Article Summary
Introduction
With the rise of cable television, new outlets like Fox News and MSNBC that are explicitly partisan have become very popular. The programming on Fox is primarily designed for Republicans, and the programming on MSNBC is primarily designed for Democrats. These channels are not intended to serve a wide audience, and so many have argued that they are less focused on presenting accurate information about the world and more focused on creating programming that aligns with party narratives, criticizes the other side, and makes audiences feel good. This article asks, “What happens to people who watch this kind of partisan news?” A large body of research in social psychology shows that most people think in terms of social groups. In the context of U.S. politics, the most relevant social groups are Democrats and Republicans. When people think in terms of groups, they identify themselves based on group membership. Republicans start to feel like being a Republican is central to who they are, and the same goes for Democrats. The author of this article suggests that, because partisan media spends so much time talking about parties, it activates these kinds of partisan social identities. When these identities are activated, the author argues that people will become more positive toward their own party and more negative toward the other party—that is, affectively polarized. And this is especially true because so much partisan news coverage talks about how one party is good and the other party is bad. Furthermore, the author argues that similar processes make exposure to partisan media trust the other party less and be less supportive of bipartisan compromise—why would someone trust the other party and want to work together with them if that person thinks they are bad people with bad ideas?
This argument makes sense when we think about Democrats exposed to Democrat-friendly media and Republicans exposed to Republican-friendly media. What happens when someone sees content from the other side that criticizes that person’s own party? The author argues that, because messages are now coming from an out-party source who most will view as different and not credible, they will not influence people’s feelings toward their own parties. However, for people who believe that the out-party source is credible, exposure to this kind of programming will make people feel less positive toward their own party.
Analytical Approach
What were the design choices? For all quantitative projects generate either a DAG (non-experimental designs) or a full table of all randomized conditions. If multiple methods were used, please complete this section for all forms of analysis used.
The author tested these expectations about the effects of partisan media exposure with a pair of survey experiments. In Experiment 1, participants were randomly assigned to watch one of several video clips. In the first clip, they saw a story unrelated to politics (functioning as a control condition). In the second clip, they sawa story about politics from a like-minded source that shares their party (i.e., Fox for Republicans and MSNBC for Democrats). In the third clip, they saw a story about politics from a cross-cutting source opposite to their party (i.e., Fox for Democrats and MSNBC for Republicans). These clips were sourced from real broadcasts from shows on Fox News and MSNBC.
| Table of Conditions for Experiment 2 | ||
|---|---|---|
| Republicans describe bill as bipartisan + 0-100 Democrat votes | Republicans describe bill as important + 0-100 Democrat votes | Republicans describe bill as partisan+0 Democrat votes |
After watching the clips, participants in Experiment 1 were asked feeling thermometer questions capturing positivity and negativity toward each party. They were also asked how much they trust the other party in Congress to do what’s right for the country, and how much they support their party working with the other party in Congress rather than focusing on advancing only their own party’s policies.
In Experiment 2, participants were randomly shown either an apolitical control story or a treatment story from Fox News criticizing Obama’s foreign policy—there was no equivalent clip from MSNBC, so the author used only a Republican-favored story in this experiment. After viewing the clip, participants were asked to evaluate the credibility of several different media outlets including Fox News, then answer the same questions as in Experiment 1 (i.e., feeling thermometers, trust, and support for bipartisanship).
| Table of Conditions for Experiment 2 | ||
|---|---|---|
| Republicans describe bill as bipartisan + 0-100 Democrat votes | Republicans describe bill as important + 0-100 Democrat votes | Republicans describe bill as partisan+0 Democrat votes |
Main Findings
Exposure to like-minded partisan media makes people more affectively polarized and less supportive of bipartisanship. Its effects on trust and evaluation of political leaders are less clear. There is also some evidence that cross-cutting exposure to out-party sources can make people see their own party more negatively, but the relevant experiment only tested this possibility for Democrats.
More specifically, in both experiments, the author found that exposure to like-minded partisan media made people more negative toward the other party. However, it did not make people more positive toward their own party, perhaps because people tend to be more attentive to negative information (i.e., criticism of the other party) than to positive information (i.e., praise of their own party). Like-minded exposure also substantially reduced support for bipartisanship. In Experiment 1, exposure to partisan news from the other party had no effect on how people felt about their own party. However, in Experiment 2, the author found that this was only true for those who felt that the out-party source was not credible; for people who felt that the cross-cutting source was credible, exposure made them feel more negative toward their own party. However, as noted in the analytical approach section, this statistical interaction was tested only for Democrats, not for Republicans.
Implications
This article shows that partisan media like Fox News and MSNBC play an important role in shaping how people feel about the political parties and how much they support bipartisan action in Congress. Importantly, the author notes that these findings are likely not driven by incivility. Name-calling and insults can be politically influential, but it is partisan slant, not incivility, that drives the effects in this article.
Although the audiences for partisan media are a relatively small segment of the population, the effects of exposure are particularly important compared to many other kinds of media exposure. The kinds of people who watch partisan media are active partisans. They are opinionated, ideologically extreme, and especially likely to participate in politics. These are the voices heard by legislators—so their attitudes (and factors like partisan media that influence them) are important.
In the supplementary materials, the author also showed that the effects of partisan media are unlikely to be explained by the experiments showing people clips they would never have watched in the real world. In Experiment 1, participants were asked about their likelihood of watching partisan media in their normal lives. Even when these preferences were included in the models as moderators, treatment effects were identical. That is, it doesn’t seem to be the case that these effects were only obtained because people saw things they never would’ve otherwise.
Questions left unanswered
Although the author suggested that the lack of evidence for effects of like-minded media on feelings toward people’s own parties may have been driven by inattention to positive information, this is ultimately only speculation. Future research should be conducted on how partisan media (and information that favors people’s parties generally) influences the way they feel about those parties.
The author also notes that the people who see the out-party source as credible are not well understood. Who watches the other party’s programming? Who trusts it? Are these people systematically different from the rest of their parties?
The author highlights that these results were demonstrated from a single treatment in which people watched a clip once. It remains unclear whether these effects persist over time or wash out in hours. It is also unclear whether repeated exposure to partisan media has different or stronger effects than a single viewing instance.
Methods and Analysis
Was the study and its analyses pre-registered?: Study was conducted before 2015
Did the study rely on proxy variables to measure polarization?: No
Were standard p-value thresholds used (p<.05 or 95% Confidence Intervals that don’t overlap zero)?: No
- Largest p-value presented as significant: 0.1
Were correlational results interpreted with causal language?: No
Limitations / Weaknesses
Because the clips used in Experiment 1 were real, the Fox clips differed from the MSNBC clips in content, tone, and many other factors. It would be worthwhile replicating these analyses with either more carefully controlled clips or stimulus sampling from a broader pool of clips to ensure that effects were not driven by idiosyncratic stimuli. The author conducted four experiments, but only two are reported in detail in the main manuscript. The other two are detailed in the appendix.
As noted above regarding inference metrics, the author appears to interpret several marginal or entirely non-significant findings in Table 3 and 4 as practically meaningful. Binary significance determinations can be limited, but the authors’ logic for interpreting these findings as significant should have been made clear.
It is unclear why the author did not test for an interaction between out-party source credibility and cross-cutting exposure in Table 3, which details effects of the Fox News clip criticizing Obama’s foreign policy on Democrats’ and Republicans’ evaluations of Obama. In principle, these dynamics should have followed those in Table 2, which used data from the same experiment but included the interactions with source credibility. Similarly, the author argues that their findings in Table A1 (wherein Democrats watch an MSNBC clip criticizing Republicans’ foreign policy) indicate that the effects in Experiment 2 were not driven by idiosyncrasies of the clip or its Republican leanings. However, Table A1 again does not include the interaction with source credibility necessary to evaluate the potential conditional effects of cross-cutting exposure. This is not a major limitation, but including these interactions would have provided much more support for the author’s argument about cross-cutting exposure and source credibility.
Open Data & Analyses
Does the article make the replication data publicly available?: No
Does the article make the replication analysis scripts publicly available?: No
Article Citation
Levendusky, M. S. (2013). Partisan media exposure and attitudes toward the opposition. Political Communication, 30(4), 565–581. https://doi.org/10.1080/10584609.2012.737435
Bibtex
1
2
3
4
5
6
7
8
9
10
11
12
13
14
@article{doi:10.1080/10584609.2012.737435,
author = {Matthew Levendusky},
title = {Partisan Media Exposure and Attitudes Toward the Opposition},
journal = {Political Communication},
volume = {30},
number = {4},
pages = {565-581},
year = {2013},
publisher = {Routledge},
doi = {10.1080/10584609.2012.737435},
URL = {https://doi.org/10.1080/10584609.2012.737435},
eprint = {https://doi.org/10.1080/10584609.2012.737435}
}