In the United States, the role of geographical sorting as both a cause and/or consequence of polarization is often assumed, sometimes tested, and hotly debated. The starting point is that citizens’ views are affected by the political views available, dominant, or allowed in their immediate environment. More controversially, a “sorting” argument (Bishop 2012) supposes a growing geographic clustering of like-minded people (in terms of partisanship and ideology, but also regarding education, class, religion, race), hardening pre-existing dispositions. As a result, electoral districts are becoming less and less competitive, attitudes polarize, and inter-party resentment grows.
While debate continues over the size and impact of this dynamic in the US, even less is known about the role of geographical sorting in shaping attitudes and political preferences in European contexts. We study its role in the less likely case of the Netherlands. Are Dutch voters shaped by the socio-political makeup of their immediate environment? And more so in the most homogeneous areas? We use unique geocoded survey data, containing vote choices in 2017 and 2012 (without relying on recall) and a broad range of attitudes. We link this to census data and polling station results.
We expect, first, that in areas with substantial support for a party in 2012 more voters are likely to switch to the same party in the subsequent election. (Or, reversely, absence of support for a party might deter voters.) Second, we expect that in socio-cultural more homogeneous areas (in terms of e.g. education, ethnicity, income, but also ideology), vote choices became more extreme than in diverse areas which offer opportunities of cross-cutting contacts.
