Scaling policy preferences from coded political texts

摘要: Scholars estimating policy positions from political texts typically code words or sentences and then build left-right scales based on the relative frequencies of text units coded into different categories. Here we reexamine such propose a theoretically linguistically superior alternative logarithm odds-ratios. We contrast this scale with current approach Comparative Manifesto Project (CMP), showing that our proposed logit avoids widely acknowledged flaws in previous approaches. validate new using independent expert surveys. Using existing CMP data, show how to estimate more distinct dimensions, for years, than has been possible before, make dataset publicly available. Finally, draw some conclusions about future design coding schemes texts.