|
Activity Number:
|
499
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 10, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Social Statistics Section
|
| Abstract - #305844 |
|
Title:
|
Test Equating for Mixed-IRT Models: an Illustration with 3PL and Generalized Partial Credit Model Using the Stacking-Lord Approach
|
|
Author(s):
|
Yann-Yann Shieh*+
|
|
Companies:
|
|
|
Address:
|
28 Adrain Way, Fredericksburg, VA, 22405,
|
|
Keywords:
|
test equating ; IRT
|
|
Abstract:
|
Test equating is essential to reporting progress on standardized and criterion-driven assessments. In many situations, using mixed-IRT models to calibrate the test items is either desirable or inevitable. Currently, few references have demonstrated how mixed-model equating can be accomplished. In this paper, we use three-parameter logistic and generalized partial credit models as an example for test equating with mixed-models. We demonstrate how the Stocking-Lord characteristic curve approach can be use to perform the IRT scale transformation when the common linking items consist of different test types and are calibrated with different IRT models. Data are from a state's grade 3 Reading achievement tests administered in 2003 and 2004. The goal is to equate the 2004 tests to the 2003 tests.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2006 program |
