Comparing the Position of Canadian Political Parties using French and English Manifestos as Textual Data

Abstract

Recently, computer-assisted, quantitative methods have been developed to position political parties. These word-based textual analysis techniques rely exclusively on the relative frequency of words. As such they do not necessitate the knowledge of any particular language to extract policy positions from texts. However, different languages have different word distributions and other syntactic idiosyncrasies. These differences might provoke word-based textual analysis techniques to extract noticeably different positions from parallel texts that are similar in every aspect except language. How crippling is this potential disadvantage when comparing political texts written in different languages? It is this chapter’s objective to determine the effect of language on the two word frequency methods Wordscores and Wordfish by comparing the policy positions of Canadian parties as extracted from their English and French party manifestos.

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 — 2 citations exist upstream.

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