Comparison of the Dichotomous and Polychotomous Quantal Response Models.
Abstract
A class of polychotomous quantal response models is defined, including the dichotomous response model. As an application of the Missing Information Principle, it is shown that the Fisher information for the common parameters in the more elaborate model is greater. For the logistic curve and several design schemes, the above result is supported in a numerical study. For estimating the common parameters of the dichotomous-response and trichotomous-response models and the percentiles of the common tolerance curve, an extensive simulation study shows that in general the latter model gives more accurate estimates than the former one. The situations in which such gains are substantial are identified. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1983
- Accession Number
- ADA137938
Entities
People
- C. F. J. Wu
- S. K. Tse
Organizations
- University of Wisconsin–Madison