An Incompetent Research Project

Why a Notorious Twin Study Was Never Published

Posted Feb 09, 2019

Few phenomena are more inherently fascinating to the public and also to researchers than twins separated at birth and later reunited. The public is interested for ethical reasons (separating twins strikes many as inherently wrong) and for human interest reasons (finding one’s long-lost first love is inherently heart-warming). Researchers are interested because twin separation, especially when the twins are identical (and thus have the same genetic makeup), is a powerful method for studying the relative contribution to human development of nature versus nurture. Typically, the separation occurs as a result of an adoption agency or physician finding they can get more money by selling the babies separately (sorry to be such a cynic) or some other inadvertence, such as a divorcing couple dividing up their property equally. There is also a guiding belief (misplaced in my opinion) among some adoption specialists that it is better for the parents, and ultimately the children, to deal with only one child.   

     Recently, while watching the CNN documentary “Three Identical Strangers” (about reunited triplets who had been separated at birth) I became aware of an abhorrent variation on this theme, in which a researcher (the late esteemed New York child psychoanalyst Peter Newbauer) directed the separation of four sets of identical twins and one set of identical triplets, as part of a longitudinal research project conducted in the 1960s and 1970s. Newbauer was a psychiatry professor at NYU, but also consulted with a Jewish adoption agency, and—with funding from NIMH—he arranged to have the newborn babies assigned to deliberately differing (e.g., affluent verus non-afflluent) adoptive parents. As a condition of the adoption the parents were told that their child was involved in a multi-year research project and they were strongly encouraged to allow recurring annual visits from researchers. None of the parents, or for that matter the children, were ever told the true purpose of the study or even, for that matter, that all the children in the study had a hidden identical sibling (or siblings) somewhere in the world. The multiple birth aspect was exposed when two of the triplets inadvertently ran into each other, but the results of the research was never published, and study files remain locked up and inaccessible to this day in Newbauer’s papers at Yale University. An aspect of this study that struck many as ironic is that Newbauer was an Austrian Jew, who escaped the Holocaust (he had been attending medical school in Switzerland at the onset of WWII and was able to stay there) and yet carried out a project that has some superficial qualities (e.g., playing God with subjects’ lives) in common with the notorious (and scientifically worthless) studies carried out by the Nazis on Jewish and non-Jewish concentration camp prisoners.  

Various theories abound as to why Newbauer never published his data. One possibility is that as a psychoanalyst Newbauer expected that nurture (parenting and social class) would explain most of the variance in child and adult outcomes, and was dismayed by how even in the most different of rearing households, the separated siblings exhibited many dramatic similarities (such as the triplets all becoming wrestlers and two separated female twins both majoring in film studies). Another hypothesis is that Newbauer came to appreciate the likelihood that the study would be viewed as unethical, and wished to avoid the resulting blow to his reputation. (A countervailing ethical obligation requires keeping one’s commitment to funders and participants). But there is a third theory that has not been espoused before, to my knowledge. This hypothesis has to do with the fact that the study may have been poorly designed and carried out and that Newbauer, as a psychoanalst trained to do single subject descriptive studies, lacked the quantitative or other research skills to make sense of his data.  

As the study was never published, I don’t know enough about it to be able to critique it in detail, but I do know that the number of subject dyads (and one triad) was far too small to be able to apply conventional statistical methods. Another possibility is that the data were never formulated in terms of discrete dependent variables, to which analytic methods could be applied. To clean up what was likely a messy data set would have required skills that Newbauer probably lacked. Finally, the independent variable (e.g., different characteristics of the adopting families) likely was not manipulated in a very consistent or clear-cut manner, for two reasons: (a) the children were all placed by the same small adoption agency and there were thus limits to Newbauer’s ability to find adoptive families that were sufficiently different from each other, and (b) one would need at the outset to have a clear and sophisticated understanding of what independent variables one wanted to manipulate, and this probably was lacking in a study that was put together hastily and without differentiated hypotheses.  

      As understanding about multiple birth children and their gestation has improved, we now know that some widely-held assumptions are incorrect (e.g., the DNA of identical twins are not completely identical). Another confound that even today is not fully appreciated is that there are non-genetic biological contributors (e.g., maternal prenatal alcohol consumption, use of forceps) to birth outcomes, and that even for twins or triplets the prenatal and perinatal experience can differ by such things as amount of nutrition and oxygen obtained; these could have affected various birth outcomes such as length, weight and brain intactness. Furthermore, even when rearing families seem very different on a gross level, there can be unappreciated similarities on a micro level. As example, there was the famous study in which identical twin brothers were separated just before WWII with one boy growing up in the British West Indies, and the other in Nazi Germany. They were later studied in their 30s at the University of Minnesota (ground zero for twin studies) and it was found to everyone’s amazement that they both flushed the toilet just before urinating. But I grew up during the same time period and during my toilet training I was encouraged to pretend that I was a pilot making a bombing run over Tokyo. This involved flushing the toilet first to give a challenging time window, and standing far enough back to account for wind drift. So this practice of first flushing the toilet could have reflected the twins attacking in wartime the nation of their separated biological parent  and thus may have had nothing to do with genetics. (Parenthetically, I continued the pre-flushing into my 50s when I gave it up dring a water shortage, but still miss the toilet on occasion, especially when I turn off the lights when making a nightime bombing raid).

     My point is not that genetics are unimportant but rather that applied research is hard to adequately pull off or interpret. I am sure that Peter Newbauer was a lovely man, who did much good for his field and for the children and families he worked with. But competence requires one to be cognizant of one’s limitations, and the separated twins study provides a warning about how a lack of such cognizance can result in incompetence and even charges of immorality.

Copyright Stephen Greenspan