“Now, for all the skeptics in the audience: who wants to bet that the 2003 data set will not have an IQ-TIMSS correlation greater than .65 or that Ghana & Botswana will score significantly above their IQ-predicted position on the trend line?...this is a chance for the h-bd skeptics - especially those who think Richard Lynn's data is *totally* useless [1] - to venture an empirical, testable prediction.”
Yesterday, TIMSS 2003 results were released to the public. I of course whipped up a little Excel goodness. (It’s a quickie, so beware of errors, etc. I am happy to take corrections*.) So, what does the data say?
(* = white lie)
Here follows a table with:
a) IQ scores, adjusted and unadjusted from “IQ and the wealth of nations” by Lynn and Vahannen. (Gathered into a table by Steve Sailer)
b) GDP per capita from the same book
c) TIMSS math scores from TIMSS 2003 for grade eight students in those countries with an entry in the L&V dataset.
d) A few simple regressions I ran on the material. Not an econometrics wizard, so point out errors at will.
e) A nice little IQ-TIMSS graph, for your viewing pleasure.
Synopsis:
A TIMSS-Raw IQ score OLS regression yields an adjusted R-squared of 0,89.
A TIMSS-Adjusted IQ score OLS regression yields an adjusted R-squared of 0,82
Notes:
a) Couldn’t find Botswana in the L&V data set. Their 366 score did manage to beat out Saudi Arabia…
b) The Canadian TIMSS score is a mix of the scores for Quebec and Ontario.
c) The UK entry is the score for England, with a flag for too few participants.