Modifications of the Test Information Function
Abstract
A minimum bound of any estimator, biased or unbiased, is considered, and, based on that, Modification Formula No. 1 is proposed for the maximum likelihood estimator, in place of the test information function. A minimum bound of the mean squared error is considered, and, based on that, Modification Formula No. 2 in the same context is proposed. Examples are given, and the usefulnesses of these modified test information functions in computerized adaptive testing are discussed. These topics are also discussed and observed for the monotonically transformed latent variable.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 30, 1990
- Accession Number
- ADA224698
Entities
People
- Fumiko Samejima
Organizations
- University of Tennessee