I was very disappointed after reading the article because I expected a well-reasoned psychological argument as to why Amodio et al. was incorrect. Instead, much of the article was hand-waving and weak arguments against standard methods in cognitive psych. I wanted to do a quick point-by-point of the article with my own thoughts on it at the end.

1. Habitual Ways of Thinking. Fifteen minutes IS a habit in terms of how you are completing the task at hand. Studies have been done studying single neurons which show that some actions can become routine in as few as 5 trials (Pasupathy and Miller, 2001). Further, tapping a keyboard is a 'way of thinking' if a key response is a critical part of the decision making process.

2. Responsiveness to Information. The time constraints on the response are there for a reason. If the researchers gave the subjects too much time to respond then it would be impossible to measure levels of cognitive control at the onset of a stimulus. By making the subjects respond very quickly you are able to measure how they deal with conflicting information.

3. Complexity and Ambiguity. My sense is that Amodio talks about complexity and ambiguity not as words describing the results of the study, but about the information processing ability of these people in general. While there is some evidence to support this idea, I think in the context of the paper (political affiliation) this was a bit too far. This was the point that got them the press though, which is probably why is was included.

4. Maladaptiveness. The Botvinick et al. (2001) was a landmark piece on cognitive control and conflict monitoring. The idea that its definition is at odds with the one from the Amodio work is incorrect. They are both describing a system that keeps an eye out for any bit of information that would require modification of a prepotent response. Also, extra conflict monitoring sensitivity has been shown to be adaptive across a wide range of studies. This has been demonstrated developmentally (Casey et al., 1997) and in clinical disorders like OCD (Roth et al, 2007) to name two groups.

So, don't get me wrong on the above points. I am not a big fan of the study. I believe that their population sample was not truly random and that to say anything robust about liberals/conservatives you would need to sample from urban and rural areas across the whole country. A quick power analysis also indicates that they should probably have 50 people per group for a modest effect size with a standard t-test - though ERP studies tend to have their own funny methods. I have other issues with the study as well, but I think you get my point.

The Amodio paper isn't a great example of scientific progress. There may very well be differences between liberals and conservatives, but studies with more powerful designs are needed. Finally, if studies like this are to be assailed they should be criticized on their weak points, not on their premise.