What Item Response Theory (IRT) Can Reveal to Us： An Analysis of a Twenty-Item Vocabulary and Structure Test
Title
What Item Response Theory (IRT) Can Reveal to Us： An Analysis of a Twenty-Item Vocabulary and Structure Test
What Item Response Theory (IRT) Can Reveal to Us： An Analysis of a Twenty-Item Vocabulary and Structure Test
Keywords: language test, scores, item response theory, interpret, revelation
Authors
Zhang Jianmin
Chen Zhiteng
Xiao Xi
Zhang Jianmin
Chen Zhiteng
Xiao Xi
Bio Data
Zhang Jianmin is an associate professor at the School of International Studies, Zhejiang University, P.R. China.
Chen Zhiteng is a senior teacher of English at Longshan High School, Ruian, Zhejiang Province, P.R.China.
Xiao Xi is a primary teacher of English at Lingxi No.2 Middle School, Cangnan, Zhejiang Province, P.R.China
Abstract
A language test serves two basic functions: 1) it tries to measure the true language ability of a student; 2) it aims to evaluate classroom teaching. Based on the results of an English test given to one class at a high school, this paper aims to answer two questions: 1) are the scores of some students from the test compatible with their regular performance in English? 2) is the test good enough to give us useful and reliable information about the test and the test takers? By comparing the test scores with those from other tests and by using the Rasch model, the authors find that IRT shows a very strong capacity in interpreting the test scores and predicting the language ability of the students in question.
A language test serves two basic functions: 1) it tries to measure the true language ability of a student; 2) it aims to evaluate classroom teaching. Based on the results of an English test given to one class at a high school, this paper aims to answer two questions: 1) are the scores of some students from the test compatible with their regular performance in English? 2) is the test good enough to give us useful and reliable information about the test and the test takers? By comparing the test scores with those from other tests and by using the Rasch model, the authors find that IRT shows a very strong capacity in interpreting the test scores and predicting the language ability of the students in question.
[private]
See pages: 30-40
[/private]Category: Monthly Editions
