the quality of data is always imperfect, and the information gathered from surveys given to thousands of people will have more errors in it than data from closely supervised clinical trials. Surveys also depend on the truthfulness of the respondents.
