Tuesday, November 04, 2008

IQ and homicide cross-nationally: I was so excited by the cross-state correlation between IQ and murder (what election?), I decided to examine the same relationship across countries. Combining data from Nation Master and Sailer's table of Lynn's IQ data, I calculated the Pearson correlation coefficient (N = 36). It is -.47.

I suspect the association is weaker than in the state data for at least two reason. Criminologists are skeptical about comparing levels of crime across nations because criminal justice systems are so different. I agree. Frankly, I don't trust data from countries like Zimbabwe--one of the 36 cases. If the country is excluded, for example, the correlation becomes -.53. (If I find time, I'll use World Health Statistics data which is more reliable.) In addition, this is a very diverse set of countries. There are probably many other factors (e.g., level of socioeconomic development) that play an important role, thus diminishing the impact of IQ.

No comments: