Last week MARRI decried the Supreme Court’s ignoring of the data in the same-sex marriage case. This week we decry a related corruption in academia: the project of shutting down comparative research on same-sex couples, marriage, and parenting. The American Psychological Association, The American Sociological Association, and the National Council of Family Relations are all participating in this project.
Academia and Marriage
Below we quote from the American College of Pediatricians’ amicus brief. This excerpt is one example of the type of research not found in the journals of these organizations. As a result, this research is not making it into the academic discourse of most social science departments in the United States. Below we include graphs from the amicus brief that give the reader some idea of what is being deliberately kept out of the academic discourse. While no data are the final words, the following information comes from datasets repeatedly used in research articles in journals published by the associations mentioned above.
There is a need for a phrase that labels this corruption of the social sciences in the United States – a nation founded on the concept of freedom. We suggest: “Academic Jacobinism”. Pass this research on to students and professors you know for they are unlikely to have been exposed to this body of research. (Underlining added.)
Despite being certified by almost all major social science scholarly associations—indeed, in part because of this—the alleged scientific consensus that having two parents of the same sex is innocuous for child well-being is almost wholly without basis. All but a handful of the studies cited in support draw on small, non-random samples which cannot be extrapolated to the same-sex population at large. This limitation is repeatedly acknowledged in scientific meetings and journals, but ignored when asserted as settled findings in public or judicial advocacy.
Of the several dozen extant studies on same-sex parenting in the past two decades, only eight have used a random sample large enough to find evidence of lower well-being for children with same-sex parents if it exists. Of these eight, the four most recent studies, by Dr. Mark Regnerus, Dr. Douglas Allen and two by Dr. Paul Sullins, report substantial and pertinent negative outcomes for children with same-sex parents. The four earlier studies, by Dr. Michael Rosenfeld and three by Dr. Jennifer Wainright and colleagues, find no differences for children with same-sex parents because, due to errors in file coding and analysis, a large portion of their samples actually consists of children with heterosexual parents. When the sample used by Wainright’s three studies is corrected of this error and re-analyzed, these data also show negative outcomes for children with same-sex parents similar to those reported by Regnerus and Sullins. More importantly, they also show substantially worse outcomes for children who have lived an average of ten years with same-sex parents who are married than for those who have lived only four years, on average, with unmarried same-sex parents.
The family arrangement envisioned in same-sex marriage is generally understood to be the condition of actually having two parents of the same sex, not opposite-sex parents who may or may not be in a same-sex relationship with someone outside the home. Instead of comparing heterosexual parents with same-sex parents, Wainright et al’s three studies compared a group of heterosexual parents with another group of (mostly) heterosexual parents. It is not surprising they found “no differences” in child outcomes between these groups, since they are, for the most part, the same group. The findings of these three studies do not apply at all to same-sex parenting and form no reasonable basis to conclude that children of same-sex parents are not disadvantaged.
Re-analysis of the Wainright studies data, after correcting the sample flaws, reveals that adolescents with married same-sex parents fare worse than those with unmarried same-sex parents. …
Comparing the married and unmarried same-sex parents with their opposite-sex counterparts, Sullins found that, while outcomes for children with opposite-sex parents improved if their parents were married, outcomes for children with same-sex parents were notably worse if their parents were married.
Bar charts below … illustrate the results. Asterisks by a number in the charts indicate that it can be inferred with confidence to the U.S. population of adolescents; the more asterisks, the greater the confidence.
You can read the entire brief here.
Americans no longer care about hurting children. As a popular song says "I don't care. I love it." It is not really about the facts anymore in our culture of death. (For example, children are willfully put to death by their parents through abortions.) Just suppress the facts… because 'I don't care.'
The best research selects a large sample and selects them using random techniques. That is Research 101. Yes there are some flaws but it is far superior to small samples that are selected at the researchers convenience (called a convenience sample).
Whether you agree or not, random sampling would better support what MARRI is saying.
Dear Dr. Cowan,
You have had a distinguished career (http://psychology.berkeley.edu/people/philip-cowan) at Berkley in the social sciences and have had many significant publications which have made a great contribution to family therapy along lines and approaches I myself used when I was a therapist decades ago. (See second last page of a recent presentation [http://downloads.frc.org/EF/EF15E114.pdf] at the UN). I hold you in great esteem for what you have accomplished at Berkley and the contributions you (and your wife Carolyn) have made through your clinical research.
To clear the ground before responding to your charge I want you to know that my world view does affect how I approach the social sciences. Is there any social scientist for whom this does not apply? But I learn the most when the social sciences present me with robust contrary data. So I treasure contrary data. Further, as an “attempting” Catholic I hold all gays as sons of God who deserve my good will and help in all ways that I can give. I have two close friends who are gay, one has been in a relationship for many years where he and his partner are raising the biological children of his partner. The other has lived with his lover for over a decade. They were my friends before they “came out” and remain so, though as Catholics they know the Church’s traditional teaching on these matters and know I hold to them.
However I also have many other friends and colleagues who, over the years, have sinned in other ways: divorcing an innocent spouse, attending Mass and receiving the Eucharist while contracepting (if sins have rank for Catholics this one is way up there and way beyond sodomy). I have my own grave faults. In other words I am a sinner surrounded by sinners and find an old Irish saying to be a good guide for me when I am tempted to judge others:
There is so much good in the worst of us
And so much bad in the best of us
It ill behooves the rest of us
To say ought but good about any of us.
Now that you have some inkling of the standards of judgment of others to which you may hold me I hope we can talk social science and methodology.
The mission of MARRI is to make the case for the nation and for individuals that the intact married family that worships God weekly is the surest road to personal, community and societal wellbeing along all dimensions measured in the social sciences (a tall claim but one that is holding up [http://marri.us/publications/mapping-america] in simple correlational terms). Will everyone take that path? Of course not. But I think it is a legitimate social science project, to remind people from the data of what is best for children. Given that you, can understand why I track comparative research at the macro level and I hope without cherry picking. Thus my interest in macro sociological data will be different from clinicians’ interest in small sample and even in opportunity sample research, for they have to serve the interests of clients from myriad backgrounds. This is the basis of my concern with the ASA in particular, but also with APA and NCFR for they use and publish data as well.
Hoping this will become an ongoing conversation about the search for truth as the social sciences goes about this task and to facilitate a good discourse with you, let us accept your critique of Regnerus (which I do not agree with, and neither does Paul Amato [http://www.baylorisr.org/wp-content/uploads/Amato.pdf], present president of NCFR, nor Loren Marks by implication but again for purposes of continuing this conversation let us accept it [http://www.sciencedirect.com/science/article/pii/S0049089X12000580]). Drop his research from the body of research mentioned in my blog above while keeping the rest, all of which is from large, national sample, government or government-funded surveys.
(Continued in next "reply")
My contention is that very significant research is being suppressed within the academy. That is my beef. And it is more than a beef, it is a very serious accusation and awful in its implications if it holds up to scrutiny. So I ask you to demolish it. Disprove me and I will be a happy man.
For instance Sullins critique of Wainright’s data and his reanalysis of that same data not only upends their claim but reverses it. Granted, the N involved is too small for the reliability needed for national generalization and could even be wrong (but that size N did not stop the American Sociological Association from show-casing Wainright’s work in their amicus brief). But now comes the rub: It is standard scientific-journal practice to publish, if not the new evidence in article form, at least to publish a letter from the author of the new study when such corrections are made. But Sullins never even received an acknowledgment of his letter from the journals involved in publishing Wainright’s research, much less an offer to publish the contesting analysis of the same data. I contend (but would love to be proven wrong) that this is suppression by denial, by ignoring. Such defense mechanisms are to be expected in the therapist’s office but not from the APA, NCFR and ASA. That is the core of my concern.
Now let us move on to Sullins two articles using the National Health Interview Survey. In these federal government, very large data sets, lie significant data, not exploited before, on same sex parenting, both for self-identified-married homosexual parents and non-married. Sullins analysis of the relationship between different types of parent configurations is flatly contradictory of the amicus brief contentions.
Yet another study, Douglas Allen’s analysis of Canadian census data is momentous in its implications. High school graduations are up to 35 percent lower for children from such unions. You just don’t get bigger datasets than a census, and this from Canada which has a much longer history of same sex civil partnering, civil unions and gay/lesbian marriage than we do in the States.
The silence of APA, ASA and NCFR to Allen’s and Sullins research is deafening. Where is the scientific debate here? Who is guarding the guardians?
Lastly, let me take up the challenge you presented about the repeated opportunity sample research you contend is dispositive because of its repeated same findings of “no difference” between gay parenting and always-intact married parenting.
I agree there is a lot to be learned from such research but it is at the “intimate” level, not the macro or sociological level. It is valuable in building knowledge but it is far from being dispositive.
If gay parents can raise children so that there are no differences observable between their children and those raised by always intact married heterosexual couples then there has to be something very significant for the rest of us to learn from them (a potential line of research for your doctoral candidates). Even if their results are closer to those of step families (which I think more likely to be the case) there will still be much to be learned.
(Continued in next "reply")
However, sociologically, most of the studies upon which ASA relies on in its amicus brief are based on “opportunity samples”, samples built by non-randomized methods. (That would have been more forgivable coming from APA but not ASA). To use an analogy — for the lay audience to our debate — this type of sampling is akin to seeking out only the members of the National Rifle Association to know what Americans in general think about gun control. Thousands of such opportunity sample studies would be trumped by one good study from a large national randomized sample. Most would not be surprised to find very different results from a national sample of Americans vs from a group of enthusiastic gun-owners. Were the topic anything other than the present one of gay parenting and gay marriage the ASA would be the first and loudest voice pointing out this fundamental methodological flaw. Is there a single demographer who would defend such method as giving a picture of nationally representative outcomes?
Should you prefer to continue this on a personal level through email or phone conversation I would be honored to continue it that way for I really appreciate your willingness to have begun this debate and, as I said in the beginning, I hold your work in great esteem. However I hope we can continue it publicly for it is a public issue and you are a worthy representative of the differing opinion.
I hope we can continue this conversation.
Pat Fagan, Ph.D.
Director,
MARRI