> Doesn’t sound very reliable, especially being from 1986 when we generally had a much less sophisticated understanding of the intricacies of statistics and hypothesis testing.

I just wanted to note - statistical analysis is not a new or groundbreaking thing. A lot of these tests have been around for decades or centuries[1] and experimental design ditto. Honestly, it being from so long ago actually gives me a bit more faith in the results since gaming statistics to get a click-bait headline and more funding has become a cottage industry in recent years.

1. https://en.wikipedia.org/wiki/Student%27s_t-test#History - enjoy the amusing connection to Guinness, it's a favorite with new stats students.