Linking in developmental scales Public Deposited

Downloadable Content

Download PDF
Last Modified
  • March 21, 2019
Creator
  • Langer, Michelle M.
    • Affiliation: College of Arts and Sciences, Department of Psychology and Neuroscience
Abstract
  • Developmental test score scales have attracted more interest with recent legislation on educational measurement, yet there remain a surprising number of unanswered methodological questions about score scales. This research examined the effects of the number and difficulty level of linking items, the inclusion of noncommon items, the size of the mean difference between groups, and the sample size in accurately linking a score scale across groups. Data were simulated for binary items based on parameters from the 1998 National Assessment of Educational Progress Reading Assessment. Item response theory concurrent linking was used for 1,000 replications each under more than 100 conditions. These results were evaluated in terms of bias and random variance of the recovered group mean differences and standard deviations. The results indicated that linking is generally reliable, regardless of other factors, when group means are 2 standard deviations apart or less.
Date of publication
DOI
Resource type
Rights statement
  • In Copyright
Advisor
  • Thissen, David
Language
Access
  • Open access
Parents:

This work has no parents.

Items