نوع مقاله : مقاله پژوهشی
نویسندگان
1 دانشجوی دکتری، گروه مطالعات برنامه درسی و سنجش و اندازه گیری، دانشگاه خوارزمی، تهران، ایران
2 دانشیار گروه مطالعات برنامه درسی و سنجش و اندازه گیری، دانشگاه خوارزمی، تهران، ایران
3 استادیار گروه مطالعات برنامه درسی و سنجش و اندازه گیری، دانشگاه خوارزمی، تهران، ایران
چکیده
کلیدواژهها
عنوان مقاله [English]
نویسندگان [English]
The aim of the present study was to examine and compare the accuracy, stability, and efficiency of test equating methods based on Classical Test Theory and Item Response Theory in the TIMSS 2023 science assessment at the fourth-grade level. This study was designed as a descriptive–psychometric, applied, quantitative, and cross-sectional investigation. The statistical population consisted of fourth-grade students from all countries participating in TIMSS 2023, and purposeful sampling was conducted across four countries with different achievement levels: Singapore as the highest-performing country, South Africa as the lowest-performing country, the Czech Republic as above average, and Iran as below average, in order to evaluate equating methods under diverse educational conditions. The data included raw scores and background information released by the TIMSS International Study Center, and data processing and equating procedures were conducted using R software version 4.5.2 with specialized CTT and IRT packages. Analyses were carried out within a digital testing environment using a common-item nonequivalent groups design with validated anchor items. Within the CTT framework, the equipercentile method with log-linear smoothing demonstrated the lowest RMSD and SEE values and the highest stability. Within the IRT framework, the two-parameter model was examined using mean–mean, mean–sigma, Haebara, and Stocking–Lord methods, among which the mean–mean method produced the lowest RMSD and SEE values and minimal changes in discrimination and difficulty parameters. The implementation of NEAT equating enabled accurate common-scale construction and reduced gender and regional bias to ensure valid international score comparisons.
کلیدواژهها [English]
)