Item Characteristics on Pro-TEFL Listening Section

Lina Lina, D Mardapi, Haryanto Haryanto
2019 Proceedings of the First International Conference on Advances in Education, Humanities, and Language, ICEL 2019, Malang, Indonesia, 23-24 March 2019   unpublished
A good test should be proven theoretically and empirically. A test is considered as a good one if its items have appropriate content, language and construct based on theoretical analysis. Besides, the empirical analysis on the test items need to do in order to describe the quality of the test. Empirical item analysis can be done based on the framework of Classical Test Theory (CTT) and Item Response Theory (IRT). This study analyzed empirically the test item characteristics based on the 1
more » ... ter logistic (1-PL) IRT framework. The items were analyzed based on the difficulty indices. Before analyzing the test items, assessment on IRT assumptions was done; unidimensionality and local independence. The data were gathered through testing. The responses from 334 test takers on 50 items of listening section on Pro-TEFL test administered by the Center for Language Development were used as empirical data. The assessment on IRT assumptions (unidimensionality and local independence) was done by using Factor Analysis which was assisted by SPSS for windows program. The item parameter (the difficulty index) was estimated by using Program R. Results showed that the assessment on IRT assumptions can be fulfilled. Then, the analysis of item characteristics was done to estimate the item parameter based on the 1-PL IRT model. Based on the analysis on the difficulty index, there were 2 items that were considered as poor items because the difficulty indices are more than +2 and they are considered as too difficult items.
doi:10.4108/eai.11-7-2019.159630 fatcat:nydnegq62zhzbaljvsyr4fpgue