04 September 2008

The Bad Husband Gene

Scientists have identified a gene that influences brain activity that, when present in men, is significantly associated with not marrying, divorcing, and having an unhappy wife.

[The study] not only links the gene variant — which is present in two of every five men — with the risk of marital discord and divorce, but also appears to predict whether women involved with these men are likely to say their partners are emotionally close and available, or distant and disagreeable. The presence of the gene variant, or allele, also seems predictive of whether men get married or live with women without getting married.

"Men with two copies of the allele had twice the risk of experiencing marital dysfunction, with a threat of divorce during the last year, compared to men carrying one or no copies," said Hasse Walum, a behavioral geneticist at the Karolinska Institute in Stockholm who led the study. "Women married to men with one or two copies of the allele scored lower on average on how satisfied they were with the relationship compared to women married to men with no copies."

The scientists studied men because the hormone being examined is known to play a larger role in their brains than in women's brains.

Linking a single gene to such a wide range of behaviors at a subclinical level is remarkable. The vast majority of mental traits with a strong hereditary component either have a genetic (or epigenetic) cause that has not been identified, or are the product of multiple different genes that operate only in combination. The X v. Y chromosome that determined a person's genetic gender, and the "fruitless" gene in fruit flies which determines sexual orientation in fruit flies, are the only other simply inherited genes of which I am aware that have such pervasive influences on complex behaviors.

I look forward to tracking down the original study to look at the methods used and the significance of the relationships found. Typically only about half of the variation in even the most strongly genetically determined traits is hereditary, and one would expect a weaker relationship in expression of trait in a context also strongly influenced by cultural and economic factors like marital health.

It is also interesting that a trait which would seem to hurt one's reproductive chances would be so common. The title to the cited article called it an "infidelity gene" despite the text of the article not clearly suggesting that result. But, perhaps infidelity is the factor that provides those who have it with a reproductive advantage, or perhaps paying less attention to a spouse pays off through economic success achieved by paying more attention to work, or perhaps the gene was once important in helping widowed men move on to new relationships, when many women died in childbirth leaving behind young children who might perish if raised only by a father.

Of course, being a good husband has not been very important to one's ability to stay married until about a century ago, when divorce started to become more common, and the ability of women to choose a spouse freely is also relatively new by evolutionary standards.

The fact that the gene may not express meaningfully, or at least in the same way, in women may also be a relevant factor in allowing it to persist. It could be a gene that is reproductively beneficial in women, and reproductively detrimental in men, for example.


More details:

The gene in question, AVPR1a, governs a receptor that regulates the brain's production of vasopressin, a hormone that contributes to attachment behavior with mates and offspring. A few years ago, scientists found that when they added extra copies of the AVPR1a gene to the brains of promiscuous meadow voles, the animals began acting more like monogamous prairie voles, spending more time with partners and grooming offspring. A similar role for the AVPR1a gene has been observed in chimps and bonobos. . . .

A team led by Hasse Walum of the Karolinska Institute in Stockholm, Sweden, sequenced the AVPR1a gene in about 500 pairs of adult same-sex Swedish twins, all of them married or cohabiting for at least 5 years, and their partners. One variation of the gene was particularly common; about 40% of males had either one or two copies of a version--or allele--of the gene known as "334." . . .

The tests included a Partner Bonding Scale containing items that reflect affection and cooperation, such as "How often do you kiss your mate?" and "How often are you and your partner involved in common interests outside the family?"

Scores on the test were significantly lower for the men carrying either one or two copies of allele 334 than for those without it, the researchers report online this week in the Proceedings of the National Academy of Sciences. The highest score possible is 66; those without the allele scored an average of 48, whereas carriers of one copy of allele 334 scored an average of 46.3. Carriers of two 334 alleles had the lowest scores of all, averaging 45.5. Although the score differences seem small, Walum says they are statistically significant. (No connection was observed in females.)

More striking were the answers to questions as to whether the men had experienced a marital crisis or threat of divorce during the prior year. More than one-third of carriers of two 334 alleles said yes, compared with only 15% of those with no 334 allele. What's more, 32% of those with two alleles were unmarried, compared with 17% of those who didn't carry the allele. Evaluations of the relationship by the men's partners tended to correspond with assessments reported by the men themselves.

The behavior appears to be somewhat heritable. Because the researchers were dealing with a population of twins, they were able to separate genetic and environmental influences. They found that about 28% of the behavior could be chalked up to inheritance, which is similar to what other studies have estimated for the heritability of marital satisfaction and of divorce.

The last sentence is particularly notable. It suggests that a single gene accounts for essentially all of the heritability of marital satisfaction and divorce.

The abstract is here.


Another interesting study by one of the authors of this study in 1997, using an overlapping data set notes that:

Controllable, desirable, and undesirable life events were revealed [to show] significant genetic variance for women. There was no significant genetic variance for either sex for uncontrollable events. Multivariate analyses of personality (as indexed by Neuroticism, Extraversion, and Openness to Experience) and life events suggest that all of the genetic variance on controllable, desirable, and undesirable life events for women is common to personality. Thus, in this sample of older adult women, genetic influences on life events appear to be entirely mediated by personality.

A 1996 study of heritability of personality by other authors note that:

The genetic and environmental etiology of the five-factor model of personality as measured by the revised NEO Personality Inventory (NEO-PI-R) was assessed using 123 pairs of identical twins and 127 pairs of fraternal twins. Broad genetic influence on the five dimensions of Neuroticism, Extraversion, Openness, Agreeableness, and Conscientiousness was estimated at 41%, 53%, 61%, 41%, and 44%, respectively. The facet scales also showed substantial heritability, although for several facets the genetic influence was largely nonadditive. The influence of the environment was consistent across all dimensions and facets. Shared environmental influences accounted for a negligible proportion of the variance in most scales, whereas nonshared environmental influences accounted for the majority of the environmental variance in all scales.

The larger project that was the basis of the "bad husband gene" conclusion also make some conclusions about intelligence heritability in a 1994 study:

General cognitive ability yielded a heritability estimate of about .80 in two assessments 3 years apart as part of the Swedish Adoption/Twin Study of Aging. This is one of the highest heritabilities reported for a behavioral trait. Across the two ages, average heritabilities are about .60 for verbal tests, .50 for spatial and speed-of-processing tests, and .40 for memory tests. For general cognitive ability, the phenotypic stability across the 3 years is .92 and stable genetic factors account for nearly 90% this stability. These findings suggest that general cognitive ability is a reasonable target for research that aims to identify specific genes for complex traits.

About The Sample

The use of a Swedish sample in these studies have pluses and minuses. Sweden is a quite homogeneous society with a strong social safety net. The positive aspect of this is that "noise" from environmental factors like cultural differences between subgroups in the data are unlikely to cloud any genetic effect that exists.

Also, the relatively low rates of marriage and relatively high rates of divorce in Swedish society, compared to many other cultures, lead to non-marriage rates and divorce rates that are numerically high enough to produce clear differences between the subgroups in the study.

Further the relative ease with which marriages are ended, or in which relationships can be established without social condemnation in the absence of marriage, again moves the focus away from environmental and economic factors as a cause of relationship instability and towards more personal inherent personality factors.

The negative aspect of this reality, is that the importance of genetic effects are probably overstated. Sweden's more egalitarian society, for example, does not have the gross disparities in marriage rate and divorce rate based upon socio-economic status to nearly the extent that the United States does. One suspects that "mixed marriages" by any conveniently comparable measure are far more rare among Swedes in Sweden than among Americans.

In short, I would expect lower heritability estimates for the outcomes measured in more diverse studies outside Sweden than I would in this study.

Replication and Thorny Questions

This is a study that is easy to replicate. There are multiple twin studies already in place. Most of the outcome data could be obtained quickly and easily with a fairly simple survey of a modest sized population that has already agreed to participate in studies trying to resolve this kind of issue. The genetic tests involved could likewise be quick and non-invasive.

Assuming that replications of this study show that there are statistically significant effects from the gene in question cross-culturally, regardless of the precisely magnitude of the heritability estimate in different studies, at least a couple of thorny questions present themselves.

One is whether genetic testing for the trait is appropriate for the general population of people contemplating marriage, potentially improving prospects for about 60% of the population, while perhaps materially harming the marital prospects of about 40% of the pouplation, at least until benefits of this gene to those who have it are established.

Another is that the 23% share of this allele in the general gene pool, implied by the fact that 40% of men have at least one of these genes (and assuming, perhaps naively, independence of the frequency of this allele in mothers and fathers), almost certainly varies in different populations. Moreover, unlike intelligence, which is measured almost entirely by observed manifestation of a trait, rather than direct genetic testing, the prevalance of this trait in different populations could be established with precision not reasonably subject to question. In short, it would be a rehash of the controversy over The Bell Curve which considerably more bite due to the far lesser degree of methodological controversy involved. This, in turn, could fuel stereotypes about the fidelity of members of particular populations inappropriate to make about individuals.

A Rosetta Stone?

The allele studies is one of the strongest genetic links to what appears to be basically a personality trait. Like a Sudoku puzzle, a piece of information supplied with certainty in any one part of the puzzle sheds light on the entire picture.

Psychology now is where biology was when Linneas was first starting to develop taxonomies based upon observed characteristics before the theory of evolution was developed. Definitions of personality traits and mental health conditions are based upon observation and guesswork, without any definitive framework to guide researchers.

For example, pre-evolutionary biologists had to decide whether bats which shared fur and nursing with rats were more closely related to bat than birds which shared winged flight with bats. Sometimes the classifications have appeared wise in retrospect. At other times, the taxonomy has been substantially reworked.

For example, we have little way of determining whether symptoms that manifest as what we call schitzophrenia are multiple disorders, or a single basic disorder.

The ground is even muddier in the area of personality. The leading five factor theory of personality in psychology was developed basically with thesauruses, questionaires, and statistical analysis. But, words are slippery things, and words we use to describe ourselves are even more perilous.

If mastery of the scope and definition and mechanism of the single gene genetrically determined personality trait measured in this study were better understood, it might be possible to pin down more single gene or small number of gene genetically determined personality traits. This, in turn, might sharpen our understanding at a biochemical level of what a personality trait actually is, within our bodies, and how the areas of personality for which we cannot find genetic markers are best defined.

For example, one obvious study that suggests itself, given this research, particularly if it is validated, is to gather a large random sample of men, test them for these gene, and examine the differences between the groups with two, one or no alleles of it, in all aspects of their lives. This could be done with a far less difficult to assemble population than a twin study. Any several hundred men would do. Indeed, a more intensely interviewed smaller study might actually be more helpful at a preliminary stage. Once you can identify who does and doesn't have a gene with certainty, and you know that it is important, it is possible to do much more definitive research about its effects.

We know that all sorts of things are corrolated with not marrying and/or getting divorced. It would be fairly quick work to determine which of these things flow from the personality of the unmarried or divorced person and which of these things flow from the non-marriage or divorce itself, rather than the underlying personalities of those invovled in it.

One psychological trait associted with non-marraige or divorce in the United States is IQ. Given that this particular trait is not obviously synonomous with general intelligence (indeed, it comes closer to the popular pyschology notion of EQ), and that in Sweden this trait describes almost all heritable variation in relationship stability, it is likely that further study would show that in the U.S. almost all IQ impacts on relationship stability are mediated by the economic factors that are so strongly influenced by IQ in the United States. This in turn, might motivate programs by providing solid evidentiary support for the already strongly suggested notion that family economic stability can help remedy the low marriage rates and high divorce rates of low income Americans.

No comments: