Spitzer Retracts his 2001 Paper – Kind Of…

As Warren Throckmorton (amongst others) reports, Dr Robert Spitzer has asked to retract a paper he authored in 2003 examining the anecdotal reporting of almost 200 people who had been through an “ex-gay” programme.

Knowing this article was coming, I talked last evening with Bob and asked him what he would like to do about his study. He confirmed to me that he has regret for what he now considers to be errant interpretations of the reports of his study participants. He told me that he had “second thoughts about his study” and he now believes “his conclusions don’t hold water.” He added that he now believes that the criticisms of the study expressed in the 2003 Archives of Sexual Behavior issue are “more true to the data” than his conclusions were.

He told me that he had expressed these thoughts to Ken Zucker, editor of the Archives of Sexual Behavior several months ago. He wondered aloud to Dr. Zucker if there was some obligation to say the critics were right and that the study should be withdrawn. Although Spitzer said he did not recall Zucker’s exact reply, he did not feel encouraged to withdraw the paper.

Of course, reading the full piece that includes this interview with Spitzer we can see exactly what Spitzer said.

I asked about the criticisms leveled at him. “In retrospect, I have to admit I think the critiques are largely correct,” he said. “The findings can be considered evidence for what those who have undergone ex-gay therapy say about it, but nothing more.” He said he spoke with the editor of the Archives of Sexual Behavior about writing a retraction, but the editor declined.

Now, read the bit in bold carefully. What Spitzer is saying is that a survey of anecdotes can provide you with anecdotes and nothing more. What Spitzer’s saying is this – his work from a decade ago is useful for knowing what those who went through ex-gay programmes thought about what they had done, but it can’t be used to determine whether such self-perceived outcomes are actually true. Put in other words, Spitzer’s survey does tell us what people who went through ex-gay programmes thought about it, but it provides no qualified analysis as to whether the claims of change actually occurred.

This article in the Archives of Sexual Behaviour highlights some of the original critiques of Spitzer’s paper. These are,

  • The sample was not a random sample of the actual population that had gone through such therapy. It was largely self-selecting and a group who sought change for religious reasons.
  • There was ambiguity as to what form the therapy actually took. Different clients were with different therapists who had differing models they were working with.
  • There was ambiguity as to what the result of such therapy actually was. Was it orientation change? Was it “happy celibacy”?

The critiques can be summarised by this quote.

As Spitzer noted, self-reports are unreliable measures. Freund (1960, 1977) and Conrad and Wincze (1976), for example, found that clients’ self-reports of favorable reorientation outcomes tended to be imprecise, deny certain facts, and were not supported by objective data. Spitzer seemed to believe that he has assessed sexual attraction; yet, he has only measured participants’ subjective experiences of their attractions.

Now what is the point of pointing this out? Well for a while I have been emphasising that anecdotal reports are not sufficient for validating ex-gay experience. What is really needed are proper academic studies like the Jones and Yarhouse Ex-Gay Study. There is also another reason for raising these issues, namely that there doesn’t exist a single report of “harm” from ex-gay therapy that doesn’t fall foul of the same kind of critique that was (rightly) levelled against the Spitzer paper. Harm reports tend to be anecdotal, taking no account of the type of therapy involved and have very poor qualitative qualifications of harm.

But it gets better. In a landmark paper in 2009, Professor Michael King produced a paper for BACP, a Systematic Review of Counselling for LGBT clients. He reported,

We found very few papers focusing on psychotherapy for transgender people that were not solely concerned with preparing them for gender reassignment or assisting them afterwards to adapt to their new gender role, but none had data suitable for inclusion in the review. Our main finding is that the standard of quantitative research into the benefits of psychotherapy for LGBT people is poor, and there is no published evaluation of its effectiveness in either experimental (randomised trials) or quasi-experimental (cohort of clients) research. Therefore, we could not conduct a meta-analysis of the quantitative data. The qualitative studies provided a rich literature on LGBT clients’ perceptions of therapy, the kinds of therapists they prefer, and how the attitudes, training and methods of therapists can determine outcome. However, many of the findings (eg need for therapist acceptance and positive regard) did not seem specific to LGBT issues and might be concerns for anyone entering psychotherapy.

  • When clients report experiences of so-called exgay affirmative therapy, they usually describe therapy in which homosexuality and bisexuality are regarded positively, prejudice is avoided, the stress of externalised and internalised anti-homosexual bias is recognised, and there is sensitivity to LGBT development, culture and lifestyles. It was a therapy that knows what it is to be LGBT.
  • There was little consistent use of standardised measures of outcome. However, therapists were mostly highly rated in terms of general satisfaction and perceived helpfulness of the therapy.
  • Most quantitative studies involved convenience samples of LGBT people, which means that selection bias is likely, and the generalisability of existing results to all LGBT people undergoing therapy is questionable.

What can we conclude from this? Quite simply, that anecdotal data isn’t normally valid for making useful qualitative assessments of therapy, but that goes across the board. Those who criticise “positive” research into ex-gay therapy tend to employ double-standards, for the same plumb-lines they measure such research by can be equally used to critique not only “research” into the “harm” of ex-gay therapy, but also gay-affirming therapies in general. As King’s 2009 paper rightly points out, there are real issues around self-selecting sample bias in most research in this field, and there are also concerns about how self-reporting anecdotal data is not reliable.

Critics of ex-gay therapy need to accept this inconsistency as much as proponents of such therapy need to rely only upon proper research methodology. For both camps, what’s sauce for the goose needs to be sauce for the gander.

Posted in ExGay Study, Statistics Tagged with: , , , , , , , , , , , , , , ,