Title: Variability in the estimation of item option characteristic curves for the multiple-category scoring model
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00099567/00001
 Material Information
Title: Variability in the estimation of item option characteristic curves for the multiple-category scoring model
Physical Description: xi, 128 leaves : ill. ; 28 cm.
Language: English
Creator: Buhr, Dianne C., 1943- ( Dissertant )
Crocker, Linda ( Thesis advisor )
Algina, James ( Reviewer )
Hedges, William ( Reviewer )
Miller, David ( Reviewer )
Smith, Davice C. ( Degree grantor )
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 1989
Copyright Date: 1989
Subjects / Keywords: Item response theory   ( lcsh )
Psychomatrics   ( lcsh )
Foundation of Education thesis Ph.D
Educational tests and measurements -- Mathematical models   ( lcsh )
Dissertations, Academic -- Foundations of Education -- UF
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
theses   ( marcgt )
Abstract: The present study was conducted to investigate the variability of the item option characteristic curves (OCCs) produced by the multiple-category scoring model under different conditions of estimation. Factors which were expected to affect estimation were test length, sample size, option, ability distribution of the sample of examinees, and difficulty distribution of the set of items. Data were drawn from two recent administrations of the mathematics subtest of the College Level Academic Skills Test (CLAST) , a minimum competency examination required for college sophomores in Florida. To investigate the effects of test length and sample size, ten random samples of 500 examinees each and ten random samples of 1000 examinees each were selected. The multiple-category scoring model of MULTILOG was used to estimate item parameters with subsets of four, six, eight, and ten items, with four items common to all sets. Option characteristic curves were calculated for the ten samples under each condition, and the variability of the OCCs across samples was computed by a variation of the Brown-Forsythe statistic. A series of repeated measures analyses of variance were conducted to determine whether the variability of the OCCs was affected by test length, sample size, or option. This analysis was replicated for a separate set of four common items. To investigate the effects of ability and difficulty distribution, item parameters were estimated with ten negatively skewed and ten normal samples of 1050 examinees each, for five-item subsets of "easy" and "difficult" items, with three common items. Again, analyses of variance were conducted with the Brown-Forsythe measure of the variability among OCCs as the dependent variable. Significant effects due to sample size were found for seven of the eight items, with more variability among OCCs for samples of 500 examinees than for samples of 1000 examinees. Significant effects due to test length were found for three of the eight items, with four-item subsets showing the greatest variability among OCCs. Ability distribution had a significant effect for one of three items only. There was no significant effect of difficulty distribution. Differences among options were related to the number of examinees selecting the option.
Statement of Responsibility: by Dianne C. Buhr.
Thesis: Thesis (Ph. D.)--University of Florida, 1989.
Bibliography: Includes bibliographical references (leaves 121-127)
Original Version: Typescript.
General Note: Vita.
 Record Information
Bibliographic ID: UF00099567
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 001500409
oclc - 21387277
notis - AHB3137


This item has the following downloads:

PDF ( 4 MBs ) ( PDF )

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - Version 2.9.9 - mvs