Thread by @jayjoseph22

Thomas Bouchard just published a response to my 2022 critical evaluation of his 1990 Minnesota #IQ twin study. Here are the full-text links for my 2022 article and Bouchard's 2023 response. I will respond at a later point.

karger.com/hde/article/66…

cambridge.org/core/journals/…

In response to my charge that the Mx software used in his MISTRA twin study was designed to produce genetic results, Bouchard argued that Mx "is not MISTRA software." That's like being on trial for stealing a BMW and seeking acquittal on the grounds that it was actually a Lexus.

Bouchard didn't deny that DZAs were the official MISTRA control group, that DZA r's were omitted in the 1990 Science article, that full-sample DZA IQ r's were never published, and that what Nancy Segal called the "important first step" of determining that IQ rMZA > rDZA FAILED.

Bouchard used about one-quarter of his article attempting to show that the fifteen environmental influences shared by reared-apart MZ twins (MZAs) I listed in my Table 3 do not lead to MZA behavioral similarity. This is puzzling, since in the 1990 MISTRA Science article...

...Bouchard recognized, "The proximal cause of most psychological variance probably involves learning through experience, just as radical environmentalists have always believed." (He then erroneously counted these environmental influences as genetic.)

In my article, I produced this table supporting my argument that the MISTRA IQ study FAILED at the first stage of #IQ heritability determination, and that the actual MISTRA finding was 0% IQ heritability. Bouchard did not dispute my numbers, or provide any new ones.

Bouchard spent a quarter of his article arguing that earlier critics Kamin, Taylor, and Farber (KTF) engaged in "pseudo-analysis." He objected to their analyses of MZ-apart subsamples to show the importance of the environment, implying that their other astute observations...

...were invalid, and that everything KTF wrote should be ignored. Bouchard's argument has no relevance whatsoever to my analysis of the MISTRA and problems with reared-apart twin studies in general because, as he well knew, I did not engage in KTF-style subsample analyses.

As in 1990, Bouchard said that his sample of 30 DZA pairs was too small to be included in the Science IQ paper. However, N. Segal revealed in her 2012 book that the researchers submitted an early 1980s paper to Science with IQ results, including a smaller sample of 12 DZA pairs.

Bouchard said that I failed to provide "any examples of p-hacking in MISTRA" (p. 4). In fact, I listed two specific examples of MISTRA p-hacking, plus an additional probable p-hacking example.

Bouchard wrote that the 30 control group DZA pairs in the 1990 IQ study "were not included because the sample was small." He did not mention that I cited non-IQ MISTRA studies where he published full-sample DZA correlations based on similar or lower DZA sample sizes.