Application of Linear Logistic Test Model to Develop Mathematics Problem Solving Test

• Main Authors: Wang, Chia-Wen, 王佳文
• Other Authors: Wu, Yui-Yih
• Format: Others
• Language: zh-TW
• Published: 1995
• Online Access: http://ndltd.ncl.edu.tw/handle/30332784977723972764

碩士 === 國立台南師範學院 === 初等教育學系 === 83 === 　　The purpose of this study is to investigate the applicability and potential power of the linear logistic model (LLTM) for Mathematics problem solving problem test (MPST) development. Six hundred and twenty seven 6th graders were tested by the MPST. LINLOG, ITEMAN, ASCAL, MULTILOG and BIGSTEPS programs were used for the item and test analysis. The major findings are summarized as the followings. 　　1.Compare with the conventional test model or Rasch''s simple test model, the LLTM is more diagnostic and individual. 　　2.The test can logically classify the students into 5 subgroups according to the components mastery level pattern. In other words, the MPST can clearly differentiate the amount and components of help needed to reach mastery level for each subgroup. 　　3.The incremental information of multiple component IRT test scoring is substantially significant.