An earlier version of this was posted 5 days ago at https://news.ycombinator.com/item?id=15162530 . My comment from then still applies. I wrote the following three paragraphs.

This is a lightweight comment which draws from one published paper, a blog, and a NYT opinion piece. [This new version has 2 more references, but not to any research.]

Its sense of incredulity is no further advanced than the opposition to pair programming in the 1990s, when it was one of the practices promoted by XP. It does not, as I hoped, tie together the decades of research studies on the topic.

For a much more detailed summary of the topic, much more along the lines of what I wanted to read here, see the chapter by Laurie Williams on the topic in "Making Software: What Really Works, and Why We Believe It".