Environmental factors relate to homosexual and heterosexual marriage: Danish study

I just received the following study and have written a brief review of it.

Frisch, M., & Hviid A. (2006). Childhood family correlates of heterosexual and homosexual marriages: A national cohort study of two million Danes. Archives of Sexual Behavior. [Epub ahead of printing].

Department of Epidemiology Research, Danish Epidemiology Science Center, Statens Serum Institut, 5 Artillerivej, DK-2300, Copenhagen S, Denmark, [email protected].

ABSTRACT: Children who experience parental divorce are less likely to marry heterosexually than those growing up in intact families; however, little is known about other childhood factors affecting marital choices. We studied childhood correlates of first marriages (heterosexual since 1970, homosexual since 1989) in a national cohort of 2 million 18-49 year-old Danes. In multivariate analyses, persons born in the capital area were significantly less likely to marry heterosexually, but more likely to marry homosexually, than their rural-born peers. Heterosexual marriage was significantly linked to having young parents, small age differences between parents, stable parental relationships, large sibships, and late birth order. For men, homosexual marriage was associated with having older mothers, divorced parents, absent fathers, and being the youngest child. For women, maternal death during adolescence and being the only or youngest child or the only girl in the family increased the likelihood of homosexual marriage. Our study provides population-based, prospective evidence that childhood family experiences are important determinants of heterosexual and homosexual marriage decisions in adulthood.

Read my review and let’s discuss…

11 thoughts on “Environmental factors relate to homosexual and heterosexual marriage: Danish study”

  1. It might be reasonable not to dismiss the birth order effect but the study did not argue from theory. The birth order effect is compromised because in at least this population, it did not show up empirically.

    The rates of heterosexual marriage are increased in families where there are older sibs versus children with fewer older sibs. These are not huge effects but they increase probabilities by a small amount. Some heterosexuals may cohabit, some remain single.

    Even in convenience samples, the birth order effect for homosexuality is not large, explaining only about 1% of the variance. This study suggests that in an entire population the birth order effect goes away.

  2. “If there is a birth order effect, it would not differentiate … unioned and ununioned gays”

    I think you are making an assumption that is contradicted by this study.

    I think we can agree that if there were something about birth order itself that influenced whether a gay person were likely to marry, then married gay persons would not be representative of gay persons as a whole when discussing birth order. In other words, birth order would differentiate unioned from ununioned gays.

    It might initially seem counterintuitive that birth order in married gays and unmarried gays may be different It might seem that I’m grasping for reasons to discount your observation. However, the abstract of the study itself states that birth order seems to impact marriage decisions:

    “Heterosexual marriage was significantly linked to having young parents, small age differences between parents, stable parental relationships, large sibships, and late birth order.”

    We know that “late birth order” is not a trait related to being straight. 90 plus percent of the population cannot have the attribute of “late birth order”. Thus clearly birth order has an influence on the decision to marry, at least for straights.

    Given that birth order has an impact on straight marriage, wouldn’t you agree that it is not reasonable to dismiss it as having no impact on gay marriage? The impact may be different (as is the rural/urban split) but surely it can’t be dismissed offhand.

    Thus we can see – or so I believe – that any observations about married gays that discusses birth order cannot be extrapolated to the gay population at large.

    It’s somewhat like saying that we can observe whether unmarried gay people wear wedding rings by looking at what married gay people do. It seems as though there’s a built-in bias.

    So the quotes about “these homosexual men” can’t be extrapolated to anything beyond “these homosexual men”.

  3. Timothy: I am not following. How/why would these populations be different? If there is a birth order effect, it would not differentiate urban and rural Danes, or unioned and ununioned gays. At least, I cannot see what basis would account for this biological factor to operate in one population and not in another.

  4. Warren,

    I’m sorry if I was confusing. I wasn’t suggesting that the gay married people were not a subpopulation of the gay people. Just that they are probably not a representative subpopulation from which any conclusions could be derived.

    If I read this correctly, all that can be determined from this study is what type of people marry. It would seem that the type of straight person that marries is somewhat different than the type of gay person that marries. This suggests to me that the same social forces result in different responses determined by whether the subject is gay or straight.

    For example, in the country there seems to be greater social pressure for straight people to marry. But this same culture seems to discourage gay people from marrying (I suspect that would also be true in the US as well). However, this observation could not then be used to assume that people are less likely to “become” gay if they are raised in the country. That would be an extrapolation that does not seem to be evident. It is probably more likely that gay persons are less likely to ruffle feathers in a small town by getting married.

    That’s all I’m cautioning, that a look at marriage not be taken to be a look at orientation.

    We know that married straight people are not representative of straight people in general. For example, in the US they are more likely to be religious, live longer, more likely to have a “nuclear family” dynamic, and (I believe) more likely to be politically conservative than are unmarried straight people. To look at married straight people and then make assumptions about all straight people would be very poor social science indeed.

    I don’t think it a stretch to believe that gay married persons differ from gay unmarried persons in Denmark. The “more likely’s” are probably not exactly the same, but its fair to assume that they exist and probably are much more extreme.

    So to look at married gay Danes and say, for example, that birth order observations can be extrapolated to all gays, would be rather presumptive, I think.

  5. While there may be some small number of people who are homosexually married who are not gay, I think it must be safe to say that the vast majority of people marry someone of compatible sexual orientation. There are some mixed marriages, of course, but I suspect choosing a gay marriage is a pretty good tip for same-sex attraction. The population is then gay people who marry other gay people. This is a subset of all gays and as such interesting for the correlates associated with them. So I think something can be gleaned but there are some red flags that I have written to the authors about and will post when I get some answers. For now, let me temper the enthusiasm for the environmental effects by saying that the effect sizes are very tiny, probably trivial for father absence, and only slightly higher for marital disruption. Urbanization may have slightly more robust effects.

  6. Post hoc, ergo propter hoc. This error in logic is used over and over again by folks who have already made up their minds that being gay must be a disorder “caused” by something. Mr. Gatt is correct, correlation is not causality.

  7. “The study contradicts the fraternal birth order hypothesis. From the study, “…we found no indication that older brothers were particularly common in these homosexual men. Rather, older siblings, whether brothers or sisters, were positively and linearly linked to higher rates of heterosexual marriage in our study.”

    Warren, I disagree. This is not a study of gay people. Nor is it a study of those who choose straight marriage v. gay marriage like one might choose the color of ones socks.

    This simply looked at attributes of gay persons who were married and straight persons who were married. Assuming that married gay people are a representative sample of all gay people is a very weak support for thinking it refutes the fraternal birth order effect.

  8. “This is fascinating, because the sample’s SO big the clear difference between straight and gay and the absence of fathers is just undeniable…You’re also right that this data challenges the birth order hypothesis. All the indicators of environment having a role to play in sexual behaviour and preference are there – right down to the urban / rural split.”

    No, Peter, I believe you are misreading this report. As tempting as it might be to conflate gay people and gay marriage, it simply isn’t the same.

    This report tells us nothing at all about gay people. All it addresses are people who marry. And I’m quite sure that you would readily agree that “married people” is not a statistacally valid sample of “people”. We all know that there are differences.

    While this study tells us nothing at all about gay people, it does tell us some interesting things.

    1. Gay Danes raised in rural areas grow up with less expectation of marriage than those raised in urban Denmark. (this doesn’t come as much of a surprise to me).

    It seems logical that the expectations of the option to marry for gay people has not fully reached all of Danish society. It was embraced in urban areas first. I suspect that once all areas of Denmark are equally accepting of gay marriage then you will see this trend to reverse and mirror the straight marriage trends.

    2. It’s curious that family instability resulted in less likelihood for straight marriages and more likelihood for gay marriages. I’m not sure what to make of that.

    3. Interestingly, younger children – both gay and straight – were likelier to marry while older children were less so. I wonder if that is universal or is only a Danish thing.

    4. The big startling revelation is… (ta-da) socialization effects ones decision whether or not to marry.

  9. 2 million? That’s not a sample, that’s the real population deal isn’t it?

    This is fascinating, because the sample’s SO big the clear difference between straight and gay and the absence of fathers is just undeniable. Jimmy might be right that correlation doesn’t necessarily imply causation, but in this instance I’d want to look at the dataset and ask “is there any other reason?”

    You’re also right that this data challenges the birth order hypothesis. All the indicators of environment having a role to play in sexual behaviour and preference are there – right down to the urban / rural split.

    Fascinating. How DID they get 2 million people involved?

  10. Very interesting article and a good review. It certainly is not directly addressing the issue of causation of homosexuality. Rather marriage in the context of homosexuality or heterosexuality is what is discussed here. Are most homosexuals married/single by the way? any other studies?
    Very nice review. Thank you Mr. Throckmorton. Looking forward to your next post.

  11. 1. Correlation does not imply causation.

    2. The question of “why are people gay?” is not important. The question that should be asked is, “is it wrong to be gay?” The answer to that, of course, is no. Being gay, like being straight, is an amoral concept and entirely defined by our culture.

    3. The ex-gay movement is the attempt by anti-gays (people who dislike gays and want them to suffer) to give gays a path out of the suffering that anti-gays act to impose. I like to call it “Christian love”, and those scare quotes are EARNED.

Comments are closed.