Election 2016           Social Desirability Bias / Overview:  Is the Media really Biased?

In order for people to feel that an answer in a survey or poll is not socially acceptable, there needs to be a transference mechanism from which they ‘learn’ the standard for socially acceptable responses.

For national political campaigns, the media is the main mechanism by which information is transferred to the general public containing cues as to which candidates and positions should be considered socially acceptable.  Opinion leaders and social media play very large roles, but it is the media that sets the pace and tone of commentary and therefore continue to be at the center of this mechanism.

For longer term topics, such as concerning personal health and exercise, media will play a smaller role.  Respondents could, for instance, downplay the number of alcoholic drinks they have per week or increase the frequency of exercising.  But they will have likely gained insights into what answers are socially acceptable over many years of doctor’s visits, conversations with family, and even guidelines presented in government reports.  For longer term topics, media as a transference mechanism is not as important.

National elections are different in that they have a specific deadline and there is little room for subtleties.  The media with its ability to reach and influence most of the country fairly easily plays the central role in creating content and coverage for the election which ends up having a large impact on how people view the topics and candidates.  Opinion leaders, even famous and highly influential ones, end up relying on media coverage to get their messages out.  Social media can act as a counter weight to media in general, but it acts more effectively as a confirmation tool as new content created by social media is more easily dismissed due to less professional standards.  Cases of a non-journalist posting a video or commentary on something like twitter becoming a major news story exist, but they are still rare.

The main point here is that media, as the core transference mechanism, is very important in getting the word out as to what should be considered socially acceptable or not.

So, what happens if the media is overly biased and are they in 2016?

The short answers are that when the media is overly biased polls will superficially discount the socially unacceptable candidate and that, yes, the media is biased in 2016.

Note that biased does not mean that they are colluding or that there is a grand scheme of hidden communications between media outlets which attempts to control content.  In other words, there is very little objective evidence to show conspiracy theory type collusion but there is plenty of data showing how the media would be biased.  This should be seen as more or less a ‘normal’ outgrowth of longer term secular trends – which perhaps makes it easier to swallow but not less bitter.

Summarizing some of the main factors pointing towards media bias:

1. Few (and Fewer) Republican Journalists
2. Many more Liberal Media Outlets
3. Increasing and Record-Setting Bitterness between Parties
4. Echo Chamber Society

We will review these points individually, but know that the combination of them amount to greatly increased media bias on both sides.

One of the more cited reports on long-term trends in journalism and media is The American Journalist in the Digital Age.  The report is based on surveys conducted about once a decade starting in 1971, with the last survey conducted in 2013.

A main finding of this report is that the percentage of journalists identifying as Republicans in 2013 amounted to only 7.1% of the total.  This figure has declined almost every decade since its high of 25.7% in 1971.  Those identifying as Democrat amounted to 28.1% in 2013, or approximately four times as many as Republican.

We would expect that with such a low level of Republicans that many media outlets would have a bias against them.  However, this figure, though already very low, likely underestimates the amount of bias.  With the increase of conservative media, such as Fox News, many journalists identifying as Republican have likely concentrated there.  In other words, the situation has changed significantly since the survey in 1971 when the mainstream news was fairly neutral.  At that time not only did you have a more equal distribution of those identifying as Democrat (35.5%) and Republican (25.7%) but they were ‘forced’ to live together in the same news rooms making coverage less biased.  With the advent of Fox News, and other partisan media, many newsrooms are left without a counter balancing voice.

How many of the 7.1% of journalists identifying as Republican are employed or associated with Fox News?  Half seems to be a fairly low estimate given the size of the organization, but for this exercise let’s assume 3.5%.  Then take out maybe a percent each for conservative-focused and neutral-focused journalism and we would be left with maybe 2% to 3% to spread out among all other media outlets.  Now we can see just how open to bias these newsrooms are.

This is not to say that there is collusion going on between media outlets to help the Democratic Party or candidates.  But, let’s face it, if only a few percent of a company’s staff is Republican, those viewpoints will not get equal time or coverage.

This topic of the percent of journalists identifying as Democrat and Republican is very much related to the listening, reading, watching audience of each news source.  According to PEW Research Center, the bulk of the major news agencies attract liberal leaning individuals.  Almost all of the main TV stations (CNN, ABC, CBS, NBC, MSNBC, and Bloomberg) and traditional newspapers / magazines (The New York Times, Washington Post, USA Today, Guardian, Huffington Post, The Economist, and New Yorker) attract primarily liberals.  These agencies are currently seen as primarily liberal institutions.  Though specific data does not exist, we can more or less assume given the very low percent of journalists identifying as Republicans, that their staff has precious few conservatives.

More or less in the middle of the spectrum is the Wall Street Journal.  We can assume that their staff is likely fairly balanced.

On the conservative side, Fox News dominates as the largest agency.  Then smaller agencies show up including Breitbart and Drudge Report.  But these agencies are actually fairly limited, with Drudge being primarily a news consolidator.  Others listed include radio shows (Rush Limbaugh) and TV shows that appear on Fox News, which makes it seem like double counting.  In short, there are very few news agencies that attract conservatives – and the portion of media outlets that are seen as conservative are extremely few.

So, these data points more or less answer the question of media bias.  They show that there are few conservative journalists left outside of conservative-centric agencies.  This has liberated the most traditional elements of the media to operate without checks or balances causing an assumed liberal bias as those reporting do in fact largely favor the more liberal party.  Really, after looking at these numbers it should come as no surprise that an underlying liberal bias for almost all of the media is a real possibility.

But, this still does not explain why it seems to have worsened recently.  Certainly, the Republicans have been mentioning a liberal media and its bias much more so in the last decade.  What might have triggered, other than the above mentioned factors, an increase in media bias?

According to figures taken from PEW Research, the country is more divided and more negative on the opposing party than at any time since this type of data has been collected.  For instance, the percent of Republicans who view the Democratic Party as ‘Very Unfavorable’ has increased from 21% in 1994 to a high of 58% in 2016.  Likewise, the percent of Democrats who view Republicans as ‘Very Unfavorable’ has grown from 17% to 55% over the same time period.  In short, over half of each party’s members view the opposing party at the worst rating possible!  And, assuming we can use PEW’s other data on negative feelings towards the other party to see how such trends might have appeared further back, it looks like the opinions regarding the other party are, in 2016, at their worst since 1964 or when data collection for these datasets began.

Assuming that journalists are not immune to such larger secular trends would result in their opinions, on average, of the other party being equally as bad.

Another way to measure how the nation has become increasingly polarized is to look at presidential approval ratings.  By taking the average approval rating of the president from members of the in-power party and subtracting the same from the out-of-power party, we can infer a level of polarization.  For instance, if a Republican is in power, then we can assume that the president’s approval rating from other Republicans would be higher than from Democrats.  But, if the difference is too high, it means the president is very unpopular with the opposing party which is likely produced by polarization of the electorate or by very one-sided governing of the president (or both).

Unfortunately, this net approval rating has been increasing over time, showing that the country is becoming more polarized.

Chart 1:  Net Approval Rating of US Presidents, as calculated by subtracting Approval Rating of out-of-power party members from Approval Rating of in-power party members, higher numbers indicate a more polarized country

Source: PEW Research

Judging from this net rating chart, US presidents during the 1950s, 1960s, and 1970s were viewed more as individuals and less as representatives of their parties.  They were more popular within their own parties, but they could still attract fairly positive reviews from those outside their party.  From the 1980s onward, the polarization has mostly increased.  During this period, increasingly an approval or disapproval rating of the president has much more to do with your party affiliation.  Obama has hit a new record for this metric.  Whether this has something to do with him as a president or the state of the nation is another question.  Right now, it is clear that in relative terms, more Democrats like him and more Republicans dislike him than any other president.

A trend of wanting to represent your own party might actually be on the rise given statements from candidates.  There was of course Romney’s famous 47% quote which made him sound like he had little interest in representing almost half of the nation.  Then, there is Hillary Clinton’s more recent ‘deplorables’ quote that puts this figure more around 20% – 25%.  Regardless, the message appears to be that candidates want to be the president of the country for their party and are much less concerned about what those of the other party might think of them.  This creates a very pointed relationship in which treating the other side with distain is often welcomed and encouraged by party supporters.

This attitude most likely has impacted journalists and journalism as well.  Obama’s approval rating within Democrats has averaged 81% while from Republicans it has averaged 14%, according to PEW Research.  Assume a journalist who identifies as a Democrat is working at a liberal learning media agency, don’t you think such strong tendencies will impact the journalist’s work?  Knowing that there is such strong support within the readership or viewership for the president would most certainly have a direct impact on coverage.  The opposite is true for Republicans.

This situation has created a somewhat fixed echo chamber in which those with political leanings will consistently go to different news sources than those from the other party.  By consistently going to sources that tend to support your view of the world you place yourself within an echo chamber, simply reinforcing previous inclinations.

Table 1:  ‘Percent who got news about politics and government in the previous week from’ the following sources, Comparing ‘Consistently Conservative’ to ‘Consistently Liberal’ responses

Source: PEW Research

As you can see there is only one source that overlaps which is Local TV, which is likely biased but not differentiated here.  In other words, many people on both sides of the political spectrum get their news from Local TV, but the reporting must vary considerably between primarily Democratic and Republican regions.  In other words, though tagged here as Local TV, it is likely highly biased reflecting the local region.  For instance, the local San Francisco TV news show likely has a much different underlying bias than that from Tulsa.

Other than Local TV, conservatives and liberals get their news from entirely different sources.  This appears to have created an almost perfect echo chamber where those already with heavy leanings towards one end of the political spectrum repeatedly return to the same biased news sources.

This topic can be fairly complex and somewhat emotional for many people so let’s review with some dry Q&A:

• Is the media biased? Yes, there seems to be evidence of bias on both ends of the spectrum.
• Is the liberal media more biased? Tough to say using this data, but certainly there are many more major liberal leaning news agencies than conservative ones, so you could generalize saying that the ‘media’ has a liberal bias as the liberal media covers much more of the universe.  Certainly a conservative will have a higher probability of seeing or reading liberal news due to its wider availability, which likely goes a long way in explaining the ‘liberal media bias’ complaint from Republicans.
• Is the media dominated by liberals? It certainly seems to be the case given the extremely low percent of journalists who identify as Republicans.  This trend also seems to be on a deteriorating longer term track.
• Do journalists take sides? Given the state of secular trends it seems like journalists are following the rest of society in seeing the opposite party, its members, and its policies as bad and even dangerous and scary.  As journalists are people, they likely follow secular trends to a similar extent and likely allow their work to be impacted.
• Do the liberal media mostly support Clinton and the Democrats? Given historical trends in who journalists support for president (as described in another post) and how they identify politically, it is safe to say that the liberal media does support Clinton.  As the liberal media vastly outnumbers the conservative media, it would appear that the ‘media’ on average is consistently protecting and rooting for Clinton over Trump.
• Do the liberal media collude to support Clinton and the Democrats? You would need a smoking gun to prove this idea.  A perceived collusion of support for Clinton could be explained by the larger secular trends.  Just because CNN, MSNBC and The New York Times run similar negative stories on Trump and chose to ignore a negative Clinton story does not mean there was a secret phone call among the top executives of these companies to hide something from the public.  Similar talking points could provide more of an indication of collusion or that could be explained by journalists and commentators following how the same story is being covered on other similar networks and copying the approach down to the wording.  Certainly it is possible for collusion to be going on, but not likely.