Tuesday, October 21, 2014

What makes somebody liberal?

“Reality has a well-known liberal bias.”
 -Stephen Colbert

In this post, I make an argument that is, perhaps, the most partisan argument imaginable.

Ready?  Here it is:

Conservatives are biased.  But liberals are not.

Before you try pulling the beam out of my eye, however, hear me out: The Pew Research Center just released a study on political polarization and media habits.  The study includes some nifty tables detailing the relationship between audience members’ political ideology and their trust in various news sources.  After reading a Wonkblog post about the study, I wondered whether there was any relationship between trust in a news source and the political leanings of its audience.  Here’s what I found:[1]

There is a strong, positive correlation (r = 0.7379) between the general public’s trust in a news source and the percentage of liberals who get news from the source.  But there is almost no correlation (r = -0.0210) between the general public’s trust and the percentage of conservatives who get news from the source.  That means that liberals are much more likely than conservatives to get news from sources that are trusted by the general public.

Now, Pew’s sample skewed a little liberal.  It’s possible that the sample is more likely to trust news sources favored by liberals because the sample itself is slightly liberal.  To check for this, I ran the numbers again using the trust of politically moderate audience members rather than the general public.  Alas, the relationship persisted.  There is a strong correlation (r = 0.6669) between moderate consumers’ trust in a news source and the percentage of liberals who get news from the source.  There is a weak correlation (r = 0.1372) between moderates’ trust and the percentage of conservatives who get news from the source.  So liberals are also more likely than conservatives to get news from sources that are trusted by moderates.

One interpretation of this finding is that liberals can draw from a wider range of sources to reinforce their views because the media has a general liberal bias.  According to this interpretation, conservatives trying to avoid liberal bias have very few options, and must resort to a small selection of conservative news sources that are unfairly decried precisely because they are conservative.

However, I don’t think the data supports that interpretation.  First, liberals use balanced sources more than conservatives.  If moderates’ trust[2] in a news source is a good metric of its ideological balance, then the most balanced sources are CNN and the major networks.  Liberals are more likely to watch all these sources than conservatives.[3]  Additionally, liberals are less likely to use distrusted sources.  The five sources most used by liberals (NPR, CNN, MSNBC, NBC, and PBS) are all trusted by moderates.  In contrast, three of the five sources most used by conservatives (Sean Hannity, Rush Limbaugh, and Glenn Beck) are distrusted by moderates.[4]  In fact, Rush Limbaugh is the least trusted source among both the general public and moderates.  Finally, liberals draw from a greater range of sources than conservatives.

Because liberals draw from both more trusted and more numerous sources than conservatives, liberals get a fuller and more accurate understanding of the news.  I (not so) humbly submit that this better understanding is not only why liberals are less biased, it’s why we’re liberal.



[1] Methods:
I did all the calculations using Excel with the tables provided by Pew.

General public:
I calculated a net overall trust index by subtracting the percentage of overall distrust of each news source from the percentage of overall trust of each news source.  I then used the CORREL function to calculate the correlation coefficients of the net overall trust index with (1) the percentage of consistently liberal respondents who got news from each source, and (2) the percentage of consistently conservative respondents who got news from each source.

Moderates:
I calculated the net mixed trust index by subtracting the percentage of respondents with “mixed” political views who distrusted each source from the percentage of respondents with “mixed” views who trusted each source.  I then used the CORREL function to calculate the correlation coefficient of the net mixed trust index with (1) the percentage of consistently liberal respondents who got news from each source, and (2) the percentage of consistently conservative respondents who got news from each source.

[2] Unless otherwise noted, “trust” refers to the trust index described above in note 1. 

[3] Liberals are more than twice as likely to watch CNN as conservatives (52% to 20%).

[4] The most popular source among conservatives, Fox News, is trusted by moderates.  Only 10% of moderates had heard of the fifth most popular source among conservatives, TheBlaze.  2% of moderates said they trusted TheBlaze, and 1% said they distrusted it.  Because TheBlaze is owned and founded by Glenn Beck and the balance of moderates distrust the Glenn Beck Program, moderates would presumably distrust TheBlaze if they were familiar with it.

2 comments:

  1. You cannot assume people with moderate political ideology are less biased than either liberals or conservatives. Having a moderate bias does not refer to the degree of bias but to where the bias points to.

    Moderates' trust of a news source is strongly correlated to the news source's lack of a face. In other words, moderates trust networks more than talk-show hosts. That is more likely why moderates trust Fox News but not Sean Hannity; they do not trust Glen Beck but twice the number of moderates trust The Blaze compared to those that do not trust it.

    While i disagree with yoir initial statement, (really we all have some kind of bias) you make some interesting points that have got me thinking. My news sources are NPR, BBC News and a bunch of poscasts (many associated with NPR.)

    ReplyDelete
    Replies
    1. It's true that moderates aren't necessarily less biased than partisans. That's the reason I refer to the sources moderates favor as "balanced" rather than "neutral." However, the utility of a metric doesn't stem from how well it "outruns the bear" but by how well it performs against other metrics. I decided that the trust of moderates was the best available measure of "balance" because Pew's sample skewed liberal, and that skewness would affect the reported "overall" trust in source.

      http://www.slate.com/articles/news_and_politics/jurisprudence/2011/10/breyer_and_scalia_unintentionally_make_the_case_for_cameras_in_t.html

      While there is a correlation between a source's institutional (rather than personal) character and whether it's trusted by moderates, I would speculate that this correlation stems from the partisan nature of personality-driven sources, rather than the fact that they're personality-driven.

      Delete