This is esp. the case with the kind of data generally available in religious studies, data about religious practices & beliefs that people have, etc. Survey data & self reporting -- while useful, & often all we have -- has to be carefully examined & cautiously used.
Q.v., from a paper by two University of Chicago academics:
"simple manipulations can a affect how people process and interpret questions ... people attempt to provide answers consistent with the ones they have already given in the survey."The whole thing is worth a read, esp. as this kind of data is used all the time in establishing public opinion, legitimating positions, etc.
"the social nature of the survey procedure also appears to play a big role in shaping answers to subjective questioning. Respondents want to avoid looking bad in front of the interviewer."
"the most devastating problem with subjective questions, however, is the possibility that attitudes may not exist in a coherent form ... measured attitudes are quite unstable over time."