Open AccessCalibrating PIRLS Test in Sultanate of Oman Using Item Response Theory
Author(s) -
Amal K. Al-zaabi,
Abdulhameed S. Hassan,
Rashid Al-Mehrzi
Publication year - 2019
Publication title -
mağallaẗ al-dirāsāt al-tarbawiyyaẗ wa-al-nafsiyyaẗ/journal of educational and psychological studies
Language(s) - English
Resource type - Journals
eISSN - 2521-7046
pISSN - 2218-6506
DOI - 10.24200/jeps.vol13iss3pp496-515
Subject(s) - item response theory , polytomous rasch model , test (biology) , psychology , sample (material) , local independence , psychometrics , item analysis , statistics , social psychology , mathematics education , mathematics , clinical psychology , paleontology , chemistry , chromatography , biology
The study aimed to calibrate Oman data of the PIRLS test using the graded response model and to examine the psychometric properties of it, as well as identify the fit and unfit of its items. PIRLS2011 test booklets were used, which consisted of 146 test items (74 dichotomous and 72 polytomous). Items were divided into 13 booklets; each with two blocks (one literary and one informational). PIRLS test booklets were administered to 13 groups of fourth grade students in Sultanate of Oman with a total sample of 10394 students. Assumptions of IRT (unidimensionality and local independence) were examined and supported. Also, item fit was examined and supported using Samejima’s graded response model. The data was analyzed by Multilog7.03 program to estimate both item and ability parameters. Results indicated that the assumptions of IRT were proved. Also, IRT analysis revealed that 8 items showed unfit which represents only 5% of the test items. So, this result confirms that the test has good psychometric properties under the IRT.