Thursday, March 23, 2006

How to spot 'junk' science

If the sample pool consists 100% of UC Berkley faculty and staff children (a University with a 91% liberal rating), and is itself, performed by a member of that same university, it's probably not a useful study to base any conclusion concerning political ideologies on.

If the strongest correlation discovered, using this highly biased source group and unclear testing procedures, is .27 (indicating a roughly 7% affect on outcome), it's probably not going to be used as the basis for any Nobel Laureates work.

If it's results seem to directly contradict 30 years of ongoing research from the very respectable non-partisan Pew Research Center, it's probably got a little way to go before gaining credibility.

Michelle has more, as does Shrinkwrapped.


Post a Comment

Links to this post:

Create a Link

<< Home

Who Links Here