Monday, 30 June 2003

Toward a theory of perceived media bias

Why can both liberals think the media is biased toward conservatives and conservatives think the media is biased against them? (Not to mention, why do libertarians like me think the media is biased toward government action?)

Here’s a rough draft of a theory. John Zaller’s The Nature and Origins of Mass Opinion makes a case for what he calls the “receive-accept-sample” (RAS) model of opinionation. In simple terms, people receive information (which may or may not be biased; it can be purely factual, or it can be partisan or distorted in some way) from a variety of sources, such as the mass media, friends, family, colleagues, etc. This information is then either accepted or rejected through a process of perceptual screening. Finally, when people are called upon to give their opinion on a particular issue, they sample from the information they have available pertaining to that issue (in Zaller’s terms, the relevant considerations).

Now, if we focus on the accept part of the RAS model, we find something interesting. People who are more politically sophisticated generally only accept the information that is consistent with their preexisting beliefs, while less sophisticated people have less developed screening mechanisms (since they care less about consistency or balance). People who complain about bias are, largely, the politically sophisticated (those who look down on Rush Limbaugh or Bill O’Reilly listeners may find this hard to accept, but compared to the average American I suspect those listeners are more politically sophisticated; they may have political beliefs that seem reactionary, but they generally know a lot more about politics and how it works than the average citizen). This probably isn’t a coincidence.

My theory is that citizens are more conscious of rejection than they are of acceptance. People who encounter information consistent with their beliefs will simply accept it and move on, while people who encounter information that is inconsistent will be more aware of that inconsistency. Thus, people who encounter more inconsistency when they encounter information from a particular source will find it to be more biased than those who encounter less inconsistency. And, they will perceive that bias as being in the direction opposite to their preexisting belief system—because, relative to them, the bias is in that direction.

Does that mean (if this theory is correct—I suspect it is, but then again, that’s because it’s my theory) that bias doesn’t exist? No. But it does mean that liberals like Eric Alterman and conservatives like Ann Coulter will only perceive bias in opposition to their own belief system.

Conceivably, one could pick a “belief system axis” and measure the bias of various information sources relative to the origin of that axis. To test this, you’d need to have a number of raters from various positions within that belief system axis (which you could locate using standard measures of ideological belief) and then have them give some measure of the bias of the information sources. Then you could produce an “objective” map of the bias of the sources using multidimensional scaling (which, hopefully, would be fairly easy to interpret).

Sunday, 6 July 2003

Kevin Drum on moderation

CalPundit tries to figure out what it means to be moderate. Money quote:

First, there’s a difference between policy moderation and rhetorical moderation. John Kerry, for example, is probably about as liberal as Howard Dean if you look at his actual policy positions, but Dean uses more fiery rhetoric. Likewise, aside from a regrettable weakness for sarcasm, my writing tends to be pretty sober compared to someone like Atrios. But on actual political positions, we’re fairly close.

And that, in a nutshell, is why I take Kevin seriously but would rather get hit by a bus than visit Atrios again in my lifetime, and hence why Kevin actually has a decent shot at persuading me that he’s right on the issues.

James at OTB has more, including a money quote of his own:

Indeed, the mere fact that we spend a lot of time thinking, let alone writing, about politics and have developed somewhat coherent positions almost by definition puts us into the extremes.

I’d probably elide the “almost.” Having a coherent personal belief system puts you in the tails of the bell curve, at least relative to the public at large. I think moderation is more a function of whether you allow the belief system to dictate how you feel about people who don’t share your beliefs, and not so much what the content of your belief system is. Which, incidentally, gets back to what I was saying about political sophistication and perceived media bias.

Sunday, 24 August 2003

The Carolene Media

Walter Cronkite’s first column has been online for two weeks, but apparently it took this Eric Burns piece for for it to garner much publicity. The “bombshell”: Cronkite concedes most journalists are liberals. Why?

I believe that most of us reporters are liberal, but not because we consciously have chosen that particular color in the political spectrum. More likely it is because most of us served our journalistic apprenticeships as reporters covering the seamier side of our cities – the crimes, the tenement fires, the homeless and the hungry, the underclothed and undereducated.

We reached our intellectual adulthood with daily close-ups of the inequality in a nation that was founded on the commitment to equality for all. So we are inclined to side with the powerless rather than the powerful.

Perhaps we should just call it the Footnote Four justification for media bias.

Speaking of media bias, if you want to see some testable, positivist political theory, try this blog post on for size. Ah, if only I had a research grant…

Wednesday, 8 October 2003

Evidence for a theory of perceived media bias

I posted my sketch of a theory of perceived media bias a few months ago; now Gallup has done me a favor and produced a poll that suggests I may be onto something. Take a look-see at the results:








2003 Sep 8-10




2002 Sep 5-8




2001 Sep 7-10





2003 Sep 8-10




2002 Sep 5-8




2001 Sep 7-10





2003 Sep 8-10




2002 Sep 5-8




2001 Sep 7-10




Now, unfortunately, there’s nothing to show the causal mechanism here (i.e. why conservatives and liberals perceive the media’s biases differently). But it’s an interesting look at the question, nonetheless.

Link via Andrew Sullivan (although I think I saw it cited earlier somewhere else).