By now everyone has read the article that the heritability (definition 2) of "political ideology" is ~0.5, but the heritability of political party is far lower. There's a lot of chew on here (unfortunately The American Political Science Review doesn't have this paper online), but note that this similar to what Tom Bouchard found out as regards to religous attitudes and beliefs of separated twins raised apart, heritability of "religious intensity" or "zeal" was ~0.5, but there was very little concordance of religious denomination.

Two points:

• The likely distribution of these tendencies suggests to me that it is better to think of them as heritable polygenic traits with strong environmental components (plus the whole correlational and interactional factor) as opposed to "genetic" traits that can be conceived of as "modules." This distinction isn't academic, it has implications with how people approach the topics.
  
• Obviously aspects of your phenotype are going to shape how your life-history tracks in a given social matrix and how you perceive yourself and your values. I suspect for example ugly people will report on a survey that "looks aren't important to them," even though I doubt when given an opportunity an ugly guy or girl would turn down a hot date with a fine physical specimen to have dinner with someone with a great personality but who was revolting as sin. In other words, studies like this tell you a lot, but not necessarily in a transparent or simple fashion.
  

Finally, Steve brings up the point that whites in the northeast are far more liberal than whites in the south, but their genetic differences are probably minimal, at least in light of the wide ideological gap. I think the point about the social matrix is important, the complex behavorial-cognitive phenotypes don't crystallize in a vacuum. Not only does space matter, but time does. In short I think what these studies are pointing out are relative slots people will fit into given a particular norm of reaction. To use the religion analogy, Swedish Americans in North Dakota are probably genetically the same as Swedes in Sweden, but the latter are likely far more "progressive" and "secular" than their American cousins, the difference being between the social matrices in which the two similar genotypes are expressed. I suspect that what you are seeing is simply a sliding over of the distribution, more or less. Similarly, today the average "conservative" tends to espouse race blindness and the average "liberal" tends to espouse positive discrimination for groups that have suffered negative discrimination in the past, but, 50 years ago the average conservative tended to espouse, condone or tolerate negative discrimination against groups that traditionally suffered such treatment, while the average liberal espoused race blindness. The change, as a function of time, was simply the social matrix, not the frequency of genes.

The take home message for the activists is that if you change social values of the society (whatever that means in the concrete sense, I suspect it translates into coopting the elite organs of civil society) individuals will simply renormalize themselves. I think homosexuality is a good example of social attitudes renormalizing over the past generation (that is, something I've observed in my short lifetime).

Related: Reflections on the "God Module".