The study itself.

The basic shape of the program

The Neubauer twin study was not, in its day, called a “twin study” in public. It was presented to the families involved as an ongoing “child development study” — a routine longitudinal observation conducted by the same agency that had handled the adoption. There was no description of the variable that made the study what it was: that the agency was deliberately placing identical siblings into separate households, withholding from those households the fact that the child was a twin or triplet, and then comparing the children’s development against one another.

That comparison was the point. The researchers framed the situation as a unique chance to observe identical genetic material developing under different family conditions — the closest thing, in their telling, to a controlled experiment in nature versus nurture.

For the families, it was something else entirely. Adoptive parents who had been told their child was an only child were periodically visited by researchers who asked questions, administered tests, and filmed the child. Many parents, in interviews given decades later, describe finding the visits intrusive but trusting that the agency knew what it was doing. Some thought the attention was unusually thorough. None of them knew their child had been split from a sibling in another home across town.

Who and where

The study was based at the Child Development Center in Manhattan, an arm of the Jewish Board of Guardians (later renamed the Jewish Board of Family and Children’s Services). It was led by Peter B. Neubauer, an Austrian-born psychoanalyst who had directed the Center since 1951. Its intellectual co-architect was Viola W. Bernard, a psychiatrist at Columbia who had served for years as the consulting psychiatrist to Louise Wise Services, the Jewish adoption agency that supplied the subject children.

Bernard had advocated for the separation of twins at adoption on independent grounds — she believed, as a matter of policy, that identical siblings developed more healthily apart than together. Whether that policy preceded the research interest or was made to serve it remains contested. What is clear is that by the time the study was operating, the adoption agency’s practice and the research program’s data needs were aligned.

What the data was

Researchers compiled what amounted to a parallel longitudinal record on each child: home visits, structured observation, intelligence and personality assessments, films of the child at structured tasks, and interviews with the adoptive parents. The data was organized in such a way that a researcher with access could compare twin to twin without either subject ever appearing in the other’s view.

None of this was published as a finished study. Neubauer never produced the synthesizing book or major paper the program seemed designed to support. He spoke and wrote about the underlying questions, but the dataset itself — the heart of the matter — remained inside the institution.

How it ended

The program quietly stopped in 1980. The reasons given have varied. By that point, the ethical climate around research on children had shifted: federal regulations governing informed consent had taken hold, institutional review boards had matured, and a study designed in 1960 was no longer one that 1980 would readily approve. The reunion of Bobby Shafran, Eddy Galland, and David Kellman in 1980 — accidental, public, and immediately newsworthy — also closed off the program’s most basic premise: that the subjects would not learn what they were part of.

Neubauer chose not to publish. In 1990, the records were given to Yale’s Manuscripts and Archives in Sterling Memorial Library, with an agreement that they remain sealed for decades — through October 25, 2065, on the terms of the deed of gift. The institution that funded and housed the study has, in the years since, made limited and partial disclosures to subjects who knew to ask.