|
Real_Trouble -> RE: Could curvy women be more intelligent? (3/20/2008 3:33:12 PM)
|
These kinds of studies and the subsequent news reporting absolutely infuriate me; the vast majority of the time these claims are false upon review, but that will never be reported down the road. This immediately strikes me as one of those things after reviewing the details of the study I could find, as the methodology seems extremely sketchy and biased. "Attempting to prove what you already believe..." seems to be likely here.
To be fair, I have seen things that were less intelligent, but most of the people invovled with them are now dead, usually through their own doing.
Here are some key points:
- The variance in intelligence between these two groups is well within the expected variance for the intelligence tests they are using; when your variance bounds contains the zero value, this means you have not achieved a statistically significant result, even if it trends in one direction.
- The sample size is extremely small (in comparison to the population of women), non-representative (ie, not a random sample), and time limited. I would suggest that all they have established, even if the findings are statistically significant, is that "among a small biased group in the following timeframe, the following oddity appears..."
- Lacking repetition is a fatal flaw for the vast majority of studies; if you give me a single set of data, I can find all sorts of fantastic correlations, but this does not demonstrate causation. It may be (and in fact probably is) random chance. Without independent testing and verification, one should be very, very skeptical of claims in any study.
One of my favorite papers to cite, because it really agitates the people who make these kinds of claims, can be found here:
http://medicine.plosjournals.org/perlserv/?request=get-document&doi=10.1371/journal.pmed.0020124&ct=1
In short: disbelieve studies that have not been repeated and expanded upon with rigorous attempts to disconfirm them, or which are in fields where there are significant amounts of bias. You are very likely to be correct.
|
|
|
|
