Are European Parliament (EP) elections still second-order? In this article, we test the classical model at the individual level in contrast to an alternative ‘Europe matters’ model, by investigating the relative importance of domestic vs. European Union (EU)-related issues among voter-level determinants of aggregate second-order effects, that is, individual party change. We do so by relying on an original, CAWI pre-electoral survey featuring a distinctively large (30) number of both domestic and EU-related, positional and valence issues, with issue attitudes measured according to the innovative ICCP scheme (De Sio and Lachat 2020) which includes issue positions, issue priorities and respondents' assessment of party credibility on both positional and valence goals. Leveraging the concept of ‘normal vote’, we estimate multivariate models of electoral defections from normal voting separately for general and European elections, based on issue party credibility. This allows us to assess: (a) the distinctiveness of the two electoral arenas in terms of issue content; and (b) the relative impact of EU-related and domestic issues on defections of Italian voters. Our findings show that although second-order effects are still relevant in accounting for results in EP elections, vote choice in the latter is also partly due to specific effects of certain policy issues, including some related to the European dimension. This indicates that EP elections have their own political content, for which Europe matters even after controlling for the importance that EU-related issues have acquired in national elections.