Experimenting with a "Third Way" in Political Knowledge Estimation

Political knowledge is a key variable in the study of political behavior known to predict political discussion, voter turnout, issue-based voting, and support for democratic norms. Yet the appropriate method of measuring political knowledge is the subject of controversy. The traditional technique, which encourages survey respondents to say "don't know" (DK) if they are uncertain of the correct answer, has been called into question. Recent work has suggested that the DK response should be discouraged. We argue that a better design eliminates the DK option altogether. A web survey is employed utilizing an experimental design. Three different strategies for measuring political knowledge are tested: the first encourages the DK response; the second discourages the DK response; and the third omits the DK option. In its absence, nonrandom psychological factors are eliminated and estimates of political knowledge are higher in comparison to the other two strategies. Implications for measurement, methodology, and survey administration are discussed.