Testing the Test Using Rasch Measurement Model: Can Expert Vetting Ensure Quality of Test Items?

Arunodaya, Barman and Ahmad Fuad, Abdul Rahim and Md. Anayet, Ullah (2016) Testing the Test Using Rasch Measurement Model: Can Expert Vetting Ensure Quality of Test Items? International Medical Journal, 23 (1). pp. 55-57. ISSN 1341-2051

	Text FH02-FP-16-05586.pdf Restricted to Registered users only Download (723kB) | Request a copy
	Image FH02-FP-16-06917.jpg Restricted to Registered users only Download (100kB) | Request a copy

Abstract

Introduction: Item analysis using Rasch measurement model provides a way of measuring the quality of questions, assess how appropriate they are for the candidates and how well they measure their ability. Methods: Multiple choice questions generated by the subject specialists and vetted at departmental and central levels were used in assessing undergraduate medical students. Twelve questions each with five true/false responses were identified from the MCQ test items used to assess the diagnostic ability of the students. One hundred undergraduate medical students who have responded to these items were randomly selected. Scores of hundred students on the 60 items were analysed using computer software BIAS. Results: Ability of the students ranges from -.561 to 2.509 and difficulty of the items ranges from -2.948 to 2.293. Twenty one items were below the ability level of the students (-.561) and ten of them were providing very little information about the students ability. The overall person reliability was .69 and separation index was 1.50 whereas item reliability was .95 and separation index was 4.51. Conclusion: Twelve questions each with 5 true/false responses are well developed to assess the diagnostic ability of the students except ten off target responses those may be rewritten or excluded from future test.

Actions (login required)

View Item