Well-conducted surveys of social data, including sexuality, yield invaluable information.
There is good news in Tuesday’s release from the CDC’s National Health Information Survey. No, it’s not the revelation that the gay and lesbian population may be slightly smaller — 1.6 percent — than many scholars have believed; population-based surveys tend to fluctuate. The results of the report make sense to me, and since others have already described them, I will refrain from repeating them at length here.
The good news is that the survey’s questions were administered to over 34,000 Americans, randomly sampled, enabling scholars to get quality information about even small communities. And since the NHIS didn’t broadcast its interests or commitments to its survey pool (unlike some studies), we are treated to valid data — on sexuality as well as many other subjects — largely devoid of a form of social-desirability bias wherein respondents’ awareness of their own participation in a study whose topic is dear to them affects their survey answers.
This was not always so. Survey questions about sex and sexuality have a colorful political history, especially when Uncle Sam picks up the tab. My entry into the study of human sexual behavior — which turned me neither left nor secular — came by way of the National Longitudinal Study of Adolescent Health, which was born at the University of North Carolina (where I earned my Ph.D.). Professors there had been seeking federal funding for a national study of sexual behavior during adolescence. Although the details of the saga are known to few, the gist of the circulated narrative is that the original study was canned with help from the state’s senior senator, Jesse Helms, who felt the American public needed protection from a government that was bold enough to snoop into its citizens’ bedrooms. Rather than capitulate and move on, the principal investigators did what seemed most sensible to them — they added health questions to the study of sexual behavior — and thus was born the well-regarded “Add Health” study, now on its fourth wave. The extra “d” in its nickname (it should be just “Ad,” for “Adolescent”) is there on purpose, a subtle jab at the system that sought to abort its birth.
I owe my career, in part, to that study, whose complexity and thoroughness taught me that high-quality survey data — carefully collected — is a treasure, even if the populations it illuminates couldn’t care less, or actively oppose the light. Good information is just that, and scholars and citizens of all stripes value knowing the truth about Americans as it can best be discerned.
Of course, a quality sample and a good set of questions is just a start. Moving from survey construction to data collection to analysis to interpretation multiplies the opportunities for ill will and intellectual contest. That’s understandable; humans are the most challenging subjects to understand, and their behavior the most difficult to predict with accuracy. Thus surveys have limitations; they’re not poised to go deep into human motivation. But they’re good at description.
Toward that end, then, is America’s LGBT population really only 1 to 2 percent, as the new NCIS data suggest? Or is it 2 to 4 percent, as most other data sources — including two large population-based surveys I conducted — have asserted? I’m not sure it matters a great deal. The more important point is that data collection done carefully with appropriate (random) sampling strategies generates estimates far closer to the real numbers than will sloppy surveys and dubious samples, the results from which are now erroneously embedded into the American imagination. Indeed, a recent Gallup foray into public attitudes revealed that Americans think that one in four of their fellow citizens is either gay or lesbian (to say nothing of bisexuals, etc.). Even Alfred Kinsey’s maligned “quota sampling accompanied by opportunistic collection” approach elicited only an off-base 10 percent guesstimate. But today, no more than four out of 100 Americans think that the real number — which the NCIS estimates is 1.6 percent — is below 5 percent. Methinks Will & Grace has done more than just prime the country for same-sex marriage.
So, yes, the government is in your bedroom, figuratively (and indirectly) speaking. But information-gathering about Americans’ attitudes and behaviors — when obtained from their own mouths and with their consent and confidentiality — is okay. From social studies like the Add Health and the NCIS we can learn, clarify, and correct. Both sides of the aisle ought to continue to support ample federal funding of general data-collection projects like these, because a citizenry that lacks good data about itself — or worse, doesn’t care — is willingly choosing ignorance over illumination.
— Mark Regnerus is an associate professor of sociology at the University of Texas at Austin, a research associate of its Population Research Center, and a senior fellow of the Austin Institute for the Study of Family and Culture.