Mapping citizens’ preferences on European integration in a multidimensional political space

Previous research suggests that citizens at the extremes of the left-right ideological spectrum are more likely to oppose European integration. Based on a new study, Dmitri Tochkov illustrates that when citizens are placed in a multidimensional political space, the relationship between Euroscepticism and the other dimensions varies in more complex ways.

Euroscepticism is a malaise of the far right and the extreme left. Citizens with moderate political ideologies tend to support European integration, unlike those on the margins. Euroscepticism goes hand in hand with conservative (social) political attitudes, such as restricting immigration and limiting gay rights. Such statements reflect much of the received wisdom about how preferences for European integration should relate to citizens’ broader views on politics and public policy. But is the received wisdom still valid?

In one new study, we take a fresh look at received wisdom, and we find a much more complex picture. How preferences for European integration relate to other dimensions of political ideology differs depending on the country studied and how we measure preferences and ideology. Moreover, Euroscepticism is not just another aspect of social conservatism. Finally, citizens’ attitudes towards the EU and its policies do not form a coherent whole. Let’s unpack these results and explain their meaning.

For our analyses, we use data from Voting Advice Applications (VAAs) – the online tools you use to express your preferences on around 30 political issues, choose positions on dimensions of political ideology (such as left-right or conservatism-liberalism) and get a “match” with the political parties that most closely match your profile of preference. We analyze VAA data for the Netherlands, Germany, Italy and France from the period leading up to the 2019 European Parliament elections.

The advantage of this data is that tens of thousands of people filled out the VAAs, so we have a lot of records. This allows us to construct a high-dimensional picture of the political space in each country at high resolution and with good coverage even of the peripheral corners of that space. This is usually not possible with standard public opinion polls. The downside of VAA data is that the people using it are not a representative sample of the country’s population, so we don’t know how VAA user results would generalize. We come back to this limitation at the end of this article.

Let us first see how support for the EU varies across positions on the left-right and liberal-conservative dimensions of political ideology. Imagine that these two dimensions define a grid and that each citizen is placed on this grid according to their scores (for example a “4” on the left-right dimension and a “6” on the liberal-conservative dimension). Now we can calculate the average value of EU support for all citizens located at each point of this two-dimensional space. When we plot these average values ​​of EU support on the vertical axis, we get something like the image in Figure 1, which shows how EU support varies for all combinations of left-right positions and conservative-liberal in the Netherlands, as of 2019 (both panels show different rotations of the same 3D plot). As you can see, the resulting surface is quite complex!

Figure 1: The Dutch political landscape based on citizens’ positions on three dimensions of political ideology (2019)

To note: For more information, see the author’s accompanying article (co-authored with André Krouwel) in European Union Politics.

It turns out that for social conservatives, EU support diminishes as we move from the socio-economic left to the right. But for social progressives, EU support peaks at moderate positions, and not quite on the far left. For leftists, EU support is only slowly increasing as we move from the conservative corner to the progressive corner. But for right-handers, the climb is steeper, and the highest point is reached before the extreme, progressive tip of the ladder. We can take the three-dimensional political space metaphor a step further and plot the same data as a real landscape using appropriate shading and color. Figure 2 illustrates this – the figure can also be viewed as an animated GIF here.

Figure 2: Another view of the Dutch political landscape shown in Figure 1

To note: For more information, see the author’s accompanying article (co-authored with André Krouwel) in European Union Politics.

It is important to note that the political landscapes we find in Europe are quite different. Figure 3 shows Italian, for example. Unlike the Netherlands (and Germany), for social conservatives EU support is highest in the middle of the left-right socio-economic scale.

Figure 3: The Italian political landscape based on citizen positions on three dimensions of political ideology (2019)

To note: For more information, see the author’s accompanying article (co-authored with André Krouwel) in European Union Politics.

These images and analyzes are based on how people position themselves on the abstract dimensions of political ideology. But since we have data on the preferences of these same people regarding actual policies, we can check whether the abstract scales correspond to political positions, for example on redistribution or migration. It turns out that the links between concrete preferences and abstract scales are quite weak and don’t always work the way experts expect.

For example, preferences on socio-economic policy issues (think taxes, state intervention in the economy, workers’ rights, etc.) are only 0.56 correlated with self-positioning on the socio-economic left-right scale and even more (0.60) with self-positioning on the liberal-conservative scale (in the Netherlands the pattern is similar in Germany and Italy as well) . This is very problematic for the way we study politics. This essentially means that the numbers people choose on abstract ideological scales are not very predictive of the positions they hold on public policy issues.

Thus, self-placements on ideological scales and positions on policies that should relate to these scales are poorly correlated. But why stop there? It is in fact debatable whether these scales are relevant to how people’s ideas about politics and public policy are structured. To do this, we can inductively analyze how policy preferences on the thirty policy questions are clustered, based on the correlation between responses. When we do this (technically we are performing factor analysis with promax rotation of the polychoric correlations), the picture we get is even more complex. No less than six factors (or dimensions) are needed to account for the structure of preferences on the thirty questions in the Netherlands.

The most important factor includes elements related to the restriction of immigration, the limitation of the rights of asylum seekers, but also the opposition to a common EU unemployment scheme and financial solidarity. The second factor captures positions related to EU policy (exit from the EU and the euro zone, return to national borders and veto power for member states). But the third factor brings together another set of elements related to Europe, such as the establishment of a common taxation, army and foreign policy for the EU.

What is important to realize is that support for expanding EU competences is not lumped together with the rest of the EU elements, so these attitudes are to some extent independent of positions on whether the EU is good overall (factor 2) and support for individual policies (factor 1 and factor 4, which capture statements related to the environment). What about socio-economic left-right and social conservatism? These dimensions are recognizable in the data, but they only come in at places 5 and 6 in terms of importance (variance explained) and collect a very small number of preferences around them.

As you can guess by now, these structures are different in the four countries we study. The economic left-right could play a more important structuring role in France, and to a certain extent in Italy, than in Germany and the Netherlands. But in none of the four countries do preferences over the policies and future of the EU form a single dimension that can be summed up in a simple Eurosceptic-Europhile opposition.

Unsurprisingly, after seventy years of European integration, citizens’ views on the EU are complex and largely emancipated from their ideological positions on the left-right and liberal-conservative dimensions. This creates some very interesting and quite varied political landscapes across the continent that we as political scientists should explore further.

A final note on the generalizability of these findings to populations in the countries we study – a concern that people choose to use AAVs themselves rather than being (randomly) selected by researchers. We re-weigh the data to resemble the demographic profile of the populations, and we get very similar results. We also rerun the analyzes on subsets of all users, and we always get very similar results.

This is not unexpected, as we are not interested in estimating levels of support for the EU or for different policies (which will probably be biased in our samples) but in the relationships between preferences (which need not be biased). It is true that AAV users tend to be younger, better educated and more interested in politics than the average citizen. But if the political structures we discovered are already so complex based on the opinions of highly educated and politically sophisticated people, things will only get more complex when the opinions of people with little political interest or knowledge are better represented. .

For more information, see the author’s accompanying article (co-authored with André Krouwel) in European Union policy


Note: This article gives the author’s point of view, not the position of EUROPP – European Politics and Policy or the London School of Economics. Featured image credit: CC-BY-4.0: © European Union 2019 – Source: EP


About Michael G. Walter

Check Also

Scholz and Macron to relaunch the Franco-German engine of European integration – EURACTIV.com

German Chancellor Olaf Scholz and French President Emmanuel Macron have pledged to step up the …