Development Of The Computerized Adaptive Testing Program For O-NET At The Grade Nine Level

Abstract

This research aimed to analyze quality of O-NET examination, prepare Item bank of
O-NET, and develop a computerized adaptive testing system of O-NET. The examination
was provided for grade 9 level including 8 learning substance groups. The computerized
adaptive testing system was generated as a web application. Moreover, the quality of
examination was assessed by using Item Response Theory or IRT with 3 parameters and
Xcalibre Version 4.1.7. The methodologies of this paper were divided into 4 steps as follows:
1) Analyze the quality of O-NET examination for 8 learning substance groups, 2) Prepare the
test bank of O-NET for 8 learning substance groups, 3) Develop the computerized adaptive
testing system, and 4) Evaluate the efficiency of computerized adaptive testing system. The
samples consisted of 30 grade 9 level of Angsilapittayakom School, Chonburi Province. The
research tools included 1) O-NET examination for 8 learning substance groups, and 2) the
computerized adaptive testing system run on a website://www.onetcat.net/onetm3. Then
the students’ satisfaction was assessed by conducting fundamental statistics such as mean
and standard deviation.
The results were represented below.
1. The quality analysis of O-NET examination for 8 learning substance groups
showed that the average difficulty of O-NET examination (b) for grade 9 level was quite high.
2. The Item bank of O-NET was able to contain 4 multiple choices and unlimited
amount depending on size of server. The system also stored 469 items of the O-NET
examination which were analyzed regarding Item Response Theory with 3 parameters. These
items met the selection criteria of all 8 learning substance groups for grade 9 level as well.
3. The efficiency of computerized adaptive testing system of O-NET for grade 9
level was at good level. Besides, it was accepted by experts and samples who had utilized
the program, both general attributes and convenience of usage.