Abstract
The EU integration process is commonly recognised as a top-down initiative coordinated by elites. Do the differentiated patterns of EU politicisation across countries create a demarcation between two regions, namely Western Europe and Central and Eastern Europe? Through the analysis of data obtained by a standardised survey questionnaire submitted to a representative sample of national political elites in ten countries, we show that elite attitudes are uneven across countries and contingent on different issues. Although the distance between elites in the two regions is significant, elite attitudes do not necessarily qualify as irreducibly Eurosceptic in Central and Eastern Europe.
Notes
1 It would be enough to mention problems of asymmetric cooperation, as happens with the Visegrád regional group, which does not include all countries in the region.
2 The main methodological problem pertains to the limited availability of political elites to respond to surveys; for this reason, samples of national politicians tend to be smaller than samples taken to test public opinion. At the same time, however, samples should not be so small as to hamper the representativeness of the reference population, and for this reason, elite surveys cannot tolerate the higher refusal rates of mass surveys and are therefore more challenging and costly to manage. In purely operational terms, given their relatively large numbers, it is easier to achieve sizable and more representative samples by addressing MPs rather than other sectors of the political elite, such as ministers or junior ministers.
3 For a few exceptions, see the comparative surveys on EU attitudes edited by Conti et al. (Citation2010), Ilonszki (Citation2010), Best et al. (Citation2012) and Conti et al. (Citation2018); see also the collection of country studies edited by Vogel and Rodríguez-Teruel (Citation2016).
4 More information on the project is available at: https://enec-2014.wixsite.com/enec-2014, accessed 14 December 2020. The project data are available upon request.
5 Originally, France was also included in the survey but was dropped from the analysis because of our failure to reach a representative sample of French MPs. The participation rate of MPs differed between countries. In smaller parliaments, about half (for example, 56% and 46% in Slovenia and Croatia, respectively) of the population of serving MPs responded to the survey; this proportion was smaller (for example, 11% in Germany) in larger parliaments. We do not provide information about the number of contacted MPs and the response rate in every country, or about their motivations for participation or non-participation in the survey; suffice to say that we encountered the typical constraints of an elite survey, such as limited availability or contingencies—sudden change of dates—that interfered with scheduled plans, especially in the case of face-to-face interviews.
6 Unfortunately, some prominent country cases are not represented in the sample, mainly for financial reasons: the ENEC project is a transnational cooperative effort involving national teams that conducted national surveys using their own resources and support which were not necessarily available in other countries.
7 The last criterion was adopted to avoid the risk that the final sample would be too skewed towards inexperienced parliamentarians. The mismatches in the strata of the final sample were low in the case of party stratification, with the exception of France, which was excluded from the analysis, and gender, so they did not require post-stratification weights in the analysis. Such mismatches were greater in terms of political experience; however, considering that the impact of this factor does not represent a hypothesis tested in our analysis, the sample was not weighted. Furthermore, we did not use population size weights. These weights—normally used in public opinion research—correct for the fact that most countries have different population sizes but similar sample sizes, so any figures combining data from two or more countries might over-represent smaller countries at the expense of larger ones. Thus, adjustments to population size weights ensure that each country is represented in proportion to its population size. In our case, it would be more difficult to define the exact population size of national political elites. Weights could be adapted to the size of national parliaments, but it seems arbitrary to increase or decrease country representation in our elite sample based on this criterion, which simply reflects national institutional variations. For this reason, we decided not to introduce country weights in the analysis. At the same time, to guarantee the transparency of our results, in the Appendix we provide statistics broken down by country and comment on country variations in the text.
8 We did not introduce left and right ideology as, based on the same data, Makarovic and Tomsic (Citation2018) have already shown that the attitudes of MPs towards the EU do not match this divide.
9 The values of the Variance Inflation Factors (VIFs) of the independent variables are always smaller than ten with an actual range between 1.03 and 1.17.
10 For example, for exposure to the EU through direct contacts, or the divide between moderate and radical parties, see Conti et al. (Citation2018); or for the division between GAL (green alternative libertarian) and TAN (traditional authoritarian nationalist) ideology, see Hooghe and Marks (Citation2018).
Additional information
Funding
Notes on contributors
Marco Cilento
Marco Cilento, Assistant Professor of Political Science, Department of Communication and Social Research, Sapienza University of Rome, Via Salaria 113, 00198 Rome, Italy. Email: [email protected]
Nicolò Conti
Nicolò Conti, Professor of Political Science, Department of Law and Economics, Unitelma Sapienza University of Rome, Viale Regina Elena, 295 00161 Rome, Italy. Email: [email protected]