Feasibility of Applying Classical Test Theory in Testing Reliability of Student Assessment

Barman, Arunodaya (2011) Feasibility of Applying Classical Test Theory in Testing Reliability of Student Assessment. International Medical Journal, Japan, 18 (2). pp. 110-113. ISSN 1341-2051

[img] Text (1. Feasibility of Applying Classical Test Theory in Testing Reliability of Student Assessment)
feasibility-applying-classical-test-theory-testing-reliability-student-assessment - Published Version
Restricted to Registered users only until June 2020.

Download (36Kb)
Official URL: http://connection.ebscohost.com/c/articles/6304576...

Abstract

Introduction: Answer to the curiosity of test takers about the appropriateness of their students' scores is the reliability test. Sources of error variance which is one of the elements of reliability test may be random error, specific factor error, transient error and sample error. Keeping in mind the sources of error, different methods of estimating reliability coefficient such as estimation of coefficient of internal consistency, coefficient of stability, coefficient of equivalence and coefficient of equivalence and stability are suggested in classical test theory. Purpose of this manuscript is to highlight the feasibility of applying these methods in student assessment. Methods: Database pubmed.org and Ebscohost were browsed with key words "Reliability in student assessment" and "sources of error in reliability coefficient". All relevant articles were identified and searched for relevant findings. Later cross references related to source of error, reliability tests (classical test theory) and the test procedures were searched out and cited in the manuscript. Results: There are multiple sources of error influencing reliability and these are random error, specific factor error, transient error and sample error. Different reliability test are advised considering these sources of error. None of these reliability tests is interchangeable. Only reliability coefficient of stability and equivalences considers the all sources of error. Conclusion: Applying the classical test reliability approaches for complete reliability assessment is nearly impracticable in student assessment other than in research situation. Generalizability theory; G-study and D- study, and Rasch measurement model may be tried to overcome this feasibility problems.

Item Type: Article
Subjects: R Medicine > R Medicine (General)
Faculty / Institute: Faculty of Medicine
Depositing User: Prof Arunodaya Barman
Date Deposited: 08 Feb 2015 03:35
Last Modified: 28 Apr 2015 06:35
URI: http://erep.unisza.edu.my/id/eprint/2053

Actions (login required)

View Item View Item