RESEARCH

Experts, Coders, and Crowds: An analysis of substitutability (with Bernhard, M., Coppedge, M., Lindberg, S., Marquardt, K., Pemsteim, D., Seim, B. & Wilson, S.). The Varieties of Democracy Institute Working Paper Series 2017:53. Available at SSRN: https://ssrn.com/abstract=3046462

Recent work suggests that crowd workers can replace experts and trained coders in common coding tasks. However, while many political science applications require coders to both find relevant information and provide judgment, current studies focus on a limited domain in which experts provide text for crowd workers to code. To address potential over-generalization, we introduce a typology of data producing actors—experts, coders, and crowds—and hypothesize factors which affect crowd-expert substitutability. We use this typology to guide a comparison of data from crowdsourced and expert surveys. Our results provide sharp scope conditions for the substitutability of crowd workers: when coding tasks require contextual and conceptual knowledge, crowds produce substantively different data from coders and experts. We also find that crowd workers can cost more than experts in the context of cross-national panels, and that one purported advantage of crowdsourcing— replicability—is undercut by an insufficient number of crowd workers.

 

V-Dem Annual Report 2017. Democracy at Dusk? (Lührmann, Anna; Lindberg, Staffan I. ; Mechkova, Valeriya ; Olin, Moa ;Piccinelli Casagrande, Francesco ; Sanhueza Petrarca, Constanza ; Saxer, Laura). V-Dem Working Paper . Available at SSRN: https://ssrn.com/abstract=3063934

V-Dem first Annual Report. With this annual report, we release the new version 7.0 of the data, covering 177 countries, 117 years, and more than 350 indicators, as well as 52 indices measuring varying aspects of democracy. Most of these data are also made available for online analysis on our web-page, where users, including those without a statistical background, can produce and download their own graphs without having to download the entire dataset.

 

“Immigrants’ Representation in France” in Pathways to Power: The Political Representation of Citizens of Immigrant Origin in Europe  (Eds. Thomas Saalfeld and Laura Morales). Oxford University Press (with Chabanet, D., Cinalli, M., Sanhueza, M.J. & Van Hauwaert, S.)

In this book chapter, the representation of citizens of immigrant origin (CIOs) in the National Assembly between 1993 and 2012 is examined. At first, the study presents the context of immigration to France and the main characteristics of the assimilation policies designed to integrate the newcomers. Then we summarize and discuss the literature about immigrants’ representation in France and we identify the main gaps. In the analysis section, we examine how electoral systems, party electoral dynamics and the opinions of the French public affect the selection and election of CIO candidates. Subsequently, we study the extent to which CIOs are represented in the National Assembly and whether IO MPs represent immigrant voters. Finally, we explain the main patterns and trends in the political representation of CIOs in France.

“When do MPs of Immigrant Origin Talk about Immigration?” in Pathways to Power: The Political Representation of Citizens of Immigrant Origin in Europe (Eds. Thomas Saalfeld and Laura Morales). Oxford University Press (with Morales, L.)

This study investigates to what extent MPs of immigrant origin differ from those without an immigrant background in how much they talk about immigration and integration in national legislatures, how do they frame the issues, and whether they present similar or different positions in their questions than in the party programs. The chapter examines the written parliamentary questions on immigration asked by individual MPs to members of national governments in the 8 European legislatures covered by the Pathways project.

“Descriptive Representation of Constituency Interests: the Role of Institutional Variables and District Composition” in Understanding the Patterns of Political Representation of Citizens of Immigrant Origin in Europe (Eds. Thomas Saalfeld and Laura Morales). Oxford University Press (with Janssen, C., Geese, L. & Schacht, D.)

This book chapter examines the impact of electoral institutions on the descriptive representation of CIO at the district-level in seven European countries between 2005 and 2015. It combines three PATHWAYS data sets at the level of electoral districts to study if and how differences in district magnitudes, ballot structure, and voter to seat ratios impact on the descriptive representation of CIO. The results show that in all kinds of districts higher CIO shares translate into more elected immigrant-origin MPs, hence suggesting that the electoral connection works in different types of electoral systems. The strength of this link is, however, affected by district magnitude, with smaller magnitudes showing significantly more descriptive representation. Moreover, multi-member-districts that allow for preferential voting tend to fare better in terms of descriptive representation than under conditions of closed-list voting. Electoral institutions do neither seem to matter due to their consequences for party representation or legislative turnover, while we find some evidence that the relative size of a district decreases the descriptive representation of CIO in multi-member district elections. On the whole, then, this chapter suggests that even though electoral systems do not matter in accordance to a majoritarian-proportional divide as suggested elsewhere, they still matter in other ways. Hence, future research should study electoral system characteristics that go beyond the distinction of PR vs. majoritarian electoral institutions. An especially fruitful path for future research may be a stronger examination of electoral systems’ consequences for personal vote seeking incentives.

 

2011. Parliamentary Questions on Immigration in Britain, France, and Germany: Is MPs’ Background Relevant? Working Paper 2/2011, University of Bamberg (with Saalfeld, T. & Wüst, T.) Online here. 

One of the main premises of representative democracy is that all those subject to political decisions should have a voice in the policy-making process. However, the legislatures of electorally accountable governments often fail to mirror the interests and perspectives of some social groups. In this paper, we are interested in learning more about the presence of immigrant-origin Members of Parliament (MPs) in Britain, France and Germany and the representational bond they create with their constituents. Focusing empirically on behavioural indicators such as issues relating to immigration and the interests of immigrant-origin minorities in written parliamentary questions, we examine the extent to which variables such as party membership and constituency composition ‘determine’ MPs’ parliamentary activities and whether parliamentarians’ background as immigrants or descendants of immigrants is of additional explanatory value for voicing concerns relating to immigration and the interests of immigrant-origin minorities. The study analyses the content of written questions regarding migration-related issues in the House of Commons (2005-2010), the Assemblée Nationale (2007-2010) and the German Bundestag (2005-2009). We examine the behavior of immigrant-origin MPs and contrast it with the behavior of native MPs from comparable constituencies.

 

 

WORKING PAPERS

MEASURING MULTICULTURALISM

Measuring Multicultural Democracy (with Lindberg, S.)

Why do groups of citizens of immigrant origin, historical national minorities, and indigenous people participate more, and have representation in elected bodies to a higher degree, in some democracies than in others? This paper conceptualizes a new variety of multicultural democracy (MCD) as a regime characterized by three core characteristics beyond the basic requirements of electoral democracy (polyarchy): 1) inclusive political institutions, 2) the protection of minority rights and, 3) minority de facto representation. The study details and justifies this conceptualization of MCD and then addresses how to go about to measure it presenting a novel MCD index with global coverage based on expert-coded indicators developed by the Varieties of Democracy (V-DEM) project and other data sources. Additionally, it demonstrates theoretically and empirically how ethno-linguistic fractionalization, levels of economic development and education inequality and political stability, colonial heritage, and institutional characteristics relate to MCD. Finally, it shows that when multicultural democracies are stronger the levels of ethno-linguistic fractionalization become less relevant in terms of defining access to resources or trusting fellow citizens.

 

ELECTIONS 

The 2017 French Presidential and Legislative Elections Campaigns in Social Media: Trends, Popularity, and Ideological Distance (with Steven L. Wilson)

In France, the 2017 presidential and legislative elections marked a turning point in the use of Twitter in French politics compared to previous elections. For the first time, all presidential candidates extensively used this tool to communicate their thoughts and views during the campaigns, and parties devoted greater resources to publicize rallies, present their policy positions, and engage in online debates. In order to evaluate the role of social media in these elections, we collected the full text of sixty million tweets that either were geocoded as originating from within France, or matched a set of a hundred keywords salient to the French election. This work yielded several important contributions to understanding how social media is used in elections by political actors in addition to demonstrating how social media can be used to measure aspects of election campaigns. First, it uses computerized content analysis via Latent Dirichlet Allocation techniques to examine topic trends across candidates and parties, identifying the emphasis on different policy areas (especially immigration, one of the primary issues of these elections), target groups addressed, and collocations (pairs or group of words that co-occur). Second, it presents a measure of message popularity based on the number of likes and re-tweets, which allows to analyze candidates and parties’ social media outreach. Finally, it includes a measure of ideological distance based on the degrees of lexical variations that allow comparing the ideological distance across candidates and parties, and between candidates and their parties.

 

Do immigrant voters support parties that are more diverse?  (with Daphne van der Pas, University of Amsterdam)

Do voters of immigrant origin vote for parties that nominate more diverse candidates? This paper explores the voting behavior of voters of immigrant origin in eight European democracies. First, we study whether and how individuals’ socio-demographic characteristics such as immigrant background, age, gender, and education influence preferences for parties that have more diverse candidates running for legislative office. In particular, we examine whether the descriptive representation hypothesis holds for immigrant voters from all world regions of the worlds and across contexts. Second, we explore some of the mechanisms that influence voters’ preferences for descriptive representatives and show that experiences of discrimination based on race, religion, and ethnicity influence the voting preferences of voters with an immigrant background. Finally, we explore the mediating effect of parties’ ideological position and their position on immigration issues. The empirical analysis combines data on the background and voting behavior of more than 20,000 European Social Survey respondents with data on the origins and background of more than 3,000 European parliamentarians gathered by the Pathways project, and the ideological position and position on immigration of parties using from the Chapel Hill Expert survey.

 

GENDER & CORRUPTION

Do women MPs care more about corruption? New data from the UK, France, and Spain (with Alexander, A. & Bagenholm, A., University of Gothenburg)

It is well-established that women’s presence in parliaments positively correlates with control of corruption (Dollar et al. 2001; Esarey and Chirillo 2013; Esarey and Schwindt-Bayer 2015; Goetz 2007; Sundström and Wängnerud 2014). To explain this link scholars theorize that female MPs are more likely to resist or actively fight corruption due to gender differences in socialization or a higher likelihood to be marginalized from power. However, few studies actually evaluate how female politicians behave as MPs as concerns corruption. Thus, evidence for establishing causality and sorting out the mechanisms behind the link between women’s presence and lower corruption is seriously limited. This paper unpacks this “individual-level black box” with new data on the extent to which female MPs differ from male MPs in their concern for corruption when posing questions to members of government during their last full mandate periods. More specifically, we examine the written parliamentary questions (WQs) on corruption asked by individual MPs to members of the Government in the national 2010 British, 2007 French and 2011 Spanish parliaments. We ask whether female MPs are more active in bringing the issue of corruption up than their male colleagues and look into the possible mechanisms behind these gender differences. We use a dataset on WQs (Sanhueza Petrarca 2015) for the United Kingdom, France and Spain. WQs are a tool with which MPs can hold governments accountable (Wiberg 1995) and an instrument to direct policymakers’ and voters’ attentions to the issues that they care about (Ferree et al. 2002:14). Moreover, the content of the question shows the representation orientation and framing adopted by individual deputies. WQs thus provide an insight into individual MPs’ concerns especially because they are not subject to such strict control by party leaders as compared to other activities (Martin 2011). To capture the full of extent of corruption and corruption related topics, we have produced a refined list of keywords that are either clearly related to corruption or have a high probability of being related to corruption. These key words have also been checked manually as a validity test. The coding of the questions has followed strict procedures combining automated and manual coding methods. Using existing official documents and considering dominant national discursive frameworks, words (including stems) relating to corruption have been identified for all three countries. Following an automated process that searches for all the keywords identified for each country, each individual WQ has been coded as related to corruption (value of 1) or not related to corruption (value of 0). In a second step, we have examined the data and searched for false negatives and positives, being able to expand the keywords related to corruption in several iterations of this process. MPs’ WQ activity is measured counting the number of written questions tabled by individual deputies on corruption issues. Hence, the number of questions captures the saliency (importance) of the issue of corruption for a given deputy.The preliminary results show that female MPs pose more WQs on corruption than male MPs in the UK, while male MPs pose more WQs on corruption than female MPs in France and Spain. These are, however, absolute numbers. If we adjust for women’s lower percentages in parliament in all countries, than in relative terms, the positive results in the UK are even stronger and female MPs are now more likely than men to engage against corruption in France and Spain. Thus, in relative numbers, female MPs are significantly more likely to show concern for corruption over male MPs across all three countries.This investigation, therefore, makes an important contribution to the larger research on women’s inclusion and lower corruption by bringing in supportive micro-level evidence. Furthermore, ongoing research (planned for completion in spring 2017) will use the data to explore and test the mechanisms behind the gender differences. This will not only add to the evidence base but also expand the theory on our understanding of the link between gender and corruption among MPs.

 

 

 

RESEARCH PROJECTS 

Civic Engagement in Berlin during the Refugee Crisis (with Wendy Pearlman, Northwestern University): the aim of this project is to examine whether and how civil society organizations contribute to refugees’ integration to Germany. At present, we are collecting data in Berlin. 

 

DOCTORAL DISSERTATION

(2015) Voters, Parties, and Representation in European Multicultural Democracies. Doctoral Thesis. Mannheim: University of Mannheim

My doctoral dissertation investigates the transformations that national electorates, parties, and political institutions are experiencing in the so-called “age of migration” (Castles & Miller 2009) in Western Europe. How does immigration affect electoral democracies? Are immigrant and native voters alike? Or, does immigration emerge as a new social cleavage? Are parties concerned about immigration and open to include immigrant candidates? And, how are immigrants represented in national legislatures? The six empirical studies compiled in the thesis examine and compare European democracies using existing datasets and original data.