Many people who he categorized as having been raised by a gay or lesbian parent had spent very little time with that parent or with his or her same-sex partner. Even Regnerus admitted that his data included only two people who said they had been raised for their entire childhoods by a same-sex couple. -And some of the data were...questionable.
By eliminating suspect data — for example, a 25-year-old respondent who claimed to be 7’8” tall, 88 pounds, married 8 times and with 8 children, and another who reported having been arrested at age 1 — and correcting what they view as Regnerus’ methodological errors, Cheng and Powell found that Regnerus’ conclusions were so “fragile” that his data could just as easily show that children raised by gay and lesbian parents don’t face negative adult outcomes.I guess some respondents were just messin' with him.
Cheng and Powell determined that of the 236 respondents whom Regnerus had identified as having been raised by a lesbian mother or gay father, one-tenth had never even lived with the parent in question and an additional one-sixth hadn’t lived with that parent for more than one year. Still more had provided inconsistent or unreliable responses to survey questions, throwing their reliability into doubt. That means, Powell says, that over one-third of the 236 people whom Regnerus classified as having been raised by a lesbian mother or gay father “should absolutely not have ever been considered by Regnerus in this study.”
Reanalyzing Regnerus’ data after eliminating respondents who offered dubious biographical information and recategorizing people who clearly were not raised by gay parents, Cheng and Powell found only three statistically significant differences between the respondents raised by a lesbian mother and those who reported having been raised in “intact biological family” households. Only one of those differences could be considered a negative adult outcome — those respondents were more likely to have had an affair while married or cohabitating. Even that is hardly a smoking gun, says Powell: “If you study 40 different variables or outcomes…just by the law of chance, a few of them should be statistically significant.”
Cheng said that in taking on “one of the most controversial articles published in the history of social science research,” they tried to stay away from the debate about Regnerus’ ideology or the source of his funding. “What we can do is analyze the data,” he said.