Survey respondents answer less consistently when anonymous than when identified

Evidence

Degree of Anonymity. 1 first noted the effect of anonymity in a 1974 experiment (RWA, p. 244) that found students who served anonymously answered with less internal consistency than did those who put their names and booklet numbers on an attendance sheet for “administrative purposes.” The latter group in turn proved less consistent than students who put their names right on the booklets. These findings have been replicated several times (see RWA, p. 326). … If you suspect students care less, and become careless when they know their answers cannot be linked to them, I agree with you. But I have always been leery of the way internal consistency sparkles when subjects put their names directly on the materials; social desirability motivations and a wish to look consistent might be pumped up too much. — Bob Altemeyer, The Authoritarian Specter, 866