Item Bias Detection: Classical Approaches
Gideon J. Mellenbergh
Abstract
Gideon J. Mellenbergh
Abstract
Abstract Under the assumptions of Classical Test Theory (CTT), a person's observed test score is an estimate of his (her) true test score. Therefore, CTT‐based methods to detect differentially functioning items use the test score to equate members of different groups. An item is considered to be differentially functioning between two groups when item responses differ between members of the two groups at the same test score. Five methods for the detection of differentially functioning dichotomous item responses and their generalization to ordinal‐polytomous item responses are discussed.
OpenAlex reports 1 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Abstract Under the assumptions of Classical Test Theory (CTT), a person's observed test score is an estimate of his (her) true test score. Therefore, CTT‐based methods to detect differentially functioning items use the test score to equate members of different groups. An item is considered to be differentially functioning between two groups when item responses differ between members of the two groups at the same test score. Five methods for the detection of differentially functioning dichotomous item responses and their generalization to ordinal‐polytomous item responses are discussed.
Key concepts: Polytomous Rasch model, Generalization, Test (biology), Item response theory, Psychology, Classical test theory, Test score, Statistics