has the opportunity to access large amounts of information. These citizens gain a much more holistic understanding of a news story by following its coverage across a range of media outlets with different perspectives and contributions to the story. For example, a political “junkie” can follow any story, from the most hyper-local to the most international. They can see many journalistic perspectives on the issue at hand, as well as view the opinions of other citizens blogging or commenting on news stories. This scenario seems like an idyllic model for what Michael Schudson calls the “informed citizen” (Schudson, 1998). He proposes that good citizens must be educated about the issues in order to properly “self-govern”. However, one must also consider the possibility of the other extreme. A person with no interest in politics or news has other media options, in that they can avoid almost all political messages. In comparison, 40 years ago, a politically uninterested person would still experience a good deal of political messages through accidental exposure, as there were often few, if any, alternatives in the media (Prior, 2005). Even in an increasingly polarized society, there is still a large center of voters who are indifferent or ambivalent toward politics (Bernhardt, Krasa, & Polborn, 2008). The idea that this group can completely detach from the political world is unsettling at best, and detrimental to the functioning of our democracy at worst. Prior explains, “Since political knowledge is an important predictor of turnout and since exposure to political information motivates turnout, the shift from a low-choice to a high-choice media environment implies changes in electoral participation as well” (Prior, 2005). Highlighting this point, Mutz points out that the average size of the audience watching prime-time presidential addresses and news conferences has steadily decreased in the late 20th century (Mutz, 2006). Clearly, selectivity leads to fragmentation of people’s news packages, widening the gap between the informed and the unin

formed citizen (Tewksbury, 2005).

Another measure of media fragmentation that is of paramount importance to this study is selectivity based on political ideology. It is not surprising, that with so many options, a person’s media choices increasingly reflect their partisan considerations. Evidence suggests that people tend to seek out information consistent with their own beliefs. This body of evidence can be traced back to some of the first studies on selective exposure from the 1940s. Paul Lazarsfeld linked his work to a psychological study by Leon Festinger on the theory of cognitive dissonance. Festinger suggested, “that people want to avoid information that conflicts with their preexisting beliefs, and that they seek out information—through activities such as selective exposure—that confirms their current beliefs” (Mutz, 2006). Accordingly, a more recent study showed that the stronger a person’s partisan affiliations, the more likely they are to select media outlets that confirmed their beliefs (Iyengar, & Hahn, 2007). A study of self-reported media exposure during the 2000 and 2004 campaigns showed significant fragmentation of media use among Republicans and Democrats. Republicans gravitated toward talk radio, a medium known to have a conservative slant; while Democrats avoided talk radio and watched television newsmagazines and late-night entertainment, two predominantly liberal media outlets (Iyengar, & Hahn, 2007). Another supporting study found that individuals’ political predispositions could predict their exposure to specific media (Lee, 2009).

The results of media fragmentation based on political ideology can have many deleterious effects on our society and political system. If citizens can shield themselves from opinions that conflict with their own, they

will receive a one-sided, partisan view of an issue. This

can lead to greater problems among media consumers interested in politics, as there will be no common frame of reference about a topic. Citizens will lack information on broad issue topics, and ignore the opposing point of view on issues they consider important (Tewksbury, 2006). Fragmentation according to political ideology creates “echo chambers” in the media. Citizens’ media packages will only include consonant ideas, excluding any competing opinions. Studies on small group interactions reveal another negative effect of “echo chambers” in the media. The study shows that people who are part of groups comprised only of like-minded individuals tend to move toward greater attitude extremity (Mutz, 2006). Sunstein also notes that, “A possible consequence [of media fragmentation] is considerable difficulty in mutual understanding. When society is fragmented in this way, diverse groups will tend to polarize” (Sunstein, 2007). This notion answers a central question of this study—is increased media fragmentation related to increased political polarization?

Polarization

Before attempting to determine a relationship between media fragmentation and polarization, it proves useful to examine research about the level of polarization in American society. Polarization refers to the trend where people move further from the center of political debate and toward the extremes of political ideology. The Economist writes that, “the 50-50 nation appears to be made up of two big separate voting blocks with only a small number of swing voters in the middle” (“On his high,” 2002). Diana Mutz says, “patterns of aggregate opinion suggest that partisanship is a driving force in how people perceive, interpret, and respond to the political world” (Mutz, 2006). Clearly, scholars suggest that we are living in an increasingly polarized political society. Observations about political debate in Washington today seem to only further elucidate this assertion.

Still, it is difficult to establish a causal relationship between higher media fragmentation and an increasingly polarized citizenry. Instead, a cyclical model seems to emerge where the media may be reflecting a more polarized citizenry, while the citizens may also be affected by changes in the media. The latter seems more probable when one considers the agenda-setting role of the media (Lee, 2009). One particularly relevant area of this research is found in a study by Dan Hallin, which discusses the role of the mass media in “bracketing” the range of acceptable opinions for the public (Hallin, 1986). In his view, the media defines the range of legitimate opinions that the public can adopt. Applying this theory to the new media environment would suggest that the proliferation of new media, often with blatant and extreme partisan biases would expand the range of acceptable partisan opinions (Hindman, 2009). Ultimately this could lead to an increasingly polarized society where more extreme opinions are considered acceptable. Further, countless media studies have shown that the content of political news has become much more sensational, more focused on horse race journalism, and more concentrated on scandal and personal intrigue (Owen, 2009). What type of agenda is being set by a media that largely focuses on contentious debate between political parties, infighting, and personal scandal? Likely, it is an agenda of partisan fighting and scandal finding, rather than moderation and issue based compromise.

Several other studies also support the notion that media selectivity and fragmentation are factors in the higher levels of polarization. A 2007 study analyzing Pew research data showed that seven out of ten voters who receive most of their election news from Fox News supported Bush, the Republican candidate, while just 21% supported Kerry, the Democratic candidate. In contrast, voters who got most of their election news from CNN favored Kerry 67% to Bush 26% (Bernhardt, Krasa, & Polborn, 2008). Jamieson and Cappella also support the assertion that confining oneself to an echo chamber of news is likely to lead to increased polarization within one’s own group (Jamieson, & Cappella, 2008). David Jones examines a more specific indicator of polarization—attitudinal change as a result of media exposure. This is paramount, as it could be argued that Democrats were more likely to watch CNN than Republicans, and that their news intake did not affect their ideology. However, Jones collected aggregate attitude reports in 1992 and then again in 1996 asking people about their political ideology, as well as several measures of media consumption. He found that devoted listeners of Rush Limbaugh’s radio talk show were more likely to call themselves conservative in 1996 than in 1992 (Jones, 2001). These findings illustrate how media selectivity leads to audience fragmentation and ultimately to increased polarization of political views.

Television News

Television news analysis is a central component of this study. “Television is the most frequently used source of news for Americans” (Baum, 2003). For this reason, research into the effects of selectivity and polarization across TV media is especially important. Before the birth of the cable news network, people received news from nightly broadcasts on the three network news channels, CBS, ABC, and NBC. This media environment offered a relatively homogeneous news product across the networks, where tuning in to the nightly newscast a “social ritual” (Lee, 2009). In 1980, the television news environment changed drastically with the birth of CNN, followed by MSNBC and Fox News in 1996 (Morris, 2005). From the beginning, these cable news networks wanted to differentiate themselves from the traditional network newscast. MSNBC’s promotional slogans elucidate this goal. They included, “It’s not the same old news” and “It’s not your father’s newscast”, a play on the Oldsmobile advertising campaign of the day. Cable news catered to a relatively small audience, and over time began to attract unique viewers because of the type of news content they provided. Data from Jonathan Morris’ research in The Fox News Factor shows the migratory trend of TV news consumers. “In 1993, 60% of the American public reported watching network broadcast news on a regular basis. By April 2004, that number had been almost halved to 34%” (Morris, 2005). The data also shows that network audiences are increasingly older while Fox News and CNN audiences are increasingly polarized (Morris, 2005). Additionally, another study shows that the audience for network newscasts has steadily decreased while CNN’s audience has steadily increased over the last few decades (Bae, 2000).

“Television news powerfully influences which problems viewers regard as the nation’s most serious.” (Iyengar, & Kinder, 1987) This quote, along with the previous research, illustrates that television is a major outlet for agenda setting and that there is a great deal of media selectivity based on politically ideology. Thus, it becomes important to understand how news differs across these TV outlets as a means of understanding how fragmented our society may become. Morris explains that “Today’s television news market is more heterogeneous than ever before. Thus, the probability that audiences are getting exposed to differing political messages increases” (West, 2001). An early content analysis from 1998, showed that there was already a significant difference in the types of stories and levels of analysis found, both among cable news shows and between cable news and network news (Bae, 2000). A more recent study from 2007 demonstrated that Fox News’ reputation for having a conservative bias was well deserved. They found that its news coverage showed a consistently pro-Republican slant (Iyengar, & Hahn, 2007). Further, Morris’ analysis shows that the Fox News audience is made up of a unique composition of viewers, and that they are exposed to different coverage than the CNN and network news watchers. The results also showed that Fox News watchers were less likely than CNN watchers to follow stories critical of President Bush and were more likely than non-watchers to underestimate the number of casualties from the Iraq war. Overall, he finds that both Fox News and CNN contain a partisan bias, but perhaps more importantly, he concludes that both audiences are moving farther from the ideological center (Morris, 2005). For this reason, it seems likely that the news product coming out of these outlets will become increasingly unique, as they continue to cater to their partisan audiences. Iyengar’s study expounds on the economic motivations of news organizations to cater to their viewers’ political preferences. His evidence shows that between 2000 and 2004, Fox News increased the size of its regular audience by almost 50%, as proof of the viability of the “niche-media” model today (Iyengar, & Hahn, 2007).

Methods

This paper employs a statistical analysis to study the trend of media fragmentation along the lines of political ideology. The data originated from the Pew Research Center for the People and the Press. The specific data set under examination is their Biennial Media Consumption Survey from 2008 (Datasets, 2009). The survey asked 3,600 adults a wide range of questions about their demographics, political opinions, and most importantly, media consumption habits. I used a bivariate crosstabulation to generate a contingency table comparing the frequencies of people’s political preferences with how often they watched certain partisan news programs.

This analysis looked at several of the survey questions. The first question asks respondents, “In general, would you describe your political views as…” They were then provided with responses including, very conservative, conservative, moderate, liberal, and very liberal. This question was chosen because one of the goals of the study was to show that people seek out agreement in their news and that ideology would likely cast a wider net than party membership in revealing people’s political preferences. Not everyone who took the survey answered this question and the n value for very conservative was only 232 and very liberal was 171. Therefore, in order to have a sufficient size that would allow my results to be statistically significant, I recoded the response categories, combining very conservative with conservative and very liberal with liberal. The next set of questions asked people to rate how often they watch, listen, or visit the website of different media outlets. Respondents again had a scale of choices including, regularly, sometimes, hardly ever, or never. To emphasize the hypothesis of ideological media fragmentation, I chose two cable news networks with strong reputations for having a partisan slant–The Fox News cable channel on the right, and MSNBC on the left. These two media outlets were also selected because of the breadth of research conducted on television news, indicating that TV is the most popular source of information for Americans and showing that there is abundant political bias in TV news.

Next, I ran a crosstabulation to show that not only do people select news that agrees with their ideologies, but also that consuming this type of news is related to their opinions about partisan issues. I used the same media consumption question about Fox News as above, but here I compared the frequency with which people watch Fox News with their self-reported approval of George W. Bush, the Republican president. This survey question read, “Do you approve or disapprove of the way George W. Bush is handling his job as president?” My final statistical analysis was more general in scope, but very important in showing that people’s self-reported media habits do not always reflect the actual trends in media consumption. I ran a frequency analysis of responses to the following question, “When getting news, what’s more important to you…” The response choices were, “getting an overview of the top news of the day”, “getting news about topics of particular interest to you”, or “both”.

Results

The statistical analyses described above resulted in many interesting observations about media fragmentation and polarization. My hypothesis predicted that media consumers tend to select media outlets that agreed with their preexisting political ideology. Table 1 supports this hypothesis by showing that people who self-identify as conservative watch, read, and visit the Fox News web site at higher rates than respondents who say they are moderate or liberal. Conversely, in table 2 we see that those people who identified themselves as liberals were more likely to tune in to MSNBC than moderates or conservatives.

Tables 1 and 2 also support the idea discussed in the literature claiming the stronger a person’s partisan affiliations, the more likely they are to select media outlets that confirm their beliefs (Iyengar, & Hahn, 2007). The tables show that the results are more significant in examining the extreme ideological positions. Of people who regularly watch Fox News, 37.8% are conservative while only 14.4% are liberal. And conversely, among regular MSNBC viewers, 20.6% are liberal compared to 12.5% who are conservative.

A closer examination of respondents who identify as moderates also proves interesting. The moderate respondents diverge from the general trend of ideological selectivity. Instead, across both Fox News and MSNBC, the moderates are clustered in the “sometimes” and “never” categories. Considering the existing research about politically uninterested media consumers, this is not altogether surprising. First, it is likely that the majority of politically interested respondents (who are all of voting age) would have some affiliation, if not with a political party, than at least with a political ideology. Thus it is probable that a majority of the people who selected moderate are in fact uninterested in political news. As discussed extensively above, in today’s high choice media environment, these individuals would avoid political content by steering clear of 24-hour cable news networks like Fox News and MSNBC, which both devote a large majority of their airtime to political issues. While this can account for the approximately one third of moderate respondents who said they “never” watch Fox News or MSNBC, we still must account for another third who reported watching these outlets “sometimes”. It seems apt to assume that these people were both politically interested and truly moderate. Thus, they would occasionally access political news from a wide variety of sources with varying ideological slants. A final observation from the data in tables 1 and 2 is that the trend towards ideological selectivity seems to be stronger in table 1 where people were asked about Fox News. A possible explanation for this disparity can be found in the existing literature on television news. Morris explains that his study showes, “the Fox News audience is slightly more Republican than the CNN audience is Democrat” (Morris, 2005). My study supports and adds to this assertion by showing that the Fox News audience is more conservative than the MSNBC audience is liberal. These findings support yet another study, which compares Fox News to other new media outlets and shows that they are in fact more biased than other outlets (Iyengar, & Hahn, 2007).

The general trend in table 3 begins to support the hypothesis that selective exposure to a partisan media source will affect people’s political views. The “approve” column shows an almost perfectly linear trend, indicating that the more someone watches Fox News, the more likely they will be to approve of their Republican president. The “disapprove” column is not as significant, so I will refrain from drawing any broad conclusions from that data. There is an aberration to the trend found in the “never” category. This is a logical result, as it can be assumed that if someone has no exposure to Fox News it cannot affect their political perceptions. Of people who said they never watch Fox News, 22.5% approve of the President while 40.1% disapprove. This data can be explained partly outside the realm of media consumption. This survey was taken at a time when President Bush had very low approval ratings overall, which is likely the cause for this discrepancy.

While table 4 does not directly relate to the stated hypothesis, it elucidates an important component of this research. I have argued that media selectivity and fragmentation are predominantly negative trends. These results may suggest that most survey respondents agree. 64.1% say that getting an overview of the news is most important to them, while 26.3% say they value getting news about topics of particular interest to them. However, data from tables 1 and 2, as well as countless other studies about media selectivity, show that this is not the actual distribution, but instead, far more people seek out news based on their own interests and ideologies. Perhaps this incongruity stems from the fact that, while people do filter their news according to their interests, they understand that this is not a positive habit and therefore over-report the desire to get an overview of the news. Another, and perhaps more alarming possibility, is that many people are filtering and selecting their media intake without even realizing they are doing it. While two-thirds of respondents claim that getting an overview of the news is most important to them, there is still 26.3% or, 924 out of 3512 people, who admit to active media selectivity. This is paramount to the basis of my study, as it implies that media selectivity is already happening with at least a quarter of the population.

Discussion

My hypothesis of media selectivity based on ideological preference was generally supported by my data. Liberals did report more exposure to MSNBC than moderates or conservatives, while the opposite trend was found with respect to Fox News. These findings contribute to the existing literature in several ways. First, the results of the statistical analysis support the general theory of media selectivity. Clearly, different citizens are utilizing different media outlets, with over a quarter of respondents admitting that getting news about topics of particular interest to them is most important. Further, these findings support the conclusions of many polarization studies by showing that peoples’ partisan affiliations are predictors of which news outlets they will use.

In addition to supporting many existing studies, I set up my research in a manner that allowed me to produce some novel results. Some studies have used Fox News as their sample of a conservative news outlet, yet most compare it with CNN, NPR, or network news. I chose to compare Fox News with MSNBC, as in today’s television market it is becoming clear that MSNBC has established itself as the left leaning cable news channel, with Fox News on the right, and CNN in the center. Evidence for the Fox News-MSNBC dichotomy is found in a recent interview with White House adviser, Valerie Jarrett. When asked, “do you think Fox News is biased?” Jarrett said, “Well of course they’re biased, of course they are, ” to which the journalist immediately followed up with, “Do you also think that MSNBC is biased?” While CNN may also have a liberal bias, it is very important to understand the reputation of the news outlet; as this research is trying to show that citizens are likely to gravitate towards media they think will agree with their opinions. I also chose to analyze the frequency of media exposure compared with political ideological affiliations rather than political party membership, as many previous studies have done in order to cast the widest possible net of political opinion.

There are many avenues of future research on this topic that have yet to be explored. While there are a few studies that attempt to establish a causal relationship between selective media exposure and polarization, more research in this area would prove very useful. Perhaps a longer study that focuses on changes in people’s political attitudes as they are consuming different types of media over time would begin to support this relationship. Also, the body of work related to high choice media environments can definitely be expanded with studies focusing on the effects of Internet media. Just as many studies have shown greater levels of selectivity with the advent of “old” new media, such as talk radio and cable television, research into the proliferation of “new” new media like websites, blogs, and social networking sites will likely reveal similar findings. Considering that many of these media outlets advertise their break from objectivity as a positive development, it can be assumed that they will include more partisan news, possibly expanding the range of public opinion in the direction of the extremes. Thus, further investigation into the role of today’s new media in agenda setting and political polarization will prove very important. Another interesting avenue for continued research on the subject of media fragmentation would be to look at this issue from a uses and gratifications perspective. This type of study would employ qualitative interviews to study people’s motivations for using specific media outlets for political news.

References

Bae, H. (2000). Product differentiation in national tv newscasts: a comparison of the cable all-news networks and the broadcast networks. Journal of Broadcasting & Electronic Media , 44, 62-77.

Baum, M. . (2003). Soft news goes to war: public opinion and american foreign policy in the new media age. Princeton, NJ: Princeton University Press.

Baum, M.A., & Groeling, T. . (2008). New media and the polarization of american political discourse . Political Communication, 25(4), 345 – 365.

Bernhardt, D. , Krasa, S., & Polborn, M. (2008). Political polarization and the electoral effects of media bias. Journal of Public Economics , 92, 1092-1104.