Survey researchers and those who rely on survey research only accept answers of a certain sort; i.e., answers consistent with what they consider to be true or correct. Unwanted answers are rejected and efforts made to produce the sort desired. For instance, columnist Heidi Stevens (Chicago Tribune, 9/6/15, sec. 6, p. 3) didn’t like answers to a survey that asked parents two years after the birth of their first child: “How satisfied are you with your life, all things considered?” Too many, she wrote, reported a sharp decrease after they became parents. Based on her preconceived notions about these matters, that decrease in happiness after parenthood is incorrect and indicates that the right question is not being asked. Therefore, Heidi entreats: “stop asking whether parenthood makes parents happy”. This makes perfect sense to survey researchers and those who rely on survey research: when you don’t get the answers you want you must stop asking the question that produces the unwanted answers. Ask questions that produce desired answers which, in this case, means positive answers about parenthood. No problem, says Heidi. All you have to do is “ask parents whether parenthood makes them better people”. Then you’ll get “a truer [more positive] picture of what it means to parent”. See how easy it is? Reject answers you don’t like and ask questions that produce answers you do like.
For an extended discussion of survey researchers accepting only answers of a certain sort and other limitations of survey research, see my book, The Problem with Survey Research.