Junk Political Science

Kevin Drum notes a survey (PDF) that says, among other things :

The extent of Americans’ misperceptions vary significantly depending on their source of news. Those who receive most of their news from Fox News are more likely than average to have misperceptions. Those who receive most of their news from NPR or PBS are less likely to have misperceptions. These variations cannot simply be explained as a result of differences in the demographic characteristics of each audience, because these variations can also be found when comparing the demographic subgroups of each audience.

The problem is they have found a correlation, and tried to make it causal link. They present no evidence as to causation. It’ is opinion polling dressed up as a scientific study.

I’m not disputing their numbers, but instead the inferences they are trying to make. They simply have not shown a statistical approach worthy of the kind of absolute causation claims that they are trying to make. An economist trying to make similar conclusions would be laughed out of the room and flunked on their thesis. I can think of all manner of criteria that they did not control for but a look at one of the comments highlights the same kind of junk science that would be possible if the misperception set were altered.

If the survey was on the following three misconceptions:

  1. Halliburton and other administration business interests was behind the decision to go to war in Iraq.
  2. The U.S. government knew the 9/11 attacks were going to happen, and did not stop them so they could further their global war strategy.
  3. The UN Weapons Inspection Program completely disarmed Iraq.

The study results would be completely reversed. Those who receive most of their news from NPR or PBS are more likely to have misperceptions. Those who receive most of their news from Fox News are less likely than average to have misperceptions.

BloggerCon (Job) - Day 1 Report
Hans Blix Is A Fucktard