Reliability of the Currently Administered Language Tests in Bangladesh: A Case Study

Abstract

This study aims at investigating the reliability of the language tests presently in use in Bangladesh, taking the Higher Secondary reading and writing tests as sample. To achieve this purpose, the English question papers of the H.S.C examinations of the last few years have been analyzed with a view to determining the extent to which the test formats used can affect candidates’ scores, and involve subjectivity in scoring. Besides, the marking scheme provided to the markers by the education board has been examined in order to find out whether and how far it is able to reduce the element of subjective judgement involved in the assessment of the responses  and thus to ensure inter-marker and intra-marker reliability of the tests. Furthermore, an empirical survey has been carried out in the form of questionnaires among the English language teachers of a few colleges in the country to collect quantitative data on various aspects of reliability of the H.S.C language testing. The study reveals that the Higher Secondary reading and writing tests, as a whole, are far below the satisfactory level in terms of reliability as they have deficiency in both test reliability and scorer reliability. Therefore, the scores produced by these tests cannot be considered reliable indicators of the test takers’ reading and writing abilities. The shortcomings of these tests, as identified in the investigation, include poor test items, inadequate test instruction, ambiguous marking criteria, insufficient marking guideline, etc. On the basis of the findings, some recommendations have been made for the improvement of the overall reliability of the H.S.C reading and writing tests.

Key Words: Language test, test reliability, scorer reliability, subjectivity, quantitative, marking criteria.