We would like to thank Kathleen Bawn, Ken Benoit, Jim DeNardo, Tim Groseclose, James Honaker, Thomas König, Jeff Lewis, Will Lowe, Burt Monroe, George Tsebelis, several anonymous reviewers, and participants at the UCLA Methods Workshop and the 2007 annual meeting of the Midwest Political Science Association for their comments and suggestions. In addition, we thank the Zentralarchiv für Empirische Sozialforschung at the University of Cologne, Germany, for providing us with the German party manifestos in electronic format. The order of authors' names reflects the principle of rotation. Both authors have contributed equally to all work.
A Scaling Model for Estimating Time-Series Party Positions from Texts
Article first published online: 9 JUL 2008
©2008, Midwest Political Science Association
American Journal of Political Science
Volume 52, Issue 3, pages 705–722, July 2008
How to Cite
Slapin, J. B. and Proksch, S.-O. (2008), A Scaling Model for Estimating Time-Series Party Positions from Texts. American Journal of Political Science, 52: 705–722. doi: 10.1111/j.1540-5907.2008.00338.x
- Issue published online: 9 JUL 2008
- Article first published online: 9 JUL 2008
Recent advances in computational content analysis have provided scholars promising new ways for estimating party positions. However, existing text-based methods face challenges in producing valid and reliable time-series data. This article proposes a scaling algorithm called WORDFISH to estimate policy positions based on word frequencies in texts. The technique allows researchers to locate parties in one or multiple elections. We demonstrate the algorithm by estimating the positions of German political parties from 1990 to 2005 using word frequencies in party manifestos. The extracted positions reflect changes in the party system more accurately than existing time-series estimates. In addition, the method allows researchers to examine which words are important for placing parties on the left and on the right. We find that words with strong political connotations are the best discriminators between parties. Finally, a series of robustness checks demonstrate that the estimated positions are insensitive to distributional assumptions and document selection.