Abstract
Famous schizophrenia genetic adoption studies were conducted decades ago in Denmark by investigators using four different research designs. The most often cited study, led by neuroscientist Seymour Kety, used the “Adoptees’ Family” design that began with adults who had been adopted away as children. These adoptees became the study’s first-identified relatives (probands). The researchers then assessed “schizophrenia spectrum disorders” among their biological and adoptive relatives in two consecutive population samples (the “Copenhagen” and “Provincial” studies). They published their results between 1968 and 1994. The author provides evidence that in the spring of 1967, Kety and colleagues discovered that their long-planned group comparison did not yield a statistically significant difference in the genetic direction. Instead of accepting and reporting these negative genetic findings, the evidence suggests that they switched to an unplanned group comparison that did yield a statistically significant difference. The author argues that in a 1996 publication and in his 1999 Lasker Award description of how he conducted his studies, Kety inadvertently revealed his neveracknowledged 1967 comparison-group-switching behavior. This practice is now known as phacking. The author describes additional ways in which the study was p-hacked. Although we can evaluate researchers and their methods in the context of the era in which they performed their research, we must evaluate their publications and conclusions based on today’s standards. The author goes on to argue Kety’s p-hacked study helped revive the then-fading psychiatric genetics field and represented a reaction to the “anti-psychiatry” movement of the late 1960s and 1970s.