Study claiming prejudice leads to early death for gay people retracted
A coding error rendered findings null, and efforts to fix things have ended. Andrew Masterson reports.
Prejudice against sexual minorities is, prima facie, a terrible thing, but a highly cited journal paper claiming that prejudice alone can shorten the lives of gay people by more than a decade has been retracted after the key conclusion was found to be the result of a coding error.
The paper, Structural stigma and all-cause mortality in sexual minority populations, was written by a team led by Mark Hatzenbuehler of Columbia University in the US and published in 2014 in the journal Social Science and Medicine.
The researchers set out to determine whether “living in communities with high levels of anti-gay prejudice increases risk of premature mortality for sexual minorities”.
They found that for people belonging to these groups within intolerant communities suicide, homicide and cardiovascular disease rates were significantly above the norm. “This result translates into a shorter life expectancy of approximately 12 years,” they concluded.
The paper attracted a lot of attention from both media and academia, not all of it positive. In 2016, the academic watchdog site RetractionWatch reported that a controversial researcher who had previously testified against same-sex marriage, Mark Regenerus of the University of Texas, attempted to replicate the study’s findings and was unable to match the result.
Regenerus demanded that the original study be retracted, but Hatzenbuehler and the journal held firm. Instead, Hatzenbuehler asked a colleague of his, another Columbia researcher called Katherien Keyes, to conduct a review of his method and findings.
Keyes eventually came back with bad news. A coding error in the way the original data had been processed had introduced major mistakes into the findings. With the error fixed, the results came back null: they did not show any effect on mortality rates among gay people living in highly prejudiced communities.
In 2017, Hatzenbuehler asked for, and was granted, a correction notice from the journal editors. The paper remained in print. It was a response which surprised many in the field, given that the central finding of the research had completely vanished.
However, the journal editors revealed that the researchers had promised new data, which in time would be subjected to peer review.
Whether or not that ever happened is unclear. In a new post, RetractionWatch reveals that on 7 February this year, without fanfare, the paper was finally and formally retracted. Neither the researchers nor the journal editors have commented.