A Poisson Scaling Model for Estimating Time-Series Party Positions from Texts

Abstract

Recent advances in computational methods for extracting party positions from political texts have provided scholars promising new ways for estimating party positions. We add to this literature by developing a Poisson scaling tech-nique to estimate positions over time based on word frequencies in political texts. Our approach is more systematic and less time consuming than hand coding of party manifestos or conducting expert surveys. Unlike the computer word scoring methodology by Laver, Benoit & Garry (2003), our method does not require the use of reference texts, the selection of which may greatly influ-ence the estimation process. Lastly, we believe our approach is the first which produces party position estimates which can be used accurately as time-series data. Using simulations, we first demonstrate that our scaling technique pro-duces accurate ideal point estimates. We then estimate the positions of German political parties from 1990-2005 from word frequencies in party manifestos us-ing our R program WORDFISH. The extracted positions reflect changes in the party system more accurately than existing time-series positions. ∗Corresponding author. Acknowledgement: We would like to thank James Honaker, Jeff Lewis, and George Tsebelis 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. 1 1

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 1 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.