Abstract

The purpose of this study was to evaluate the population invariance of equating functions across country subgroups in TIMSS 2015 mathematics tests in relation to the raw-score distribution, DIF, and DTF. We used equipercentile and IRT observed-score equating methods. The results of the study indicate that there is a relationship between the population invariance and raw-score and skewness differences for both equating methods. The association between the population invariance and the number of DIF items is explicit in the equipercentile equating but not clear in the IRT equating. Furthermore, we have observed that when the number of DIF items is high, if the DTF value is relatively low, the equating errors tend to be moderate to low.

摘要

本研究的目的是评估 TIMSS 2015 数学测试中与原始分数分布、DIF 和 DTF 相关的等同函数在国家/地区子组中的总体不变性。我们使用等百分位数和 IRT 观察分数等式法。研究结果表明，两种等式方法的总体不变性与原始分数和偏度差异之间存在关系。总体不变性和 DIF 项目数量之间的关联在等百分位数等式中是明确的，但在 IRT 等式中不明确。此外，我们观察到，当 DIF 项的数量较多时，如果 DTF 值相对较低，则等式误差往往是中等到较低的。