As I believe Clark has argued, there is often tradeoff between income and other desirable aspects of the job. Many university professors make much less than they would doing something else, but like independence and job security and schedule flexibility. Another reason why correlations aren’t higher.
We need more metrics about each person. The most powerful approach is genetic. For some large enough sample of people sequence high income people and academic people and test their IQs. For the same IQ range what separates them? Also, measure them on personality attributes, beliefs, and values.
Interesting. Maybe ideally we'd have measures of income plus benefits, including intangible benefits like those. Or I think there's a whole literature on measuring "status" per se, of which income and wealth are components.
Tangentially related, I wonder what would happen to the IQ-income correlation before and after including all taxes and net transfers.
> not adjusting for the fact that some years yield more income than others
Do you mean adjusting by imputing outlier incomes by fitting an income trajectory for each individual?
> other biases in the IQ~income correlation in other datasets like restriction of range or measurement error in intelligence
You mean that measurement error in intelligence is particularly low in NLSY and the IQ range particularly broad, leading to a high estimate of this correlation?
Excellent article! I was just pointing this out to someone the other day, but you made the point much better than I could. I also like to point out to people who are overly critical of IQ that they never offer a better measure to predict across life domains. The only one someone ever offered was g, which I then told him was extremely highly correlated to IQ and wouldn't change the results much.
Until they can find a better construct which explains as much individual, group, and national success in the modern world, their criticisms of IQ, while valid considerations, ring a lot more hollow. For example they bring up cultural bias, which in some ways is valid but in other ways is a bit silly. What if some cultures are better able to build economies and integrate into Western or globalized economies? I think this is certainly true but it seems the implication is offensive to some.
As I believe Clark has argued, there is often tradeoff between income and other desirable aspects of the job. Many university professors make much less than they would doing something else, but like independence and job security and schedule flexibility. Another reason why correlations aren’t higher.
We need more metrics about each person. The most powerful approach is genetic. For some large enough sample of people sequence high income people and academic people and test their IQs. For the same IQ range what separates them? Also, measure them on personality attributes, beliefs, and values.
Interesting. Maybe ideally we'd have measures of income plus benefits, including intangible benefits like those. Or I think there's a whole literature on measuring "status" per se, of which income and wealth are components.
Tangentially related, I wonder what would happen to the IQ-income correlation before and after including all taxes and net transfers.
The correlation would not change because the rank-order is largely unaffected, but the net effect would reduce due to decreased variance in income.
Nice analysis.
> not adjusting for the fact that some years yield more income than others
Do you mean adjusting by imputing outlier incomes by fitting an income trajectory for each individual?
> other biases in the IQ~income correlation in other datasets like restriction of range or measurement error in intelligence
You mean that measurement error in intelligence is particularly low in NLSY and the IQ range particularly broad, leading to a high estimate of this correlation?
Excellent article! I was just pointing this out to someone the other day, but you made the point much better than I could. I also like to point out to people who are overly critical of IQ that they never offer a better measure to predict across life domains. The only one someone ever offered was g, which I then told him was extremely highly correlated to IQ and wouldn't change the results much.
Until they can find a better construct which explains as much individual, group, and national success in the modern world, their criticisms of IQ, while valid considerations, ring a lot more hollow. For example they bring up cultural bias, which in some ways is valid but in other ways is a bit silly. What if some cultures are better able to build economies and integrate into Western or globalized economies? I think this is certainly true but it seems the implication is offensive to some.