Dodgy social science in the Utah same-sex marriage case

graduationIn their recent submissions to a federal court judge who’s being asked to mandate same-sex marriage in Utah, the plaintiffs included a declaration from a sociologist, Charlotte Patterson, who commonly weighs in on litigation with the message that mothers and fathers are interchangeable. The problem, of course, is that there are now studies with larger sample sizes and employing better methodologies than have been conducted in the past which indicate there are real differences in outcomes for children raised by married mothers and fathers compared to those raised by same-sex couples.

For instance, a very recent Canadian study that suggested poorer educational outcomes for the latter. The plaintiffs’ expert tries to minimize this study.

We asked the author of the Canadian study, Dr. Douglas Allen of Simon Fraser University, for his response to the treatment of his work in the “expert” affidavit. Read on for his response:

In a recent declaration in Kitchen v. Herbert, Charlotte Patterson has stated the following:

“A new study by Allen (“High school graduation rates among children of same-sex households”, in Review Econ. Household, 2013), presents data that also purport to discredit the consensus, but also fail to do so. This study examined high school graduation rates among young adults in Canada, and reported that those who described at least one parent as gay or lesbian had lower rates of high school graduation than did their peers with continuously married heterosexual parents. This comparison is not relevant here, however, because almost all the parents characterized as gay or lesbian had also undergone divorce or separation, which is a known correlate of school problems, whereas other parents were more likely to be continuously married. Thus, the main conclusion that would appear to be warranted on the basis of Allen’s data is that young adults whose biological parents divorce or separate are also likely to report lower educational attainment than their peers with continuously married parents. This conclusion is already well-established, and it has nothing to do with parental sexual orientation. ”

There are several errors in this statement.

First, the 2006 Canada census has couples identify their type of relationship.  Hence, the paper I wrote uses data from gay and lesbian couples only. It is not the case that “at least one parent [is] gay or lesbian”.

Second, it is not the case that “almost all the parents characterized as gay or lesbian had also undergone divorce or separation.”  A large number had, but not “almost all.”

Third, and most importantly, the conclusion is incorrect.  She states that because the parents were divorced this explains the low graduation rates.   However, one of the contributions of my paper was to control for marital status (as opposed to Rosenfeld (2010) who did not).  Hence, if a parent was divorced or separated, this was controlled for in the estimations.  A reader who looks at columns (2) vs (3) in Table 5, or columns (1) v. (2) or columns (3) v. (4) in Table 6 of the paper, is seeing the effect of controlling for parental divorce and separation.  It is true, that not controlling for marital history reduces the odds of graduation for children of same sex couples, but when marital status is controlled for the odds of graduation of these children is still low.  The paper only discusses results based on the full control regressions.

Hence, the claim made by Prof. Patterson is false, and one has to wonder if she actually read the paper.

This entry was posted in Family, Gay Rights and tagged , , , . Bookmark the permalink.