Ideal Point Estimation with Partial Rank Preference Data

• In Event: Scaling and Ideal Point Estimation

Abstract

Many public opinion surveys and polls include question items in which respondents select and rank order a couple of alternatives they most prefer out of multiple choice options. While such partial rank (or top rank) question items have their benefits, they have been underutilized by public opinion surveys conducted by political scientists and especially by political science research using these data, partly because there are currently no easy-to-use methods commonly known among political scientists. Building on the previous literature on statistical methods to deal with partial rank data in other fields, this paper introduces a method to estimate respondents' underlying ideal points from answers to partial rank survey/polling items. This paper also examines the performances of the proposed method using both simulated datasets and existing public opinion survey data.

Author

• Ikuma Ogura, Georgetown University