Psychometric Properties for Selected Forms of International Computer Driving License Tests in Jordan, and Equating Their Scores

Abstract

This study aimed at investigating the psychometric
properties of a selected sample of ICDL tests in Jordan, and equating
these tests. To achieve this aim, the concept of equating tests was
defined, the required conditions for conducting the equating process
were identified, and the design group was explained. In addition, the
methods used in equating, were presented. Linear and equipercentile
horizontal methods of equating in classical theory and the Rasch
model in item response theory (IRT) were used. The psychometric
properties (difficulty and discrimination parameters) were presented
for the three forms of the seven subtests, which consisted of basic
concepts of information technology using the computer and
managing files, word-processing, spreadsheets, presentation, and
information and communication. With 60 items in the first one and
36 items in each of other subtests, the results of analysis of form
properties indicated that their scores should be equated. The sample
of study consisted of 1200 examinees for the subtest selected by
systematic randomization, distributed into three equale groups
randomly. The item parameters and ability parameters were
estimated, and the relative effectiveness was discussed for each
subtest and all models, depending on both standard error and cross
validation to judge the efficiency of equation. The results show that
the equation according to the Rasch model is better and more
effective according to the previous indicators, followed by the linear
and equipercentile respectively.