How can you say it's easier for humans to reason about when your test subjects almost all exclusively got started in programming with imperative programming? It's a slightly biased data set.