From a e-friend pursuing his PhD in neuroscience:
Kids at UCLA are liberal.
I’m looking at the original paper and the liberal-conservative scale.
They say 43 subjects but I caught only 37 data points. Ok, no big deal.
On a scale of Liberal 5 to Conservative 5, 26 people identified themselves as Liberals, 4 are at 0 (moderates?), and 7 are on the conservative side.
The most common score was a Liberal 3. I know how these things work, you don’t want to appear like a 9-11 truther, but you still want to impress the slightly older hip grad student. So you answer negative to the questions about proletariat uprising, but affirmative to questions about universal health care.
Thing is when comparing groups is that you generally want to have equal numbers in each group. It’s not easy to do, but you at least have to try. 26 to 7 is over three times as many liberal subjects. Further, 7 subjects is a pretty small sample in general. They could only do a correlation because the sample is just to small to run inferential stats.
Here’s the kicker, basically two moderates (score of 0), and two conservatives dragged down the average. Basically, you find two guys with thick skulls (the measure is the electrical potential on the surface of the head), lump them in with the group in which you want to show an effect, and voila, front page news. Also, with these kind of numbers, you can’t even control for gender or social economic status, etc.
I haven’t even gotten into supplemental information, where all the real bugs reside.
No Conservatives stronger than a score of Conservative 3. Not enough conservatives on campus!
I think the researchers and journalists hit their goal, which was to get us to talk about it.