Positions and saliency of immigration in party manifestos: A novel dataset using crowd coding

Abstract

Immigration is one of the most widely debated issues today. It has, therefore, also become an important issue in party competition, and radical right parties are trying to exploit the issue. This opens up many pressing questions for researchers. To answer these questions, data on the self‐ascribed and unified party positions on immigration and immigrant integration issues is needed. So far, researchers have relied on expert survey data, media analysis data and ‘proxy’ categories from the Manifesto Project Dataset. However, the former two only give the mediated party position, and the latter relies on proxies that do not specifically measure immigration. The new dataset presented in this article provides researchers with party positions and saliency estimates on two issue dimensions – immigration and immigrant integration – in 14 countries and 43 elections. Deriving the data from manifestos enables the provision of parties’ unified and unfiltered immigration positions for countries and time points not covered in expert surveys and media studies, making it possible to link immigration and immigrant integration positions and saliency scores to other issue areas covered in the Manifesto Project Dataset. Well‐established criteria are used to distinguish between statements on (1) immigration control and (2) immigrant integration. This allows for a more fine‐grained analysis along these two dimensions. Furthermore, the dataset has been generated using the new method of crowd coding, which allows a relatively fast manual coding of political texts. Some of the advantages of crowd coding are that it is easily replicated and expanded, and, as such, presents the research community with the opportunity to amend and expand upon this coding scheme.

Politicizing immigration in Western Europe

“…As the issue categories of the Manifesto Project do not include an issue category for immigration (see Lehmann and Zobel [2018]: 2), we provide novel indicators for the issue attention of parties and their positions on this topic in party manifestos to measure the concept. For this purpose, we use the man-ifestoR corpus which enables applying text mining approaches to the manifestos covered by the MARPOR project (Lehmann et al 2017;Volkens et al 2017).…”

“…To validate this method, we use data on parties' issue attention and positions on immigration also measured in party manifestos using a crowdsourced coding approach (Lehmann and Zobel 2018). As many manifestos are covered in both studies, it is possible to use this study to validate our indicators.…”

Validating a sentiment dictionary for German political language—a workbench note

“…But this only works if the underlying sentiment weights adequately reflect term usage in the political context of interest. The technically more advanced literature recently offered context-specific machine-learning approaches (e.g., Ceron, Curini, & Iacus, 2016;Hopkins & King, 2010;Oliveira, Bermejo, & dos Santos, 2017;Van Atteveldt, Kleinnijenhuis, Ruigrok, & Schlobach, 2008), sometimes paired with crowd-sourced training data (Lehmann and Zobel 2018;Haselmayer & Jenny, 2017), in this regard. Yet, especially in projects where expressed sentiment is only one variable in a broader analytical setup, the computational, financial, or human resources required for such approaches can quickly offset the comparative advantages that led to conducting an automated analysis in the first place.…”

Does Radical-Right Success Make the Political Debate More Negative?

“…To ensure valid and reliable data production, it is essential to build in several quality control measures. Similar to recent studies using crowd coding for data production (Benoit et al 2016;Lehmann and Zobel 2018;Rudkowsky et al 2018), the validation process included different test questions before and during the coding process to assure a high quality of coding (see Online Appendix B for details). The coded validation sentences serve now as "true" answers and allow us to measure the accuracy of the ed8 dictionary.…”

“…("I hate the government and all these idiots!") Anger Fear Disgust Sadness Joy Enthusiasm Pride Hope 2 1 1 1 0 0 0 0 To validate the dictionary at hand, we have chosen crowd sourcing (or crowd coding) which is one of state-of-the-art methods used to code text in the social sciences (Benoit et al 2016;Lehmann and Zobel 2018). Crowd sourcing refers to using the internet to distribute large amounts of small tasks to a very high number of workers around the world that receive financial rewards per task.…”