So are you saying that no response data ever should be published, or people should be forced to take surveys and what not?

Getting good survey data requires that the subjects agree to take a survey before you tell them what will be surveyed. This still biases respondents, obviously, but it puts all surveys on an equal playing field.