As a rule of thumb (ha ha), in men the index finger is shorter than the ring finger, and in women the index finger is longer than the ring finger.
This appears (at least in preliminary experiments with other mammals) to be due to hormonal balances in utero, to wit androgen, testosterone et al.. This means that many other sex-linked traits (biological ones, not cultural) can be correlated to finger length-- for example, one study* supposedly found that women with shorter index fingers (normally a "masculine" trait) were less likely to get lost.
*I don't have a reference; the Scientific American article just said "studies have found" and went on from there. And now that I look at it, I'm REALLY damn skeptical. I couldn't track down the actual study, but Wikipedia, funnily enough, does have an impressive list of links to other studies involving second/fourth finger ratios, such as this one (http://www.ehbonline.org/article/S1090-5138(99)00026-4/abstract) investigating musical ability.
Haven't read any of these studies, but one question to ask when evaluating correlational human research involving large numbers of subjects is how much of the variance is accounted for by the main variable.
ReplyDeleteIn other words, how likely is it that someone does not fit the pattern. Often it is quite.
Such research may be relevant to populations and theories, but it is not of any practical significance on the individual level.
"Often it is quite"
ReplyDeleteShould be 'quite small'.
The key word is 'correlated'.
A difference can be -statistically- significant but still very small in actual measures. The larger the number of subjects, the smaller the difference needed for statistical significance.
ESP research is a good example.
If you run 10,000 trials (not unknown in the field) and subjects average 50.5% correct when chance predicts 50.0%, that could be statistically significant. Such a small effect, however, is more likely due to an experimental artifact that a real difference.
In this case, the artifact could involve sampling, since this is data mining rather than a 'real' experiment where subjects are randomly assigned to treatment conditions and given different treatments.