Data sets are our friends

mathematicsDahlia Lithwick, an intrepid senior editor for Slate, has a piece up (a version of which also ran in the Washington Post yesterday) that makes the case for gay marriage and laws protecting gay parents. Reporters frequently cover gay marriage issues, but gay parenting rights are also in need of thoughtful analysis (the links below are to PDF files):

A heads-up to those of you still fretting about the alleged evils of gay marriage: The parade has moved on. Try as you may to vote or legislate your way out of a country that solemnizes such relationships, committed gay couples are already giving birth to, adopting, and fostering children. Whether or not same-sex marriage becomes widely legal in America, same-sex parenting is a done deal. . . .

According to the 2000 census, 34 percent of female same-sex households and 22 percent of male ones include children. Good data are extremely hard to obtain here, but the Lambda Legal Defense Fund estimates that 6 million to 10 million gay parents are caring for 6 million to 14 million children in this country.

And that does seem very decisive. One of my close gay friends is in a relationship in which he helps care for his partner’s children every other weekend. Another friend and former colleague moved with his partner to Ohio to be near his offspring. In both cases the men with children were previously married and seem to be very loving fathers. In neither case do the men serve as primary caregivers. But the data made me pause because the vast, vast majority of gay men and women I know who have absolutely no children in their lives. It could be that I know an extremely non-representative sample of gays, but let’s look deeper at the data.

There are only 78.1 million children under the age of 18, according to 2000 census data. Even though reliable data suggest the percentage is much smaller, let’s be generous and say that 5 percent of the United States population is gay. Does it really make sense that 5 percent of the population is caring for up to 18 percent of the children?

The census data that Lithwick links to says:

A reflection of changing life styles is mirrored in Census 2000′s enumeration of 5.5 million couples who were living together but who were not married, up from 3.2 million in 1990. These unmarried-partner households were self identified on the census form as being maintained by people who were sharing living quarters and who also had a close personal relationship with each other. The majority of these unmarried-partner households had partners of the opposite sex (4.9 million) but about 1 in 9 (594,000) had partners of the same sex. Of these same-sex unmarried-partner households, 301,000 had male partners and 293,000 had female partners.

mathbookOkay, so if 22 percent of male-male households have children, that’s 66,220 male gay couples with children they care for. And if 34 percent of female-female households have children, that’s 99,620 female couples that have children they care for. Same-sex households that aren’t gay are also included here, I should note, but let’s just leave that out of the equation.

So if gay couples are caring for up to 14 million children, as Lathwick states, that means that the 165,840 gay households with children in America care for an average of 36 to 84 kids.

Now perhaps she meant to include gay households such as the ones I mentioned earlier — which would completely change her argument since her point is “Try as you may to vote or legislate your way out of a country that solemnizes such relationships, committed gay couples are already giving birth to, adopting, and fostering children.” Also, from both a judicial and social analysis standpoint, caring for children and caring for children occasionally are two very different things.

Anyway, perhaps it is true that the average gay household with children includes 84 kids under the age of 18. Or, perhaps, Lithwick got a bit overzealous in her attempt to make a point.

Print Friendly

  • Erik Nelson

    I had a similar reaction when I read that story. Advocacy groups spit out statistics all the time, and are usually quite artistic in their use of exaggeration. But even the quick math I did in my head told me that these numbers couldn’t possibly be correct as they were stated.

    I don’t know if this is overzealousness or just another example of reporters not doing very basic fact checking before they make an argument. There is far too much reliance on talking points by some, and it is foolish. It doesn’t seem to be asking too much of reporters to do a quick check to make sure the numbers make sense.

  • matt

    many people can’t do math. sad.

  • Michael

    Since the census doesn’t count sexual orientation and those unmarried couples are completely self-identified, it’s fair to say those numbers are also far from accurate. If you live in many parts of the country, the chances you are going to self-identify to a census taker or even on a census form is pretty slim, especially if the census taker lives in your same town.

    Coupled with the fact that despite a PR campaign, the census identification is confusing even to a more “out” same-sex couple, it seems that the census data is probably even more suspect than LDEF’s data.

    The little data on same-sex parents shows that it is more common in non-urban settings, among African Americans and Latinos, and even in small towns. That could explain why looking around DuPont Circle or Chelsea is probably not the best indicator of the amount of same-sex parenting.

  • http://www.getreligion.org Mollie

    Michael, it’s funny you say DuPont is not where to look for same-sex parenting because one of the funniest nights of my life was walking around DuPont with my precious niece Anja in a Baby Bjorn. You would have thought I was carrying a previously undiscovered species of animal or something. She got so much attention it was ridiculous.

    Anyway, as for the data — census data is actually considered some of the most reliable data out there. Part of that is because there are multiple ways of asking the same census questions.

    But even if it has problems, do you think that gays are raising almost 20 percent of kids in America? Don’t you think the data have some problems?

  • Michael

    Mollie, I agree that Lithwick’s data is suspect, but I think the Census data is suspect in this new category. I think even the Census would agree with that. Counting gays and lesbians (and same-sex couples) is alway difficult because it requires self-identification which is then skewed to people who feel they don’t have any ramifications from self-identifying.

    Your point about DuPont is a good one because you don’t see many kids around there. The same-sex couples with kids that I know in the DC area live on Capitol Hill, in Takoma Park, in Arlington, in Columbia.

  • http://theaccidentalanglican.typepad.com Deborah

    Okay, maybe I’m slow on the uptake, but does part of the statistical problem lie with Lithwick’s assumption that the 6-14 million gay-parented kids all have two parents? The Lambda study says that 6-14 million is the number of children who have “a” gay parent. Couldn’t that number just as easily be counting single-parent households as well as dual-parent households? And what does Lambda do with joint-custody situations where one natural or adoptive parent is gay and the other straight?

    Or did my cursory reading miss something? Somebody correct me, please.

    In any event, Lithwick’s parsing of the data appears to have a “lies, damned lies and statistics” veneer to it.

  • http://onlinefaith.blogspot.com C. Wingate

    I have no idea what the actual numbers are, but an extremely rough calculation suggests that the percentage of households with children isn’t higher than 30% and probably a lot closer to 20%. Isn’t that a bit close to the numbers being claimed?

  • Micah Weedman

    mollie said:
    “So if gay couples are caring for up to 14 million children, as Lathwick states, that means that the 165,840 gay households with children in America care for an average of 36 to 84 kids.”

    deborah replies:
    “but does part of the statistical problem lie with Lithwick’s assumption that the 6-14 million gay-parented kids all have two parents?”

    but, lithwick actually says:
    “but the Lambda Legal Defense Fund estimates that 6 million to 10 million gay parents are caring for 6 million to 14 million children in this country.”

    Lithwick, in fact, never states or implies that 6-14 million gay couples have or care forchildren, but that 6-14 children have a parent who is gay and who cares for them, which she does by citing a research piece. This is a pretty significant diffence in verbage and would seem to indicate that Mollie’s own math is based on her misreading of Lithwick’s own use of data. Perhaps Dahlia’s not the only one being a little overzealous.

  • Micah Weedman

    I should have said “…but that 6-14 million children have a gay parent who cares for them…”

  • http://www.getreligion.org Mollie

    Micah,

    I’ll refer you back to the penultimate paragraph of my piece:

    “Now perhaps she meant to include gay households such as the ones I mentioned earlier — which would completely change her argument since her point is “Try as you may to vote or legislate your way out of a country that solemnizes such relationships, committed gay couples are already giving birth to, adopting, and fostering children.” Also, from both a judicial and social analysis standpoint, caring for children and caring for children occasionally are two very different things.”

  • Micah Weedman

    sorry, posting before editing, apparently:

    anyway, I should have said “lithwick never states or implies that gay couples care for 6-14 million children…”

  • Micah Weedman

    Mollie, your math still depends on lithwick’s assertion that 6-14 million children are cared for by gay couples. otherwise, why take the time to do this long division:

    “Okay, so if 22 percent of male-male households have children, that’s 66,220 male gay couples with children they care for. And if 34 percent of female-female households have children, that’s 99,620 female couples that have children they care for. Same-sex households that aren’t gay are also included here, I should note, but let’s just leave that out of the equation.

    So if gay couples are caring for up to 14 million children, as Lathwick states, that means that the 165,840 gay households with children in America care for an average of 36 to 84 kids.”

    The assertion that you base this math on, the one you attribute to Lithwick, doesn’t exist. In fact, nobody said that gay couples are caring for up to 14 million children, at least no one that is being engaged here. Lithwick simply quoted a source saying that that many children have a gay parent who cares for them. The assertion that doesn’t exist is a key component of your math.

    So, your penultimate paragraph doesn’t solve the problem, because Lithwick’s argument isn’t based on your math, so it doesn’t change. She never argued anything on the basis of gay couples caring for 6-14 million children, but that 6-14 million kids have a gay parent. I just think that, mathematically, that’s a pretty significant difference.

  • http://www.getreligion.org Mollie

    Micah,
    Well, then, we could look at it another way, I guess.

    She’s saying that gay parents are raising children all over the country and that 14 million kids have a gay parent. But the census data she links to says only 165,840 gay households have children in them. So even if we use a generous 3 kids per gay household figure, that means that gay parents are not taking care of 13.5 million of their offspring — which raises the social question of whether gay parents are very good at taking care of their children.

    Is that what you meant to say?

  • william james

    A+G+E+N+D+A= agenda, do the math people!

  • Micah Weedman

    I’m certainly not saying anything about gay children, nor is Lithwick, I think.

    But there certainly is a difference between having a gay parent and living in a gay household–I think this is key here. Having a straight father and a lesbian mother counts as having a gay parent, but not necessarily as living in a gay household. Lithwick doesn’t bother to distinguish between “having a gay parent” and “living in a gay household” because it is not crucial to her argument (which is in part that gay parents already have and care for children in a variety of situations). However, it is crucial to your math. For your math to make sense, you’ll have to show how, statistically, having a gay parent/caregiver and living in a “gay household” are the same thing for Lithwick’s argument. I don’t think you can do that though, becuase the census data is refering to households that have two unmarried adults in them and does not take into account the variety of other ways a child might be said to have a gay children.
    So, your assertion that gay parents are statistically bad parents seems a little unfounded at this point.

  • http://www.getreligion.org Mollie

    Micah,

    I fixed that typo about “gay children” in my previous comment already and hoped no one saw it! But I was too late.

    Anyway, I KNOW that she’s mixing data here — one set of which is so on-it’s-face ridiculous and made up as to not merit inclusion.

    But that’s why I said in the second-to-last paragraph that if she meant to include households like the two I mentioned (where a gay parent has children but is not the primary caregiver) that that undercuts her argument.

    No one questions whether homosexuals CAN impregnate or be impregnated. The whole story is about whether gay parenting is this huge social phenomenon.

    Her data do not bear it out.

    Let’s just go back to the simple thing of wondering whether it seems feasible that 14 million children have a gay parent.

    Do you think it’s feasible that almost 20 percent of children today have a gay parent?

    I mean, gays tend to have FEWER children — for hopefully obvious reasons — than straight people.

    It’s actually somewhat difficult to get pregnant or impregnate someone if you’re gay. So the notion that 1 out of 5 children in America was parented by a gay person just should give us pause . . .

  • Micah Weedman

    Mollie–sorry about the gay children thing–should’ve guessed that was a type-o.

    I just don’t buy what your after, becuase I don’t think that you are reading Lithwick’s piece correctly (which is, admittedly, most fully about family law in America, which makes sense given that Lithwick is a Slate legal commentator on a regular basis, as well as showing up at NPR and other places commenting on legal issues in American culture).

    You state this:
    “Anyway, I KNOW that she’s mixing data here — one set of which is so on-it’s-face ridiculous and made up as to not merit inclusion.”
    But which is so farsical–the information from the 2000 Census, or the information on gay adoption from the US’s Department of Health and Human Services? (she may be citing the Lambda Legal Defense folks, but the link goes to HHS) Actually, pointing to HHS was, rhetorically, a pretty deft move on Lithwick’s part, since her ultimate point in the piece is that the machinations of family law and services have, in fact, skipped caring about gay marriage at all but are developing a legal tradition that deals with gay parenting issues. Case in point–well, just look at what HHS itself has to say about gay adoption and parenting!

    Anyway, your antidote to Lithwick’s zealotry seems to be to disprove that there are as many gay parents as she claims there are. My point is that she isn’t making the claims you accuse her of making, or at least originally accused her of making. She cites what would seem to this layperson to be fairly reasonable sources, that is, two branches of the guvm’t that, you know, track things, (while offering up the dislaimer herself that this stuff is hard to track).

    Anyway, do I think its possible that 1 in 5 kids has a gay parent? Sure, but its irrelevent what I think, becuase I’m not an expert. What I do know is that there is some data that would suggest that, while admitting that its hard to know ultimately. Am I willing to believe that 1 in 5 kids lives in a gay household? Not so much, but then, I’m certain no one has asked me to believe that.

  • http://www.getreligion.org Mollie

    Micah,

    I feel a bit uncomfortable spending more time discussing this data issue with someone who has no trouble believing that one in five children in America has a gay parent.

    I mean, we all know that gays comprise at most 5 percent of the population and that their sexual activity does not lead to procreation.

    I guess all I can say is that I encourage you to be more skeptical of the numbers that journalists use to bolster their arguments.

    And otherwise, I think we have irreconcilable views about the goodness of the data sets used by Lithwick to make her case that gay parenting is commonplace.

  • Micah Weedman

    Mollie,
    The conflict here is that you used numbers to bolster your own argument based on something you said Lithwick asserted in her article. I was skeptical of not only your numbers but also your reading of Lithwick’s argument and demonstrated where, in fact, she did not make the assertion you were basing your numerical argument on in the first place. You are free to make the critique that there is something between the lines in Lithwick’s piece, but you have not done so. My original post, and all subsequent ones, have dealt with nothing other than the contents of your original post and the article to which it responded.

    You are certainly welcome to move on to bigger and better things, and I for one am certainly not interested in discussion about whether gays procreate or are suitable parents, but at least one reader here is skeptical of the numbers this journalist used to bolster her own arguments.

  • njones

    … the real number to debate would something else from Lithwick’s link.

    The very next sentence after the 6-14 million number you all have been debating estimates that between “8-10 million children are being raised in gay and lesbian households.” The citation goes to an book published by the editors of the Harvard Law Review. This is probably in the form of a 95% confidence interval so we can use 9 million as a working guestimate.

    If there are 78.1 million children under 18 and 9 million of them are being raised in same sex households, how probable is it that 11.5% of the children under 18 are being raised in households of a group that makes up between 1-10% of the country. If you take the higher estimates of the number of gays (those closer to 10%) then this is quite likely. If you take the lower estimates (those closer to 1%) then this is very unlikely.

    The real question to ask before we can evaluate these numbers is how many gays and lesbians are there?

    -n

  • http://www.getreligion.org Mollie

    NJones –

    For your “smell test” on the numbers you mention, you might ask yourself whether the procreation habits of homosexuals are any different than for heterosexuals . . .

    I mean, let’s just be super generous and use numbers that only Al Kinsey could approve and come up with a ten percent gay population rate . . . and let’s lowball the number of children with supposedly gay parents at 10 percent.

    Okay, now I mean, it could just be the gay people I know but two gay men or two gay women engaged in regular every-day sexual contact don’t normally conceive. Again, that could just be 100 percent of the gay people I know and not representative of the entire gay population.

    But is anything kind of making you pause here? Anything you know about, say, the mechanics of homosexual sex?

    OR IS IT JUST ME?

    Also, that Harvard Law Review article I have not read and would like to check — although I seem to recall the publication date was 15 years ago . . .

    Finally, being gay and operating a gay household — the latter being that phenomenon that Lithwick is trumpeting — are two totally different things. By way of example, I’m heterosexual but I do not live in or operate a heterosexual household. When I get married later this year, that will change.

    So for your point to work in context of Lithwick’s “gay households are everywhere” argument, this would mean that 100 percent of gay people operate or live in gay households. In other words, for you to answer the question “how probable is it that 11.5% of the children under 18 are being raised in households of a group that makes up between 1-10% of the country,” that group would have to only exist in gay households not as single gays.

    And again, I could be wrong, but I’m pretty sure most gay people don’t live in gay households.

  • Michael

    By way of example, I’m heterosexual but I do not live in or operate a heterosexual household. When I get married later this year, that will change.

    Of course you do, unless you assume that sexual orientation only has an impact when you are partnered. The mere fact you are planning a wedding to a man means you live in a heterosexual household. It just happens to be a heterosuxal household of one.

    Households can include just one individual.

  • http://www.getreligion.org Mollie

    Michael — I don’t live by myself. I have housemates (and won’t go into detail about their sexual lives).

    When the government surveys our house — and they did once — they considered us a household.

  • Sarah

    Michael-

    Try this math problem. Assume that each of the 165,840 gay or lesbian households with children has an average of 2.3 kids. If there are 6 million children being raised by a gay or lesbian parent, that means that only 94% of these kids are being raised by a single parent.

  • Michael

    You are putting much too much faith in the census numbers. I believe the number of gay and lesbian households is actually much higher, maybe as much as five times higher. It was a disasterous question that yielded questionable results. No other subset of the population is as likely not to self-report than gays and lesbians.

    While I think the 6M number is probably high, it is probably more reliable than the Census data.

  • Micah Weedman

    The American Academy of Pediatrics has said that the majority of children with gay parents are the products of heterosexual marriages that ended in divorce (presumably with one member coming out of the closet before or after the divorce.) Mollie is right, of course, that gay sex doesn’t procreate. However, Mollie seems to ignore the possibilty that gay men or lesbians having heterosexual sex with members of the opposite sex might lead to procreation.

    Mollie, this may not fit with your anecdotal trip to ol’ Gay D-C, but if anecdotes mean anything, it fits quite nicely with my own(limited)experience. Over this past weekend, I worked a retail establishment in conjunction with the Men’s Big Ten conference tournament. I worked with a grand total of 8 different people. Of those eight, three were openly gay (that I know of). One fits your stereotype of DuPont Circle gay culture–in fact, this person used to work at a same-company-owned retail establishment in, you guessed it, DuPont Circle. The other two are lesbians, and both are involved in same-sex households that are primary caregivers for children. Person A lives with her partner who helps her care for Person A’s two biological children, who were conceived and born in Person’s A previous heterosexual marriage. I don’t know what the custody arrangements are between (lesbian) mother and father, but I know that mother and her partner provide the majority of care. Person B lives in a long standing committed relationship with her partner. Person B helps to care for biological children produced in her partner’s previous heterosexual marriage, in addition to foster children unwanted by their heterosexual parents. Person B was also previously involved in a heterosexual marriage, but as far as I know, does not have biological offspring.
    So, your repeated assertions that gays don’t procreate does not match up with my own anecdotal experience. And remember, I said Big Ten tournament, not ACC, Big East, Ivy League or Pac Ten. I live in middle ‘Merica, full of pick-ups and soy beans, in a town whose paper still quotes Holy Scripture on its masthead.

    I think ther are plenty of ways to go after Lithwick, if you want to. She misquotes and mislinks her source on the number of gay persons caring for children (though as I mentioned above, the mislink is effective to her argument; besides, we all know that this may have as much to do with copyediting as with agenda). Then there is, of course, engaging the content of her argument, which is that the legal system is making provisions for gay parenting, which she barely manages to argue, since part of her article deals with those state trying to prevent gay adoptions. The numbers, which she herself admits to being hard to measure, just do not seem to be important to her argument, which is that gay parenting has arrived on the family law scene. I’d be interested to know if her representation of the legal situation is trustworthy or not.

  • http://www.getreligion.org Mollie

    Micah,

    You lost me after your first paragraph where you said, “Mollie is right, of course, that gay sex doesn’t procreate. However, Mollie seems to ignore the possibilty that gay men or lesbians having heterosexual sex with members of the opposite sex might lead to procreation.”

    Did you even read my post? Because I seem to recall writing that both gay men with children that I know were previously married.

    In case you missed it, here is what I wrote:

    “One of my close gay friends is in a relationship in which he helps care for his partner’s children every other weekend. Another friend and former colleague moved with his partner to Ohio to be near his offspring. In both cases the men with children were previously married and seem to be very loving fathers. In neither case do the men serve as primary caregivers.”

    Yep, I’m clearly ignoring this possibility.

  • http://www.getreligion.org Mollie

    As Barbie once said, “Math is hard.” But that’s no excuse for this treatment of data. Like it or not, census data are considered some of the most reliable data on planet earth. So if you’re going to dismiss census data, you might as well dismiss all data. In fact, because of its relative accuracy, you would have to dismiss all other data — if you care about statistical integrity at all.

    The reason why census data are considered so reliable are because of everything from the large sample size to the multi-pronged approach to asking the same question (which balances out incorrect answers or attempts around answering or answering correctly). Now, if you want to believe that it makes MORE sense that one in five children are born to gay parents (sigh) than that census data are pretty accurate, well, then, again, I honestly don’t think this conversation will be fruitful.

    Anyway, since Dahlia Lithwick based her entire argument on how prevalent the practice of gay people raising children was, and yet failed to prove how prevalent the practice was, that’s what this post was about.

  • Michael

    The reason why census data are considered so reliable are because of everything from the large sample size to the multi-pronged approach to asking the same question (which balances out incorrect answers or attempts around answering or answering correctly).

    Of course, none of that applies to a brand-new question that doesn’t even say the word gay and lesbian and lacks duplicator questions and no way to know whether there is statistical or histgorical undercount. Thus the PR campaign within the gay community to get people to understand the oblique question.

  • Micah Weedman

    Mollie,

    My fault on forgetting your post about your gay friends who have children. It seemed like when you said this:
    “I mean, let’s just be super generous and use numbers that only Al Kinsey could approve and come up with a ten percent gay population rate . . . and let’s lowball the number of children with supposedly gay parents at 10 percent.

    Okay, now I mean, it could just be the gay people I know but two gay men or two gay women engaged in regular every-day sexual contact don’t normally conceive. Again, that could just be 100 percent of the gay people I know and not representative of the entire gay population.

    But is anything kind of making you pause here? Anything you know about, say, the mechanics of homosexual sex?

    OR IS IT JUST ME?”

    you seemed to be suggesting that homosexual sexual mechanics are the reason there can’t be that many kids with gay parents.

    At any rate, I share your thoughts on the fruitfulness of this discussion. However, I do need to say that never did I indicate that I believed any statistics, but that I could imagine a possible scenario–by way of indicating that I don’t personally, or really theologically, care about the statistics. They are irrelevent.

    But my posting is about your final statement:
    “Anyway, since Dahlia Lithwick based her entire argument on how prevalent the practice of gay people raising children was, and yet failed to prove how prevalent the practice was, that’s what this post was about.”
    My point in posting is not to enact any agenda to promote gay marriage (which does mean that I do/do not have said agenda). My agenda here revolves around what I think is a misreading of Lithwick’s article. Her argument is *not* based on the prevelence of gay parenting—the prevelence of gay parenting IS her argument, and her argument is based *not* on statistics–which she herself admits are troublesome–but rather on the cases that are appearing in US family courts struggling with gay parenting. Furthermore, she is clearly suggesting that family law in America is more suited to gay parenting than it is gay marriage. She has an agenda in suggesting that “liberal” judges who rule in favor of gay parents/families are protected from conservative criticism by the rule of family law. This strikes me as the interesting point of entry into her argument, theologically and culturally, especially for conservative culture warriors.

    At any rate, I am happy to admit that the statistics may be important, and I just may not be interested in them. Either way, my original point in posting was not to dispute or confirm anyone’s numbers, but to point out that Lithwick at no time states or asserts that 6-14m gay couples are caring for children, which was the basis for your whole mathematical argument.


CLOSE | X

HIDE | X