Homosexual parents: testing "common sense"--a literature review emphasizing the Golombok and Tasker longitudinal study of lesbians' children

Counter to claims by the American Psychological Association and the National Association of Social Workers as well as numerous reviewers that children raised by homosexuals and married heterosexuals do not differ, the elaborate social-personality theory called "common sense" predicts that because "like produces like" and because psychopathy/sociopathy informs the major expressions of social deviance including homosexuality, children of homosexuals will (1) be more frequently subjected to parental instability (of residence and sexual partners) and (2) have poorer peer and adult relationships. Also, as is held to be true of their parents, homosexuals' children will be more apt to (3) become homosexual, (4) be unstable (have emotional problems and difficulty forming lasting bonds) with reduced interest in natality, and (5) be sexually precocious and promiscuous. Differences between homosexual and heterosexual comparison groups that bore on "common sense" were considered suggestive "bits" of empirical evidence. Differences that emerged within studies conducted by sympathetic researchers utilizing volunteer samples were considered bits of adverse evidence. Of 171 bits, 82 adverse and 55 nonadverse bits supported, while 34 bits fell against "common sense." From this tentative method of counting, support was found for common sense beliefs that children of homosexuals will be more apt to become homosexual and have poorer peer relationships, while weaker support was found for some of the other predictions. As assessed in this way, the empirical evidence in the literature tended to lean against claims of "no differences" between children raised by homosexuals and heterosexuals. In particular, the strongly worded official claims of there being "no differences" are overstatements. They amount to the organizations and some prominent researchers asserting that they have proven the null hypothesis, which is fundamentally impossible. It is likely that the nonsignificant statistical findings stressed thus far include Type Two errors created by use of volunteer samples, inadequate identification and measurement of likely differences, and refusal to interpret results in ways contrary to the sympathies of subjects, investigators, and the organizations.