Project leader

James Dennison

Project start and end date

July 2019-December 2021

Funding Source 

Swedish Research Council


During the 21st century, the party systems of European democracies have undergone transformations on a larger scale than at any other time in the post-war period. Most notably, populist right parties have either built on earlier middling successes in countries such as France, Italy and Austria or come from nowhere to disrupt the party system in the likes of Sweden, Germany and the United Kingdom.  

New, populist left-wing parties have made similar gains in countries such as Spain and Greece, while other centrist and green challenger parties have enjoyed intermittent success. Meanwhile, all across the continent, the established centre-right and, particularly, centre-left parties have seen fairly consistent declines. These changes continue to have major political and policy ramifications, in some cases affecting the integrity of the European Union and questioning both the liberal, rights-based consensus and free-market economic consensus that had seemed victorious in the aftermath of the Cold War.

In spite of the obvious importance of these changes for the future of European societies, there remain no commonly accepted academic explanations for them. Existing academic work has tended to focus on one party family at a time or changes in just one country, despite there being impressively consistent pan-European trends. Moreover, most existing explanations for voting behaviour use variables such as voter socio-demographics, voter political attitudes or, more fundamentally, the socio-economic structure of society, all of which we know are far too stable and slow-changing to be effective explanations for the volatility in Europe’s recent electoral results. Other, external, proposed causes such as immigration rates or policy changes necessarily are, alone, insufficient because we know that changing voter behaviour requires as a prerequisite some form of cognitive change amongst voters, which remains unidentified.
This project takes a number of different approaches. The project’s primary hypothesis is that these trends result not from Europeans having dramatic changes of heart ideologically but from variation in what they see as the most important political issues of the day, also known as issue salience. These issues could include the economy, unemployment, health, the environment, education, immigration, Europe, amongst many others. The project will theorise and test the relationships between the ‘salience’ of certain issues and the electoral success of corresponding party families across Europe. For example, it may be that populist radical right parties do particularly well when a high percentage of the electorate see immigration as ‘salient’, even though few members of the public have become more in favour of or opposed to immigration. The project will first map what Europeans see as the most important issues in all EU28 member states—using data that asks representative samples ‘what do you think are the most important issues affecting your country today?’—and show how this has changed throughout the 21st century. It will then use advanced statistical models to identify whether changes in issue salience is the cause of the dramatic changes to Europe’s party systems, or whether it is some other potential explanation.

The project will then identify what causes variation in issue salience itself, using such explanatory variables as real-world trends, events and the actions of media and politicians. For example, do voters start to see immigration as very important because of it being mentioned in the media more, or does the media mention immigration more because voters already consider it more important. This will point us to the more fundamental causes of electoral change. Finally, it will consider when and where across Europe these salience-based explanations work and where they do not, and why, by considering so-called ‘interaction effects’ and using combined models that take each stage of causality, from real-world events, to media and politician actions, to voter salience, to voting behaviour, into account. By using cutting-edge panel data models and structural equation modelling to adduce causality, the project is likely to contribute significant advances to our scientific understanding of not just voting behaviour but the causes of human behaviour and cognition moreover, while solving some of the most important political science puzzles in Europe today.