Is crime x g is positive among US Africans?

Sat 26 Jul 2014 10:33

Admin

Was reading an old review of two books by Lynn:

The Attack on The Bell Curve, By Richard Lynn
Personality and Individual Differences 26, (1999), pp. 761-765

http://www.eugenics.net/papers/RLSPECREV.htm

The relation between intelligence, crime and race is considered by Manolakes. She accepts H and M's contention that among whites IQ is negatively associated with crime. H and M did not consider this relationship among blacks. Manolakes faults them for this and finds that among blacks in the NLSY sample IQ is positively associated with crime. She criticises Hand M for not discovering this themselves, failing to note that Hand M were not primarily concerned with race differences. She has certainly made a remarkable discovery considering the large research literature showing that crime is predominantly committed by the less intelligent. Before taking this result too seriously it should be noted that the data consist of self-reported crime and people do not invariably report their crimes truthfully. Nevertheless her apparent discovery that IQ is positively related to crime among blacks certainly deserves further research.

This can't be right. It would also seriously question g as an explanation of the crime difference between US Euros and US Africans. Does anyone know of some dataset to try this analysis on?

Philbrick Bastinado

Sat 26 Jul 2014 16:34

Perhaps more intelligent blacks were simply more truthful –– dumber blacks might have worried that the "study" was a source of entrapment.

Emil O. W. Kirkegaard

Sat 26 Jul 2014 16:45

Admin

Before proceeding to various proposed models, it seems wise to first attempt to replicate the finding.

Curtis S Dunkel

Tue 29 Jul 2014 01:07

I took a brief look at the ADD Health data and there doesn't seem to be much of a difference between the correlations for Blacks and Whites.

Emil O. W. Kirkegaard

Tue 29 Jul 2014 01:22

Admin

Normally crime x IQ is reported as around -.2. But those are near zero, sometimes even positive. What's happening?

Curtis S Dunkel

Tue 29 Jul 2014 13:28

Normally crime x IQ is reported as around -.2. But those are near zero, sometimes even positive. What's happening?

My guess is a couple of things. The Peabody was given years prior to the self-report on crime and the crime measure is not aggregated across specific crimes. I could play with it to tighten things, but these initial results supply a slight indication that pattern is the same for Blacks and Whites.

Meng Hu

Wed 30 Jul 2014 18:30

Admin

I took a brief look at the ADD Health data and there doesn't seem to be much of a difference between the correlations for Blacks and Whites.

Thanks. Is it from the full sample Add Health ? Anyway the correlations can be small because of individual variables, not aggregated. But even aggregated when I did that before (John Fuerst asked me to do this) in the NLSY97 and Add Health (public data) I did not find a good relationship in black, white or hispanic sample, because the variable, even aggregated was strongly non-normal. That is, the values range between (if my memory is still correct) 0 and 6 or 7 maybe. But most values range between 0 and 1. Most people say "I never did that" or they say "only once". I think however that my method of doing simple correlation was wrong. I should have used regression (not necessary multiple regression) and "save" the predicted value of the dependent var (crime) and then correlate it with IQ (ASVAB) and making separate plot for each race (that could have been entered in the regression, or not). Alternatively, you can simply make cross tabulation and look at mean score of IQ and mean score of crime/delinquency variable.

Concerning the Manolakes study, I'm glad to see I was not the only one who has spotted this thing when reading Lynn's review, but I wasn't interested in it. I know, however that no one else has replicated the study, and google scholar has only 10 references who cited it. So, you can attempt to replicate that if you wish.

If you are interested, I still have my syntax for these analyses I have made 1 year ago now...