Read on Twitter
A bizarre but surprisingly common argument is some version of: "SAT/IQ/etc is merely about parental income and nothing else."
This is simply empirically false. First off, the correlation between wealth/income of parents and offspring score is not even that high! [1/n]
![A bizarre but surprisingly common argument is some version of: "SAT/IQ/etc is merely about parental income and nothing else."This is simply empirically false. First off, the correlation between wealth/income of parents and offspring score is not even that high! [1/n]](https://pbs.twimg.com/media/EuSLugnWYAMHMqy.png "A bizarre but surprisingly common argument is some version of: \"SAT/IQ/etc is merely about parental income and nothing else.\"This is simply empirically false. First off, the correlation between wealth/income of parents and offspring score is not even that high! [1/n]")
This is simply empirically false. First off, the correlation between wealth/income of parents and offspring score is not even that high! [1/n]
2/n For example, a correlation of ~0.24 between parental income and SAT score.
Even being overly generous about measurement error etc, this fact alone shows that the vast majority of variation in SAT scores is not shared with parental income.
Source:
https://twitter.com/ZachG932/status/1275537038859796481
Even being overly generous about measurement error etc, this fact alone shows that the vast majority of variation in SAT scores is not shared with parental income.
Source:
https://twitter.com/ZachG932/status/1275537038859796481
3/n So that alone shows the argument to be wrong.
More importantly however, the correlation that does exist cannot just be attributed to supposed beneficial effects of being raised in a wealthy environment.
"Correlation ≠ causation" is a cliché, but true, statement.
More importantly however, the correlation that does exist cannot just be attributed to supposed beneficial effects of being raised in a wealthy environment.
"Correlation ≠ causation" is a cliché, but true, statement.
4/n A large body of evidence shows that the association between socioeconomic environment (SES) and IQ is not just a cross-sectional artifact. IQ longitudinally predicts *mobility* in SES. An individual's likelihood of moving up/down in SES is associated, in part, with their IQ.
5/n Thus, the association between IQ and SES is an example of a selection effect: in every generation, people who move up or down in SES differ, on average, in IQ. This is at least part of the explanation for the observed association.
6/n An example of such a study (of which there are many) is:
Nettle 2010 ( https://doi.org/10.1348/000712603322503097)
Similar associations are seen between siblings of the same household. The better testing sibling is more likely to earn higher income ( https://doi.org/10.1016/j.paid.2019.05.045).
Nettle 2010 ( https://doi.org/10.1348/000712603322503097)
Similar associations are seen between siblings of the same household. The better testing sibling is more likely to earn higher income ( https://doi.org/10.1016/j.paid.2019.05.045).
7/n Adoption studies (which control for systematic genetic confounding) represent another way to test for causal beneficial effects of rearing socioeconomic status. Such studies find little to no correlation between rearing SES and IQ.
Source:
https://link.springer.com/article/10.1007/s10519-007-9142-7
Source:
https://link.springer.com/article/10.1007/s10519-007-9142-7
8/n Conclusion:
Performance on cognitive tests is not merely about wealth/SES. The correlation is too weak for it to be even close to the only factor.
Furthermore, the (already modest) correlation is an overstatement of the effects of SES; much of it is driven by selection.
Performance on cognitive tests is not merely about wealth/SES. The correlation is too weak for it to be even close to the only factor.
Furthermore, the (already modest) correlation is an overstatement of the effects of SES; much of it is driven by selection.
