Jump to content

User talk:169.231.178.183

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Hi, new to wikipedia, sorry if I inadvertently don't pose this correctly. I'd like to point out a problem with both references [57,58] appearing in

"Males tend to show greater variability on many traits, for example having both highest and lowest scores on tests of cognitive abilities,[9][54][55][56] though this may differ between countries.[57][58]"

[57] is a study into a particular set of high school mathematics aptitude tests...not IQ tests. I think that should be emphasized. As such, it's unclear, at least from reading the paper, whether the reversal in variance ratios is due to cultural differences between the countries or limitations of that particular test (e.g., score ceiling effects, non-gaussianity in test score distribution). Hopefully there's a more robust study into IQ differences in variability between countries. I recommend replacing 57 with a more appropriate set of references. Here is a start, comparing a large US + UK cohort and finds good reproducibility in variance ratios between these two countries: They don't look at IQ but make some effort to correlate results from their standardized CogAT tests with IQ Br J Educ Psychol. 2009 Jun;79(Pt 2):389-407. doi: 10.1348/000709908X354609. Epub 2008 Sep 25. Here is another meta analysis claiming that variance ratios are not reproducible between countries, although it seems that different tests were used in different countries, and the author admits that s/he was unable to control for ceiling effects or pathologies https://link.springer.com/article/10.1007/BF01420741

[58] is a relatively small study (n=150), a fact the authors acknowledge. It doesn't mention variance/variability once. It's true that figure 1 shows a narrower distribution in scores for men vs women. However, 1) The distributions are truncated gaussians, so likely to suffer ceiling effects 2) the main claim of the paper is that males have a statistically significant better score than females. The combination of 1) and 2) could well be responsible for a narrowing in the male score distribution (bunching near the top). FWIW, the M:F ratio in the top 5% is 3:1, which supports the hypothesis that males are more frequent in the tails. I think [58] is a weak reference, and should be dropped.

Welcome to this talk page

Start a discussion