This study has some serious methodological issues: http://chronicle.com/blogs/percolator/controversial-gay-parenting-study-is-severely-flawed-journals-audit-finds/30255
As a specific example:
"Because of how the paper was written, Sherkat said, it would have been easy to miss Regnerus’s explanation of who qualified as “lesbian mothers†and “gay fathers.†If a reviewer were to skip ahead to the statistics in the table, it would be understandable, he said, to assume that the children described there were, in fact, raised by a gay or lesbian couple for a significant portion of their childhoods.
In reality, only two respondents lived with a lesbian couple for their entire childhoods, and most did not live with lesbian or gay parents for long periods, if at all."
It appears that he was comparing statements from children whose parent had a string of relationships, including at least one same-sex relationship, to children from intact two-parent families. I would not be AT ALL surprised to find negative outcomes from that. As far as I know, his data set has yet to be released, but I have been following this with interest since I saw the rebuttal blog in the Chronicle.