Interestingly, there's already some research about how bad people are at generating random numbers, so if the author tests that at the follow-up, I'd expect him

Interestingly, there's already some research about how bad people are at generating random numbers, so if the author tests that at the follow-up, I'd expect him to be both repetitive and predictable. I haven't looked at the paper in a couple decades, but iirc, if asked to pick a random number between 1 and 20, almost a third of the answers will be either 7 or 17.