Dozens of studies in the past few years have linked single genes to whether a person is liberal or conservative, has a strong party affiliation or is likely to vote reguarly. The discipline of “genopolitics” has grabbed headlines as a result, but is the claim that a few genes influence political views and actions legitimate?
We don't think so. The kinds of studies that have produced many of the findings we question involve searching for connections between behavior and gene variants that occur frequently in the population. Most of the 20,000 to 25,000 human genes come in hundreds or thousands of common variations, which often consist of slight differences in a gene's sequence of DNA code letters or in repeats of a particular segment. For the most part, scientists do not know what effect, if any, these common variants, known as polymorphisms, have on the functioning of the proteins they encode. Genes predict certain well-defined physiological diseases—such as hereditary breast cancer and the risk of developing Alzheimer's disease—but when it comes to complex human behaviors such as voting, the link is tenuous at best.
One of the most prominent papers showing a link between a few polymorphisms and political behavior was published by James Fowler and Christopher Dawes in 2008 in the Journal of Politics. They concluded that people who possess certain variants of a gene called MAOA are more likely to vote than those who do not and that people with a particular variant of a gene known as 5-HTT who regularly attend religious services are also more likely to vote. We do not believe that these conclusions are right.
On supporting science journalism
If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.
Like most claims that a specific gene predicts variations in a particular behavior, the findings were based on what is known as a candidate gene association study. Instead of surveying all the genes in the human genome for possible associations with a given trait, such studies look for potential links between polymorphisms for one or two candidate genes and a specific trait. This type of study can be a relatively inexpensive way to conduct research because it usually depends on large databases of information that already exist, but it can lead researchers astray.
We identified two major problems with the study of Fowler and Dawes. First, they misclassified the genes they were studying in a way that amplified the statistical significance of their findings. Second, their methods fell short of adequately taking into account population stratification, in which the frequency of polymorphisms varies from one ethnic population to another as a result of unique ancestral patterns of migration and mating practices. (This is a common problem in the field.) When we analyzed the different ethnic groups in detail, we found inconsistencies. For instance, in the case of Asians, Native Americans and nonwhite Hispanics, we saw the opposite trend—toward less voting.
Yet we have more fundamental issues with these kinds of studies. The same polymorphisms of these same two genes that have been tied to voting are also said to predict variation in other behavioral and physical traits—irritable bowel syndrome, schizophrenia and premature ejaculation. Such broad findings beggar belief. The idea that a pair of genes could be responsible for so many disparate behaviors is biologically implausible.
Recent research provides growing evidence that genetic influences on human behavior involve thousands of different genes, which influence one another and the environment in intricate ways. Differences in aggression among fruit flies, to take just one example, entails the activity of more than 4,000 genes. The chance that any complex human behavior—such as voting—might have one or two major predisposing genes is practically zero.
SCIENTIFIC AMERICAN ONLINE
Comment on this article at ScientificAmerican.com/nov2012
