Why Bother Lying When You Know So Few Care? Party Contact, Education and Over‐reporting Voter Turnout in Different Types of Elections

Studies on individual‐level voter turnout in the European Parliament elections rely solely on self‐reported turnout data. At the same time, a long tradition in public opinion research examines the impact of individual and contextual variables on over‐reporting of voter turnout. The ultimate goal of these efforts is an assessment of the bias introduced when turnout models are estimated from self‐reported turnout data. In this article, it is proposed that certain characteristics related to turnout (and turnout over‐reporting), like university education or party contact, should also be positively associated with political awareness and knowledge. If so, they might contribute to respondents' greater ability to distinguish between different salience levels, significantly increasing non‐voters' propensity to over‐report turnout in high‐salience elections, but not in low‐salience ones. This hypothesis is tested using data on electoral participation in Sweden, comparing patterns of turnout over‐reporting in the (high‐salience) national parliament elections and the (low‐salience) European Parliament elections. The results of this test, the first one to give an account of patterns of over‐reporting of turnout in the European Parliament elections, largely support the above hypothesis. Finally, the consequences of this phenomenon for the validity of inferences made from self‐reported turnout data are analyzed.