We’ve been watching with interest an unfolding flap about a controversial 2003 paper in the Archives of Sexual Behavior (ASB) by a prominent mental health researcher, Robert Spitzer, which suggested that gays could be deprogrammed by so-called “reparative therapy” to change their sexual orientation.
Spitzer, who was instrumental in the effort to extradite homosexuality from the realm of mental illnesses, apparently had developed serious doubts about the validity of his paper, which has been cited 47 times, according to Thomson Scientific’s Web of Knowledge. His regrets came to light recently in a piece by Gabriel Arana, of The American Prospect, detailing his own unfortunate experience with “ex-gay” therapy. In his article, Arana says Spitzer requested that he
print a retraction of his 2001 study, “so I don’t have to worry about it anymore”?
(Of note: Well-known sex researchers William Masters and Virginia Johnson also claimed to have “converted” 12 gay and women of homosexuality, but those cases have also been questioned.)
We think a better choice of words than “retraction” in the Spitzer case might be “repudiation,” since Arana obviously would not be in a position to retract anything Spitzer wrote, unless of course he’d published it in The American Prospect.
And it doesn’t look like the ASB has any intention of doing so, either. Northwestern’s Alice Dreger, who blogs for Psychology Today, spoke with ASB editor Ken Zucker, who said he told Spitzer that controversy alone did not make a paper retraction-worthy — a sentiment with which we heartily agree:
A few months ago, Zucker told me, Spitzer had called Zucker wanting to talk about the latest DSM revision. During that call, according to Zucker, Spitzer “made some reference to regretting having done or publishing the study, and he said he wanted to retract it. My recollection of the conversation was something like this: I said, ‘I’m not sure what you want to retract, Bob. You didn’t falsify the data. You didn’t commit egregious statistical errors in analyzing the data. You didn’t make up the data. There were various commentaries on your paper, some positive, some negative, some in between. So the only thing that you seem to want to retract is your interpretation of the data, and lots of people have already criticized you for interpretation, methodological issues, etc.’”
Zucker went on: “Did he ask me whether, if he submitted a letter to the editor, I would say no? No. I didn’t say no, I didn’t say yes. I basically think that, in the conversation, I was pushing back in terms of what exactly he wanted to say.” In other words, Zucker was trying to get Spitzer to articulate exactly what he wanted to say now, publicly, about his 2003 article. “And that was the end of the conversation. Now had Spitzer a week later submitted a letter to the editor saying ‘I no longer agree with my own interpretations of the data,’ would I have published it? Of course. Why not?”
We think the questions of whether and when researchers should be able to call do-over, as Spitzer is trying to do here, are complicated. To be sure, discovery of a fatal flaw in an experiment — ordering the wrong mice, for example — is a no-brainer. But does questionable interpretation of the data meet the threshold of retractability? And if not, how should the record be corrected, so that anyone who finds the original paper also finds the doubts and questions? Finally, does that threshold vary by field?
We look forward to a robust discussion.