ERIC Number: ED652532
Record Type: Non-Journal
Publication Date: 2020
Pages: 296
Abstractor: As Provided
ISBN: 979-8-6912-2656-4
ISSN: N/A
EISSN: N/A
Available Date: N/A
Application of Item Response Tree (IRTree) Models on Testing Data: Comparing Its Performance with Binary and Polytomous Item Response Models
Yixi Wang
ProQuest LLC, Ph.D. Dissertation, The Ohio State University
Binary item response theory (IRT) models are widely used in educational testing data. These models are not perfect because they simplify the individual item responding process, ignore the differences among different response patterns, cannot handle multidimensionality that lay behind options within a single item, and cannot manage missing response appropriately. Some studies applied polytomous IRT models in educational testing data. Although the polytomous IRT models can estimate the differences among different incorrect response patterns, the item-responding process was still simplified, the multidimensionality among different options within a single item cannot be handled, and the missing data still cannot be managed appropriately. In recent years, IRTree models reported reasonable model-building processes that can be applied to testing data and manage missing data appropriately. Although designed for psychometric surveys and have not been applied in testing data yet, the stepwise modeling strategy of IRTree models can be applied to educational testing data. This study applied the IRTree model to educational testing data to evaluate its applicability, advantages, and disadvantages compared to binary and polytomous IRT models. The results indicated IRTree models are applicable to testing data with a large sample size, can estimate reasonable and interpretable item and person parameters, can provide researchers an insightful perspective to discuss the dimensionalities in individual item responding process, and can help to explore the dimensionality of missing responses. However, disadvantages of applying IRTree models to testing data also exist, such as low convergence rate, high dimensionality, parameters that are hard to be managed and interpreted, and uninterpretable information functions. The polytomous IRT models also reported unique advantages, such as stable parameter estimations, interpretable item and test information, item parameters that are easy to be interpreted, and direct evidence for optional design. Detailed results from IRTree models are reported, together with discussions about the advantages and disadvantages of IRTree models and suggestions to applied researchers. The limitations of this study and implications for future studies are presented. [The dissertation citations contained here are published with the permission of ProQuest LLC. Further reproduction is prohibited without permission. Copies of dissertations may be obtained by Telephone (800) 1-800-521-0600. Web page: http://www.proquest.com.bibliotheek.ehb.be/en-US/products/dissertations/individuals.shtml.]
Descriptors: Item Response Theory, Educational Testing, Data, Models, Comparative Analysis, Test Items, Responses
ProQuest LLC. 789 East Eisenhower Parkway, P.O. Box 1346, Ann Arbor, MI 48106. Tel: 800-521-0600; Web site: http://www.proquest.com.bibliotheek.ehb.be/en-US/products/dissertations/individuals.shtml
Publication Type: Dissertations/Theses - Doctoral Dissertations
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A
Grant or Contract Numbers: N/A
Author Affiliations: N/A