Reassessing conceptualization, data and causality: A critique of Boomgaarden and Vliegenthart’s study on the relationship between media and the rise of anti-immigrant parties Teun Pauwels * Centre d’e´tude de la vie politique, Universite´ Libre de Bruxelles, Institut de Sociologie, CP124, Avenue Jeanne 44,1050 Brussels, Belgium article info Article history: Received 1 July 2009 Received in revised form 22 February 2010 Accepted 23 February 2010 Keywords: Populist radical right News content Voting behavior abstract This article critically assesses the study of Boomgaarden and Vliegenthart (2007) on the relationship between media content and anti-immigrant party support. With regard to conceptualization, it is argued that their dependent variable is flawed, because it groups two parties together that do not belong to the same party family. Some data-related issues, such as measurement equivalence, are also discussed. Finally, it is argued that the causality might be easily reversed. Because anti-immigrant parties are able to exploit issues which are neglected by mainstream parties, it is possible that the media will give more attention to these issues. This study concludes that a causal relationship between news content and anti-immigrant party support is nonexistent in the Dutch case. Ó 2010 Elsevier Ltd. All rights reserved. Studies addressing the question of why populist radical right parties emerge in contemporary democracies have proliferated in the last decade (Arzheimer and Carter, 2006; Carter, 2005; Eatwell, 2003; Lubbers et al., 2002; Rydgren, 2005; Van der Brug et al., 2005). 1 Many of these scientific contributions focus on why populist radical right parties are successful in some countries while they are ephemeral or nonexistent in others (e.g. Golder, 2003). More recent research also explores populist radical right support in a longitudinal perspective (e.g. Arzheimer, 2009). Even though these numerous empirical studies have taught us a great deal, it is frustrating that the findings are often very inconsistent and at times even completely contradictory (Mudde, 2007b). Many of these differences might be explained by a difference in time span, level of analysis or case selection, but many scholars seem to be unaware of the impact of these choices. In some cases it was demon- strated how a flawed conceptual and methodological approach yields spurious empirical findings (Arzheimer and Carter, 2009). In this study, I will argue that more attention should be paid to the conceptualization, data and causality assump- tions in explaining the rise of the populist radical right so as to reach valid research results. This will be illustrated by critically assessing the article by Boomgaarden and Vliegenthart (2007) ‘‘Explaining the rise of anti-immigrant parties: The role of news media content’’, in which it is argued that news content should be brought into the equation in order to understand the rise of anti-immigrant parties in the Netherlands. The authors find that the more the news media reported about immigration-related topics, the higher the aggregate share of intention to vote for anti- immigrant parties, controlled for real-world events such as the level of immigration and unemployment for the period from 1990 to 2002. They also claim that the relation between news media content and anti-immigrant party success probably holds across the borders of the Benelux, and should therefore be included in future research. * Tel.: þ32 2 650 33 88. E-mail address: tepauwel@ulb.ac.be URL: http://www.cevipol.site.ulb.ac.be/fr/membres_pauwels-teun.html 1 Populist radical right is my preferred term to denote parties such as the Vlaams Belang, Front National or Freiheitliche Partei O ¨ sterreichs. Because the article I discuss uses the term anti-immigrant party, the two concepts will be used interchangeably. Contents lists available at ScienceDirect Electoral Studies journal homepage: www.elsevier.com/locate/electstud 0261-3794/$ – see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.electstud.2010.02.006 Electoral Studies 29 (2010) 269–275