We instinctively assume that differences in behaviour that are in fact due to culture must be linked to – even caused by – characteristics of appearance. That is what the traditional notion of race is all about. But genetics has found no such innate origins of behavioural differences between “races” – and it is highly unlikely, given what we know about genetic variation, that it would. (my emphasis)
Because the term "race" is so poorly defined and because this categorization has been and still is an immensely important source of hate and discrimination, I will use the term "human populations" (despite the term being at least as vague). The fact that the term categorization into "races" does not make in any way a good representation of the partitioning of the genetic variance in human populations (e.g. see this post) is irrelevant to our discussion.
Are there any study who investigate the partitioning of genetic variance for a behavioural trait (e.g. drug dislike, aggressiveness, agreeableness, openness, reading performance, etc...) among "human populations" (defined one way or another)? Did they find a significant correlation? What is their correlation coefficient?
The article I refer to says
We instinctively assume that differences in behaviour that are in fact due to culture must be linked to – even caused by – characteristics of appearance.
In no way, do I expect that the phenotypic variation used to define races could "cause" much of an eventual genetic structure for a behavioural trait. I know it is a politically sensitive topic. Let's raise above it and discuss the evidence.
Estimating partitioning of genetic variance
If need be, see this post for a definition of heritability.
I realize that estimating the partitioning of genetic variance is a difficult endeavour and can easily be confounded with other sources of phenotypic variance such as the ones caused by social and educative systems. Any article that does its best to estimate a correlation coefficient between a behavioural trait and "races" is welcome.
Of course, the term "behaviour" is not perfectly defined. This has never prevented us from massively using this concept in biology and I don't think it should prevent us from using this vague categorization of phenotypes to answer the current question.
What behaviour am I interested in? ... pretty much anything that one might want to call behaviour. I am not particularly interested in behavioural differences that might be directly cause by a difference in the prevalence of a specific disease.
In the book IQ and the Wealth of Nations, Lynn and Vanhanen attempts measures of IQ among countries. These methodology has been highly criticized. I double that IQ could be considered as a behaviour but, while I would welcome anyone to include such work in an answer (along side its criticisms on their methodology; e.g. Witcherts et al., 2010; see also Rindermann 2007), I would rather like to focus on evidence that are not directly focused on IQ. I am fine with evidence that are on behavioural traits that somehow relate to our intuitive notion of intelligence (e.g. reading performance, spatial abilities, ...). This answer in the post Genes and Intelligence is very relevant and lists three articles that looks for loci that explain some variance in IQ.