Was a Gay-Marriage Paper Fabricated?

A recent paper on same-sex marriage appears fabricated; an earlier, disparaged one on same-sex parents was not.

There’s been much ado the past couple of days about the possible retraction of Michael LaCour and Donald Green’s 2014 article in the journal Science on how attitudes about same-sex marriage can be changed.

The original study, labeled by one outlet as the biggest political-science study of the year, purported to reveal that when gay and straight canvassers attempted to encourage voters to support same-sex marriage, each successfully altered respondents’ attitudes, but only gay canvassers’ efforts appeared to persist over time. They also reported seeing sustained within-household transmission of opinion change, again only among the gay canvassers, not among the straight ones. It made a big splash. And Science is kind of a big deal as far as prestigious journals go.

The results reinforced gay-rights organizations’ interests in the personal-contact hypothesis — the idea that face time spent with “out group” individuals by those who disagree with them can reduce the latter’s negative perceptions and, in this case, change their minds about same-sex marriage. It’s certainly a plausible theory, and many anecdotes seem to reinforce it. But we will no longer know from this embattled study whether it’s as accurate as it appeared.

What prompted the concern about this study? Several things. First, attempts to replicate a portion of the study returned a dramatically lower response rate than the original. (This looks fishy.) Second, some of the numbers, including a distribution on a gay “feeling thermometer,” were found to be identical to ones in a different population-based dataset, which ought to be different, if even just a little, from LaCour’s Los Angeles sample. Third, the company that purportedly collected the data has never heard of the study. (Ouch.) There are several other anomalies as well, enough to suggest the possibility of outright fabrication. It’s not the first time this has happened. Indeed, journals are becoming wary. The American Journal of Political Science recently began requiring its published authors to produce their replication files for public access, in order to confirm study validity.

The company that purportedly collected the data has never heard of the study. There are several other anomalies as well, enough to suggest the possibility of outright fabrication.

I know how it feels to be accused of scientific malfeasance and sampling and data manipulation. I do not, however, know what it feels like to actually be guilty of those things. And yet over at the New York Times, my 2012 studies have been opportunistically lumped in with Mr. LaCour’s in an effort to tag my New Family Structures Study as tainted data. (They are not.) In other words, if Mr. LaCour’s ship is sinking, why not conveniently attempt to drag that pesky NFSS along with it?

It’s the latest in a very long string of efforts to criticize the data, together with its sample, its author (and his friends), its funders, its measures, its analytic approach, its terminology, its data-collection organization, its reviewers, its journal’s editor, and its supporters. First one, then another, university inquisition has come to naught.

At bottom, I’ve supposedly been “debunked” because of this one thing: I chose to treat (unusual) maternal and paternal same-sex relationships comparably rather than “control away” the far more numerous unstable ones in favor of focusing only on the (handful of) cases that reported household security. Too few, I held, to assess them separately with ample statistical power. Others dissented — with pitchforks in hand — claiming I’m somehow trying to make same-sex households look bad. That’s the story.

But unlike the sinking LaCour study, we’re still in scientific territory with the NFSS. And you can learn from the latter if you’re willing to wrestle with the social reality of households in which parents exhibited same-sex relationships in the era that I studied (the growing-up years of adults who were ages 18 to 39 in 2011). The reality was messy. Perhaps they’re neater and more stable today. (It’s an empirical question.)

But the financial barriers to adoption or assisted reproductive technology that are the lot of same-sex couples will always pose a higher bar to the acquisition of children, and with it a more selective and wealthy sample of those who become parents, as opposed to those (many) who already were parents prior to entering their first formal same-sex union. Step-parenting, dubbed “social parenting” among the LGBT community, remains unavoidable, complicating any notions of a “fair” comparison that might be made with planned or unplanned children born to their biological mother and father. Complicated stuff.

#related#American household life is messy, and can only be truly understood if social scientists first describe it before moving on to the more subjective judgment that comprises their explanations of how that social reality came to be. My study was descriptive, and never claimed to be something else. Others privileged jumping to explanation. Since I really do want to know what’s going on, I think both can be valuable, but the second without the first is a fool’s errand.

I recently wondered aloud, in a Public Discourse article describing the costs of leaping past description to explanation, whether “now that the Supreme Court’s oral arguments are behind us, and the justices have already privately cast their votes about the future (and the history) of marriage, perhaps it’s possible that the social science of marriage, sexuality, and child outcomes can catch its breath” and “operate without the pressure cooker of politically-acceptable narratives.”

It’s safe to say we’re not there yet.

— Mark Regnerus is associate professor of sociology at the University of Texas at Austin and a senior fellow of the Austin Institute for the Study of Family and Culture.