Computer Programs for Scoring Test Data with Item Characteristic Curve Models.
Abstract
Three computer programs are described for scoring test response data using item characteristic curve (ICC) or latent trait models. The rationale and mathematical basis of both maximum likelihood and Bayesian ICC scoring methods are presented, as well as some data comparing the two methods of scoring. The three computer programs are designed for scoring conventional (linear) test data (LINDSCO) in dichotomous response format, adaptive test dichotomous data (ADADSCO), and conventional (linear) test data scored by polychotomous ICC models (LINPSCO). Options available in these three general purpose programs are described, and examples of the input and output are given for each program. Complete FORTRAN listings of the three programs are included. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 01, 1979
- Accession Number
- ADA067752
Entities
People
- David J. Weiss
- Isaac I. Bejar
Organizations
- University of Minnesota