In my last post I talked about the Wealth Achievement Quotients of various ethnic groups.  For example, Jews, are 2% of America, but 35.75% of the Forbes 400 richest Americans list, thus Jews have a Wealth AQ of 17.88 (because 35.75/2 = 17.88).

Using data cited by iconic blogger Steve Sailer, I calculated the Wealth AQ of every American racial group for which I have IQ data.

Blacks (average IQ 85): 0.25/13.6 = 0.0184
Native Americans (average IQ 86): 0/1 = 0.0
Hispanics (average IQ 89): 0.5/17.08 = 0.029
White Gentiles (average IQ 100): 58.5/59.6 = 0.98
East Asians (average IQ 104): 2/3.62 = 0.55
Indians: (average IQ 110): 1/0.9 = 1.11

The problem is, when I used the above six data points to build a regression equation predicting Jewish IQ from their Wealth AQs of 17.88, I got a ridiculously high average IQ 250!

However I have now decided to calculate the natural logs of these Wealth AQs, excluding Native Americans because you can’t get a natural log for zero.  The natural log Wealth AQs are:

Blacks (average IQ 85): 0.25/13.6 = -4.017
Hispanics (average IQ 89): 0.5/17.08 = -3.54
White Gentiles (average IQ 100): 58.5/59.6 = -0.0202
East Asians (average IQ 104): 2/3.62 = -0.5978
Indians: (average IQ 110): 1/0.9 = 0.104

I then calculated the regression line of best fit predicting ethnic mean IQ from its natural log Wealth AQ:

log

The regression equation is:

Mean IQ =105.51289532299+4.9020538489579(natural log of Wealth AQ)

The Jewish Wealth AQ is 17.88 which has a natural log of 2.88.  Entering 2.88 into the above formula gives Jews an expected mean IQ of about 120.  This is a far more reasonable estimate than the ridiculous 250 I got before I reduced Wealth AQs to their natural logs, and shows the advantage of using natural logs when dealing with abnormal distributions.

Indeed an IQ of 120 is pretty close to the actual average verbal IQ of Ashkenazi Jews.

Advertisements