ESRA logo

ESRA 2023 Glance Program

All time references are in CEST

Cross-national multi-purpose survey data as a resource for political research 1

Session Organisers Dr Riccardo Ladini (University of Milan)
Miss Jessica Rosco (University of Milan)
Professor Cristiano Vezzoni (University of Milan)
TimeWednesday 19 July, 14:00 - 15:00
Room U6-06

The panel invites papers that illustrate the potential of large cross-national multi-purpose surveys not specifically focused on political issues to study electoral and political behaviour and politics at large, both focusing on specific countries or comparative approaches.
Some examples of multi-purpose survey programs are: ESS, EVS/WVS, ISSP, and various continental barometers.
Nowadays, social research broadly relies on data coming from these surveys, which are increasingly recognised as the infrastructure of the social sciences. Their success rests on the high quality of data, cross-national coverage, longitudinal depth, complete documentation and, last but not the least, availability.
Despite their merits and popularity, cross-national survey data are still under-used to study political attitudes and behaviour. Possibly political scientists’ cold feet toward cross-national surveys comes from their inherently multi-purpose nature, with a limited number of questions explicitly referring to political orientation and behaviour. To complicate the situation, each round of a survey happens at once in several countries, crosscutting each national electoral cycle at a different point.
A closer look, however, suggests that cross-national survey data can offer great opportunities also for the study of politics. Practically each survey programme includes basic political questions (previous vote, party identity, self-placement on the left-right scale) together with an abundant set of questions referring to value orientation and socio-political attitudes. In addition, the synchronic data collection in several countries and repetition across time give a vantage point to study cross-cultural equivalence of measures and contextual effects, by means of multi-groups and multilevel designs.
Thus, the aim of the panel is to collect contributions that present original and creative ways for political researchers to exploit cross-national survey data, enhancing their use by proposing solutions to the methodological and conceptual problems encountered while


Leaning into partisanship: Follow-up questions as a way to improve measures of political affiliation in a cross-national survey

Dr Laura Silver (Pew Research Center)
Dr Patrick Moynihan (Pew Research Center) - Presenting Author

In the United States, the Pew Research Center measures partisan identification using a two-part question. The first question asks individuals, “In politics TODAY, do you consider yourself a Republican, Democrat, or independent?” Because around four-in-ten Americans consider themselves an independent – and because the U.S. a system composed of two dominant political parties – this question can be easily followed by a question about political leaning: “As of today, do you lean more to the Republican Party or more to the Democratic Party?” In the U.S., around eight-in-ten independents will then typically express a preference for one side or the other.

In cross-national surveys, the Pew Research Center has traditionally measured party affiliation by asking, “Which political party do you feel closest to” and then presented a pre-coded list to interviewers. In 2022, the share across the 19 countries surveyed who did not offer a substantive response to this question ranged between 15% in Israel and Sweden to 57% in Japan – with the largest share in most countries saying they simply “do not feel close to any particular party.” Given the experience in the U.S., where a follow-up netted significantly more people who were willing to declare themselves aligned with a party, we introduced follow-up questions in the other 18 countries. First, we asked those who had not offered a substantive reply to the first question whether there is any party they feel closer to than others. Then, for those who said yes, we asked them to name the party. This allowed us to create a measure of "true partisans" and "leaners" for each of the countries.

In this paper, we explore the results of this change, examining what share of people offered substantive responses to the follow-up question and the characteristics of those who provided a response to the leaning question but not to the original party identification question. We also examine whether “leaners” and party supporters are similar or different on key attitudinal measures, including satisfaction with democracy, favorability of key political parties, views of international organizations, and more.

The Index of Political Inequality (IPI): Introducing a New Index of Inequality in Political Voice Based on Longitudinal Cross-National Survey Data

Dr Vardan Barsegyan (Utrecht University) - Presenting Author

The paper introduces a new Index of Political Inequality (IPI) and presents first results of its measurement and application. Political equality is a cornerstone of a democratic society. More particularly, the IPI reflects inequality in political voice measured by different forms of political participation. Based on the requirement that the sample of people that are active in politics should not differ from the general population, or, in other words, that social characteristics of people should not affect their political participation, I argue that the extent to which social characteristics of people do affect participation can be an indicator of political inequality. Therefore, I measure the IPI as a coefficient of determination (R-squared) that reflects the predictive power of social characteristics of people in their political participation. I derive the IPI for 29 European countries based on ESS data from 2008 to 2018. I measure political participation as a two-parameter logistic item response model based on nine items of political participation. I include nine social characteristics of people in the model such as age, education, gender, education, ethnic minority background, and parental socioeconomic status. The paper shows the trends in political inequality across 29 countries from 2008 to 2018. The paper also shows bivariate correlations between political inequality and other country level-variables such as country turnout level per year, Gini-index, GDP per capita, different democracy and liberalism indexes. Although this paper is mostly introducing the IPI and presenting some bivariate statistics, a set of time-series regression models predicting IPI are attached. My ambition is that IPI becomes one of the most crucial explanatory, dependent, and control variables that are included in any cross-national survey research on political inequality. The index will be renewed yearly for all countries with survey data on political participation.

Identifying the Causal Effect of Foreign Language Education on Political Attitudes Using Educational Cohorts in Survey Data.

Dr Roland Kappe (University College London) - Presenting Author

This paper provides empirical evidence that that foreign language teaching in school can have a substantial (causal) effect on political attitudes later in life.

Prior research has identified a possible link between foreign language skills and political attitudes, especially European identity (Kuhn 2015, Díez-Medrano 2017). The fundamental problem with existing research is that self-reported language skills may be endogenous to political attitudes, and that language skills are correlated with other individual and contextual factors known to affect political attitudes and values.

This paper tries to overcome these challenges by leveraging plausibly exogenous variation in foreign language learning due to education reforms. Existing survey data such as the Eurobarometer, the European Social Survey, and the British Election Study are used to identify educational cohorts. This allows us to estimate the effect of education reforms on political attitudes by comparing pre- and post-reform cohorts in a regression discontinuity framework.

For a major language education reform in Scotland, the results show that respondents who started high school just after the reform were substantially more likely to vote for remain compared to cohorts not exposed to the reform. Results using education reforms in other countries show similar patterns.