Can Overreporting in Surveys be Reduced? Evidence from Israel's Municipal Elections

Studies over the last 50 years have consistently shown that survey respondents overreport their participation in American national elections (Abelson, Loftus & Greenwald, 1992; Clausen, 1968; Parry & Crossley, 1950; Presser, Traugott & Traugott, 1990; Silver, Anderson & Abramson, 1986; Traugott & Katosh, 1979). Similar results were reported in Britain (Swaddle & Heath, 1989), Sweden (Granberg & Holmberg, 1991), New Zealand (Karp & Banducci, 1999), and Norway (Waldahl & Aardal, 2000). Overreporting is a widespread democratic phenomenon. For example, ANES data reveal that between 13% and 41% of nonvoters report that they have voted in national elections when in fact they have not (Belli, Traugott, & Beckmann, 2001). Karp and Brockington find that overreporting among nonvoters ranges from 18.7% in Swedish regional elections to 26.2% in national elections. In Britain, the national numbers are similar, somewhat lower than in Norway (35%), U.S. presidential elections (40.1%), and in New Zealand (42.4%) (Karp & Brockington, 2005).