Talk:User expectations

From Wikipedia, the free encyclopedia

Suppose I sample data on the wages of workers, compute an estimate and confidence interval, and then decide I'd like to publish it. But first I must ask, how much error is the data user willing to tolerate before rejecting the published data as non-useful, or deceptive?

Has anybody ever studied this? Has anybody ever surveyed data users, and tried to scientifically find their "pain threshold", so to speak, in these regards?

I would like to decide based on soun science, not personal preference.