Self-reported cheating in web surveys on political knowledge

Abstract

Measuring citizens’ political knowledge is important for understanding public opinion formation. In view of the increasing popularity of Web surveys, this paper examines the limitations of this interviewing facility when measuring factual political knowledge. We show that Web surveys contain a source of measurement error as respondents can “Google” the correct answers. This cheating activity is our principal concern. Past efforts are extended by: (1) offering a self-reported estimate of the share of Googling cheaters, (2) showing that the positive effect of education on factual political knowledge is most probably underestimated when cheating occurs, and (3) showing that self-reported cheating activity is inversely related to actual response time. In the concluding section, we discuss the implications of these results and the extent to which cheating can be reduced. The empirical analyses are based on a Danish Web sample from 2012 (N(Formula presented.) (Formula presented.)1,509).