Development of Computer Literacy Test as Computerized Adaptive Testing


ÖZBAŞI D. , Demirtasli N.

JOURNAL OF MEASUREMENT AND EVALUATION IN EDUCATION AND PSYCHOLOGY-EPOD, cilt.6, ss.218-237, 2015 (ESCI İndekslerine Giren Dergi) identifier

  • Cilt numarası: 6 Konu: 2
  • Basım Tarihi: 2015
  • Dergi Adı: JOURNAL OF MEASUREMENT AND EVALUATION IN EDUCATION AND PSYCHOLOGY-EPOD
  • Sayfa Sayıları: ss.218-237

Özet

The purpose of this study is to investigate the applicability of the Computer Adaptive Testing (CAT) of the exemption exam of Information and Communication Technology in computer environment (as CAT) given to the first year students at every faculty of Ankara University each year. The research was carried out in two basic stages. In the first stage, the researchers studied with 1366 university students to obtain the data to study on with CAT application and and to test the prepared items. In the second stage, a paper and pencil test was given to 142 university first year students with live CAT. The test of computer literacy was used as an instrument of data collection. It was also tested in the study if the collected data met the hypothesis of Item Response Theory (IRT). With this regard, a live CAT application was carried out with 136-item pool which was found to comply with the three-parameter logistic model. According to the findings of the study, the highest reliability estimate found in simulative CAT application was found in test termination condition depending on the fixed item number (with 30 items). Besides, with regards to the number of used item, the least item use happened when test termination condition is Standard Error (SE)<0.50. In the CAT test, students' ability estimates in CAT in which proficiency estimate is done depending on Maximum Likelihood Estimation (MLE) has come up with more reliable results in extreme values compared to those obtained in paper-pencil test, and lower standard error estimates were obtained with the use of CAT application with regards to standard error value. The average reliability obtained from live CAT application was found to be higher that of paper-pencil test. According to the findings of this study, when the SE<0.30 test termination rule is applied, MLE was found to be more reliable, when SE<0.50 and fixed item (30) test termination rule was applied, the proficiency estimate value based on Expected A Posteriori Method (EAP) was found to be more reliable. Besides, the test information amount obtained in live CAT application was significantly higher than that of paper and pencil test.