hmmm, not a good experiment- give people a mean and ask them to guess the distribution - it is not surprising they can't do it - they are asked to guess Anscombe's quartet from the mean. it is almost a deception experiment - since there are an infinite number of possible answers and no right answer

I’d say it is not a bad experiment at all for the question asked (disclaimer: I read the article above, not the original study). As I understand it, the point of the study is to disprove the „common view“ that bar charts are a good/reliable visualization in almost all use cases. It shows, how…