Item Analysis of Short Test in Educational Testing:Comparative Study on Parameter and Non-parameter Item Response Theory
Yuan Shuli
Abstract
Yuan Shuli
Abstract
As one of the significant types of tests, the test project and short test are popular in educational testing. Parameter and non-parameter item response theory being the starts, these tests were under analysis. Compared was the geography paper in inaugurated arts taken by some senior three students. During this comparison the Rasch and Mokken model were respectively selected. For analyzing software Winsteps and Xcalibre were utilized to analyze item parameters in Rasch model. Analyzed in detail were the parameters of difficulty, differential item functioning and information curve. Software MSP was for the purpose of analyzing items in Mokken model. Besides, the statistics of accurate rate and coefficients of homogeneity were also analyzed in detail. Finally, four conclusions were arrived at as the following:( 1 ) The estimate results of difficulty between non-parameter and parameter item response theory were equivalent. ( 2 ) Those items, which failed to fit parameter item response theory, succeeded in non-parameter item response theory. ( 3 ) Non-parameter item response theory is more rigorous than parameter item response theory in dimension testing and item screening. ( 4 ) The result was equivalent in the detection for differential item functioning between non-parameter and parameter item response theory.
A significance statement is not available in the OpenAlex record.
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.
As one of the significant types of tests, the test project and short test are popular in educational testing. Parameter and non-parameter item response theory being the starts, these tests were under analysis. Compared was the geography paper in inaugurated arts taken by some senior three students. During this comparison the Rasch and Mokken model were respectively selected. For analyzing software Winsteps and Xcalibre were utilized to analyze item parameters in Rasch model. Analyzed in detail were the parameters of difficulty, differential item functioning and information curve. Software MSP was for the purpose of analyzing items in Mokken model. Besides, the statistics of accurate rate and coefficients of homogeneity were also analyzed in detail. Finally, four conclusions were arrived at as the following:( 1 ) The estimate results of difficulty between non-parameter and parameter item response theory were equivalent. ( 2 ) Those items, which failed to fit parameter item response theory, succeeded in non-parameter item response theory. ( 3 ) Non-parameter item response theory is more rigorous than parameter item response theory in dimension testing and item screening. ( 4 ) The result was equivalent in the detection for differential item functioning between non-parameter and parameter item response theory.
Key concepts: Item response theory, Differential item functioning, Rasch model, Classical test theory, Computerized adaptive testing, Test theory, Statistics, Polytomous Rasch model