esmail mam sharifi; ali delavaran; azadeh boluki; somayeh shabani
Volume 2, Issue 7 , April 2012, , Pages 1-34
Abstract
Background: This study was administered with the aim of investigating the psychometric properties of the theoretical part of the driver’s license test. The study sample included the responses of a group of 350 subjects, selected through multi-cluster sampling, to the 30 randomly chosen theoretical ...
Read More
Background: This study was administered with the aim of investigating the psychometric properties of the theoretical part of the driver’s license test. The study sample included the responses of a group of 350 subjects, selected through multi-cluster sampling, to the 30 randomly chosen theoretical questions of the driver’s license test. In order to investigate the psychometric properties of the test questions, the results obtained from both the classical test theory and the item-response test theory were compared and evaluated. The study adopted a descriptive methodology and the accuracy of the sample was verified at the beginning. Factor analysis and Cronboch method were used to determine the unidimensionality of the test. Then the test questions were analyzed based on the classical and the item - response test theories and the question parameters (difficulty, discrimination and guessing) and the ability due to uncertainty were extracted using the simultaneous estimation method. Results: The results of the study confirmed the unidimensionality and independence of the test. After determining the main theory assumptions of the IRT, the model - data fitting was evaluated and the results of the two-parameter model showed better fitting with the data. In the next step, the question parameters and the ability factor were evaluated with the T-test. Results showed that there was no significant relationship in determining the accuracy of estimating the difficulty, slope and ability parameters between the classical theory and the item - response theory. In order to check the reliability and the stability of the test results on the first run, a test - retest was administered to a sample of 30 subjects. Since the present test is a kind of reference criterion, Kappa coefficient of reliability was used to settle the disagreements. Results showed that there is a significant relationship between the first run and the second run and moreover the test has sufficient reliability and validity to be administered in different cases. Conclusion: The analysis of the question parameters and subjects confirmed the simplicity of the test and its high capacity to distinguish the ability of the subjects. It can, therefore, be concluded that the test questions are more accurate with subjects with lower ability. In comparison to the classical theory, the estimated ability in the item - response theory is closer to the real rate. According to the estimated abilities, questions can be selected based on subjects’ abilities which can finally lead to the creation of a question bank.
