Publications using wordfish

  • Slapin, Jonathan B. and Sven-Oliver Proksch. 2008. "A Scaling Model for Estimating Time-Series Party Positions from Texts" American Journal of Political Science 52(3): 705-722.
    Recent advances in computational methods for extracting party positions from political texts have provided scholars promising new ways for estimating party positions. However, existing methods face serious challenges when it comes to producing time-series data. This leaves a gap in the literature on estimating party ideology. We fill this gap by proposing a scaling technique to estimate positions based on word frequencies in political texts. Our technique allows researchers to locate parties in one or multiple elections. We estimate the positions of German political parties from 1990-2005 from word frequencies in party manifestos. The extracted positions reflect changes in the party system more accurately than existing time-series positions. Our method allows us to examine which words are important for placing parties on the left and the right. We find that words with strong political connotations are the best discriminators between parties. Finally, we include a series of robustness checks and demonstrate that our positions are insensitive to distributional assumptions and document selection. [REPLICATION DATA]
  • Slapin, Jonathan and Sven-Oliver Proksch. 2009. "How to Avoid Pitfalls in Statistical Analysis of Political Texts: The Case of Germany", German Politics 18(3), 323-344.
    The statistical analysis of political texts has received a prominent place in the study of party politics, coalition formation and legislative decision making in Germany. Yet, we still lack a thorough understanding of the practical conditions under which such analysis produces unbiased estimates of policy positions. We examine the asymptotic behaviour of a recent scaling method called ‘Wordfish’. For this purpose, we conduct Monte Carlo simulations to demonstrate the effects of the choice of manifesto texts on party position estimates, including the number of documents included in the analysis and their length. Moreover, we present guidelines on how to process linguistic information for researchers interested in using the technique, focusing specifically on German texts. 
  • Proksch, Sven-Oliver and Jonathan B. Slapin. 2010. "Position Taking in European Parliament Speeches", British Journal of Political Science 40(3): 587-611.
    This article examines how national parties position themselves in European Parliament (EP) speeches. We apply  new computer-based content analysis technique, Wordfish, to estimate party positions in the European Parliament using the word counts in speeches. We test three hypotheses of position-taking in the European Parliament: a left-right ideology hypothesis, a pro/anti-Europe hypothesis, and a national politics hypothesis. Surprisingly, and in contrast to studies of roll call votes, we do not find evidence that national party positions from MEP speeches reflect the parties' overall left-right ideology. Instead, these positions reflect the parties' stance with regard to EU integration and national redistributive characteristics. We test the robustness of our results in a threefold manner. First, we take advantage of the multilingual environment of the European Parliament and show that the estimated positions are robust to the choice of translation (English, French, and German). Second, we use independent measures of national party positions from analyses of roll call votes and two expert surveys. Finally, we apply a range of statistical models to account for measurement error of the independent variables and the hierarchical structure of the data. Our robust findings suggest that the entire corpus of EP speeches reflects partisan divisions over EU integration rather than left-right politics. [REPLICATION DATA]
  • Thomas König, Bernd Luig, Sven-Oliver Proksch, and Jonathan B. Slapin. 2010. “Measuring Policy Positions of Veto Players in Parliamentary Democracies”, in König/Tsebelis/Debus (eds), Reform Processes and Policy Change: Veto Players and Decision Making in Modern Democracies, Springer Press, 69-98.
    Testing veto players theory in parliamentary systems requires researchers to have accurate estimates of policy preferences of political parties. In empirical studies, however, political scientists use remarkably context-free and invariant estimates of party positions. As a consequence, they must make assumptions about the nature of policy conflict. In the extreme, the lack of accurate estimates reduces sample sizes and increases the risk of selection bias in studies that use the ideological range among veto players as a key explanatory variable. We introduce a new automated method to identify policy relevant content in political texts. Our method “smart tags” sentences in election manifestos using portfolio-specific keywords from a legislative database. Policy positions can then be estimated on the tagged subset of relevant sentences. We apply our method to analyze the legislative positions of political parties for several portfolios in the German parliament. We show that these estimates vary over time and across portfolios – variation which cannot be produced using existing methods. We believe our method simplifies preference estimation and increases construct validity of policy preference measures used to evaluate policy-focused theories.