*********************************
There is now a CONTENT FREEZE for Mercury while we switch to a new platform. It began on Friday, March 10 at 6pm and will end on Wednesday, March 15 at noon. No new content can be created during this time, but all material in the system as of the beginning of the freeze will be migrated to the new platform, including users and groups. Functionally the new site is identical to the old one. webteam@gatech.edu
*********************************
Name: Yan Yan
Master’s Thesis Proposal Meeting
Date: March 11, 2020
Time: 1:00 pm
Location: J.S. Coon Building, Room 148
Advisor:
Susan Embretson, Ph.D. (Georgia Tech)
Thesis Committee Members:
Susan Embretson, Ph.D. (Georgia Tech)
James Roberts, Ph.D. (Georgia Tech)
Michael Hunter, Ph.D. (Georgia Tech)
Title: Vertical Equating with Longitudinal Data: Unidimensional and Multidimensional Item Response Theory Models
Abstract:
Vertical linking of test items is essential for understanding developmental processes in many areas of education. Several approaches for equating and linking have been studied (e.g., Kolan & Brennan, 2004), involving both classical test theory and IRT based methods. Vertical equating has proven especially difficult because the tests are administered at different levels of difficulty and thus the distributions of the examinees also differ (Hambleton et al, 1985; von Davier, 2011). Expressing the scores of different test occasions on a common scale allows the tracking of the amount of change or growth in student achievement over time (Rijmen, 2009). IRT models that involve change are particularly appropriate (Rijmen, 2010). This study aims to examine multidimensional and unidimensional IRT models to link item parameters in both a longitudinal study of mathematics achievement in 7th and 8th grade and a simulation study with a longitudinal design of mathematics achievement in two adjacent grades. For each study, model fit and linking performances will be compared among four IRT models, which are two unidimensional models (Rasch and 2PL) and two multidimensional models, including MRMLC and its 2PL version, the structured latent trait model (SLTM; Embretson, 1997).
