Communicative Language Testing Its Validity and Reliability at the Addis Ababa University
No Thumbnail Available
Date
1997-05
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Addis Ababa University
Abstract
The aim of this study was to try and see if the English language can be
tested communicatively at the university and if the test produced has the required
types of validity and reliability.
So a test was prepared and administered for 120 students in the freshman
programme. And after multiple marking was applied, the papers of 80 students
(2/3 of the total number of subjects) were included in the item analysis list.
Results of the item-analysis showed the test had items of varied and acceptable
facility values (average= 54.8) and indexes of discrimination.(average =0.36)
The reliability of the test was also checked with the help of two methods:
the Spearman-Brown split-half which gave r=0.79, and the Kuder - Richardson
20(KR 21) which gave r=0.8. KR 21 was used because it also takes into
consideration the interitem consistency in the test. The results obtained in both
showed the test could be accepted as a reliable instrument.
The results of the subjects were also correlated with their results in College
English - I to check the test's concurrent validity. The two tests showed a good
correlation at r=0.75.
Two questionnaires were also prepared: one for the students who took the
test and one for instructors. The results of the analysis of the responses of the
respondents showed the test had the different types of validity.
So it was concluded that communicative language testing can be
applicable in the tertiary level, and recommendations where made for test
Description
Keywords
Communicative Language Testing