A gauge of the extent of cheating in survey research is the extensive efforts made by survey researchers to counter it. Consider, for example, the following description (edited with added emphasis) of a webinar presented by a university survey research center:
“Preventing and Detecting Interviewer Falsification
Interviewer falsification is an unfortunate reality of survey research. Falsification involves the interviewer’s intentional deviation from the study protocol, and can include fabricating all or part of an interview, or deviating from recruitment and consent protocols. This webinar will not only provide an overview of how falsification can be detected, but also how to prevent it through proper training and oversight. This webinar is useful for any investigator who is supervising research assistants to collect data, or who hires an external survey research organization to collect data.”
Falsification occurs when askers fill in blanks on questionnaires (unit falsification) and when they make up responses for for entire questionnaires, interviews, or polls (falsify entire instrument). Askers who falsify answers are sometimes called “falsifiers” and, in other instances, “curbstoners”; i.e., sitting on the curbstone and completing asking instruments with made-up information.
How much cheating in survey research? Probably a lot and, most likely, since the beginning of survey research in the 1930s, which is also the time when there began to appear articles about how to build check-ups and cheater traps into polling procedures. Cheating “plague[s] polls”, says Sarah Igo, and the same can be said for other asking efforts. However, a precise amount cannot be determined because this aspect of survey research has not been systematically investigated.
For a further discussion of cheating in survey research, see my book, The Problem with Survey Research.