Negative partisanship has been on the rise in the last decades, not only in the USA but also in many other democracies (e.g.,Iyengar et al. 2019). Previous research has demonstrated the considerable impact of negative partisanship, independently from positive party identification, for political attitudes, democratic dissatisfaction, and behaviour in the USA and worldwide (e.g. Bankert, 2020; Mayer, 2017; Ridge, 2020).
Even though the concept of negative partisanship has become a highly salient topic, its current measurements are far from optimal: they do not capture negative partisanship as an identity (mirroring positive party identification measures, as they should). Most measures stay on the level of partisan attitudes, which hinders simultaneous analyses with affective polarization that are often operationalized using the very same variables or are not feasible for multi-party systems (e.g. Bankert, 2020; Caruana et al., 2015; Mayer, 2017; Ridge, 2021), making difficult to understand its role vis-à-vis affective polarization.
We develop and validate a measurement instrument for negative partisanship in multi-party systems that allow an adequate identity measurement. Starting from previous studies, we propose new measures. We validate those with a) expert interviews (n=7), b) cognitive interviews (n=30, combining probing and think aloud responses) and c) a large-n online survey. To ensure a broad test, we use Germany and Italy as countries for validation: Both share similar levels of AP, a pronounced left-right divide but are highly different when it comes to the stability of the party system which allows to see if we can capture old and new repulsions.