In the political war over ex-gay conversion therapy, sexually conflicted people get trampled.

The 200 people included in Spitzer’s study weren’t picked at random. They were referred to him by the very advocates whose work and ideology he was testing. Forty-three percent learned about the study through ex-gay religious groups. Twenty-three percent learned about it through NARTH. Nine percent were recruited by their former therapists. The rest, by and large, were referred by conversion therapists or other subjects in the study. It’s hard to conceive of a more biased sample. And there’s reason to believe that the hidden denominator—the number of patients who didn’t turn out well and weren’t referred—was enormous. Spitzer reported:

To recruit the 200 participants, it was necessary to repeatedly send notices of the study over a 16-month period to a large number of participants who had undergone some form of reparative therapy. This suggests that the marked change in sexual orientation reported by almost all of the study subjects may be a rare or uncommon outcome of reparative therapy.

In a conversation with Arana two months ago, Spitzer used Nicolosi to illustrate his suspicions: “In all the years of doing ex-gay therapy, you’d think Nicolosi would have been able to provide more success stories. He only sent me nine patients.”

By several measures, Spitzer's subjects were wildly unrepresentative of the gay population. Ninety-three percent said religion was “extremely” or “very” important to them. Seventy-eight percent had spoken publicly for efforts to cure homosexuality. Such faithful advocates may have been driven by an exceptionally strong will to believe that their orientation could be, and was being, changed. In fact, the study explicitly excluded failures. “To be accepted into the study,” Spitzer explained in his report, participants had to demonstrate after therapy “a change of at least 10 points, lasting at least 5 years, toward the heterosexual end of the scale of sexual attraction.”

None of the things being measured—how gay the subjects had been before their therapy, what kind of therapy they got, how straight they were afterward—was clearly resolved. “Given the fallibility of memory for past events,” Spitzer wrote, “it is impossible to be sure how accurate individuals were in answering questions about how they felt during the year before starting the therapy, which on average was about 12 years before the interview.” The nature of the therapy was never defined, and in half the cases it was administered by a pastoral counselor, support group, Bible study, or other religious practice. As for subjects’ claims that they had stayed gay-free after therapy, Spitzer confessed that he “relied exclusively on self-report” and that this reliance “deserves careful examination in light of the participants’ … high motivation to provide data supporting the value of efforts to change sexual orientation.”

Within these limits, the net result was as scientific as Spitzer could make it—which is to say, not very. The “data set” he produced looked more like a collection of unverified anecdotes from people eager to vindicate conversion therapy. The widespread changes in attraction reported by this group didn’t match the broader experience of Alan Chambers, president of Exodus International—the world’s leading ex-gay ministry—who admitted this year that, regardless of behavioral change, “99.9 percent” of people he had met in the ex-gay world  had “not experienced a change in their orientation” to the point of escaping same-sex attraction. Notorious cases of behavioral relapse include the chairman of Exodus and the founder of “National Coming Out of Homosexuality Day.” When the American Psychological Association reviewed the available research in 2009, it concluded:

There are no studies of adequate scientific rigor to conclude whether or not recent [therapies] do or do not work to change a person’s sexual orientation. Scientifically rigorous older work … found that sexual orientation (i.e., erotic attractions and sexual arousal oriented to one sex or the other, or both) was unlikely to change due to efforts designed for this purpose. Some individuals appeared to learn how to ignore or limit their attractions. However, this was much less likely to be true for people whose sexual attractions were initially limited to people of the same sex. … [S]ome individuals modified their sexual orientation identity (i.e., group membership and affiliation), behavior, and values … They did so in a variety of ways and with varied and unpredictable outcomes, some of which were temporary … Based on the available data, additional claims about the meaning of those outcomes are scientifically unsupported.

The gist of the literature is that a few people might manage to change their feelings, and some can change their behavior, but most fail. In his report on the 2001 study, Spitzer argued that even a partial change in attraction after therapy should be acknowledged as significant, as it would be in evaluating a drug. But the FDA doesn’t approve drugs that work only in an unquantified, poorly defined subpopulation, based on self-reporting, when such drugs are also reported to have harmed other patients. Look what happened to Arana. After his conversion therapy, he writes, “I spent hours in front of the window of my third-story room, wondering whether jumping would kill or merely paralyze me.”

Spitzer’s study never was the smoking gun it was cracked up to be. It didn’t substantiate conversion therapy as a cure, or even as a safe treatment, for homosexuals in general. By the same token, his apology doesn’t warrant the therapy’s eradication. What he showed us, anecdotally, is that an unusual subset of highly motivated people can find ways to alter their sexual self-understanding and possibly their behavior. Those people have no grounds to say conversion therapy will work for the rest of us. And we have no grounds to say it can’t work for them.

William Saletan's latest short takes on the news, via Twitter: