Note that when discussion inequality, share of variaence explained is never. used.
If you pool all men and women from 8 billion, then compute which R^2 is income is explained by gender you don't get something impressive, but 77 cents for a dollar is impressive enough to be repeated.
Note that when discussion inequality, share of variaence explained is never. used.
If you pool all men and women from 8 billion, then compute which R^2 is income is explained by gender you don't get something impressive, but 77 cents for a dollar is impressive enough to be repeated.
How can you properly deal with non-linearity in data? For example when you compare grades with wealth.
You for example claim that:
>A test correlating .30 with performance can be expected to result in 67% of those in the top test quintile being above-average performers
But what if the reason why r=0,3 not 0,9 is big noise in top 30 % of results?
Great article.