一因子高層次試題反應理論模式之評估

An Evaluation of the One-Factor Hierarchical Item Response Theory Model

郭伯臣(Bor-Chen Kuo) ； 謝典佑(Tien-Yu Hsieh) ； 吳慧珉(Huey-Min Wu) ； 林佳樺(Chia-Hua Lin)

《測驗學刊》 59卷3期 (2012/09) Pp. 329-348

https://doi.org/10.7108/PT.201209.0329

摘要

本研究主要是探討，當資料具備一因子高層次試題反應理論模式結構時，誤用單向度試題反應理論模式估計高層次量尺分數，或多向度試題反應理論模式估計次級量尺分數，對於參數估計所造成的影響。模擬研究考慮受試者樣本數、試題數、次級量尺間相關程度，以及題間與題內多向度測驗，結果顯示：1. 當資料具備一因子高層次試題反應理論模式結構時，一因子高層次試題反應理論模式於試題參數、高層次量尺分數與變異數的估計效能，優於單向度試題反應理論模式。2. 當資料具備一因子高層次試題反應理論模式結構時，一因子高層次試題反應理論模式於試題參數、次級量尺分數與變異數－共變異數矩陣的估計效能，近似於多向度試題反應理論模式。

關鍵字

一因子高層次試題反應理論模式 ； 多向度試題反應理論模式 ； 單向度試題反應理論模式

並列摘要

In this study, some simulation experiments were conducted to evaluate how the abilities, regressions and item parameters are affected by means comparing the accuracy of one-factor hierarchical item response theory model to that of one-dimensional and multi-dimensional item response theory models. There were four different factors such as sample sizes, item lengths, correlation between domain abilities, and model specifications considered in this study. The results showed that:1. With hierarchical data structure, the accuracy of one-factor hierarchical item response theory model outperformed that of uni-dimensional item response theory model for estimating the variance of the overall ability, the overall ability, and item parameter.2. With hierarchical data structure, the accuracy of multi-dimensional item response theory model was similar to that of one-factor hierarchical item response theory model for estimating the variance-covariance between domain abilities, domain abilities, and item parameter.

並列關鍵字

multi-dimensional item response theory ； one-factor hierarchical item response theory model ； uni-dimensional item response theory

參考文獻

Adams, R. J.,Wilson, M.,Wang, W. C.(1997).The multidimensional random coefficients multinomial logit model.Applied Psychological Measurement.21,1-23.

Baker, F. B.(2004).Item response theory: Parameter estimation techniques.New York, NY:Marcel Dekker.

Brooks, S. P.,Gelman, A.(1998).General methods for monitoring convergence of iterative simulations.Journal of Computational and Graphical Statistics.7,343-455.

de la Torre, J,Hong, H.(2010).Parameter estimation with small sample size: A higher-order IRT model approach.Applied Psychological Measurement.34(4),267-285.

de la Torre, J,Song, H.(2009).Simultaneous estimation of overall and domain abilities: A higher-order IRT model approach.Applied Psychological Measurement.33(8),620-639.