I can't say I am surprised by the findings of this evaluation. The whole issue seems to be a classic example of scientific-political mutualism, in which (i) common sense-toppling research is produced by scientists, who overlook some details that go against their own preferences and/or overstate the significance of their findings. (ii) These isolated results are then packaged into a review, which further boosts their significance. (iii) At some point, the media get hold of the supposed new scientific truth and start reinforcing it, making it the new public "common sense". (iv) More reviews and popular articles appear, parroting the new consensus, while opposite views are ridiculed and suppressed with a loud "science says otherwise". But (v) continuing new science and critical thinking undermines the new common sense. The final step, yet to be realized in this case, (vi) is the collapse of the new consensus and the re-affirmation of something closer to the original common sense.
Psychol Rep. 2008 Aug;103(1):275-304.
Re-evaluation of the "no differences" hypothesis concerning gay and lesbian parenting as assessed in eight early (1979-1986) and four later (1997-1998) dissertations.
Academic and policy effects of eight early dissertations on gay and lesbian parenting are discussed with a focus on their having been cited at least 234 times in over 50 literature reviews, beginning with Gottman in 1989 and 1990. Most literature reviews, referencing these eight early dissertations and agreeing with Gottman's early conclusions, have reiterated the theme that parenting by gay men or lesbians has outcomes no different than parenting by heterosexual parents. Here it is proposed that certain potential adverse findings may have been obscured by suppressor effects which could have been evaluated had multivariate analyses been implemented. Further, several adverse findings were detected by reanalyzing data where sufficient information was yet available. Some of the dissertations' results (absent controls for social desirability and other differences between homosexual and heterosexual parents) supported the 2001 "no differences" hypothesis discussed by Stacey and Biblarz. Yet, differences were also observed, including some evidence in more recent dissertations, suggesting that parental sexual orientation might be associated with children's later sexual orientation and adult attachment style, among other outcomes. Odds ratios associated with some of the apparent effects were substantial in magnitude as well as statistically significant. Also, more recent research on gay and lesbian parenting continues to be flawed by many of the same limitations as previous research in this area of study, including overlooked suppressor effects.