A Monte Carlo examination of external unfolding

Abstract

Monte carlo techniques were used to examine regression approaches to external unfolding. The present analysis examined the technique to determine if various characteristics of the points are recovered (such as ideal points). Characteristics of the situations were manipulated, including number of stimuli, error level, and measurement level. Generally, monotonic analyses resulted in good recovery. Recovery was poor when the data were generated by a weighted Euclidean process. Negative weights were commonly encountered, apparently due to error in the data. In this approach, estimation is done by statistical techniques, and some statistical concerns should be taken into account when examining the results. Index terms: external unfolding, monte carlo simulation, multidimensional scaling, nonmetric methods, parameter estimation, unfolding.