Measuring Issue-Specific Preferences From Votes

Sooahn Shin (Harvard University)

Abstract: How can we measure issue-specific ideal points using roll-call votes? Ideal points have been widely used for measuring ideology, yet its nature of latent space makes it difficult to target a specific issue area of interest. I propose a new measurement of issue-specific ideal points developing further on a standard spatial voting model as follows: first, fit a multidimensional ideal point model such as NOMINATE or IRT; then, using bill parameters, estimate a vector that represents a specific issue; finally, project the standard ideal points to the issue-specific vector. This method enables researchers to construct issue-specific ideal points leveraging information from the entire set of voting data and to identify the bundle of different issue areas.

View Poster in a New Tab