A Multi-dimensional Comparison Of The Computer-based And The Pa Per-based English Majors’ Achievement Tests

Along with the fast development of computer and web technology, languagetesting witnesses an increasing change of testing media with the tendency ofComputer Based Language Test (CBLT) to replace Paper Based Language Test(PBLT). Many previous researches have been conducted on the equivalence betweenCBLT and PBLT, i.e. comparability study. However, most of the studies have beendone in foreign countries but few in the Chinese context and those studies haveyielded inconsistent conclusions. Given this insufficiency, this thesis attempts toexplore the equivalence of two modes of English majors’ achievement test.Based on the theory of classical true score theory, this research was conducted.According to the reform of English majors’ achievement test of a school in SichuanInternational Studies University,27seniors of English major as the computer-basedpilot test subjects are recruited in this research. This study tries to conduct amulti-dimensional comparative study between computer-based English majors’achievement test (CBEMAT) and paper-based English majors’ achievement test(PBEMAT) from six aspects: test paper preparation, cost of test administration,administration of test, scoring and score recording, rater reliability, and responsevalidity. Through test score and teachers’ and test-takers’ questionnaires, this studyinvestigates the equivalence between CBEMAT and PBEMAT. Quantitative andqualitative forms of data are gathered and SPSS16.0is employed to do the statisticalanalysis.The results are as follows:(1) By using test software platform on the computer,CBEMAT, compared with PBEMAT, is faster and convenience with less time andenergy put in.(2) In the long run, CBEMAT can reduce test administration costs.(3)The CBEMAT can not only save time and energy but also raise the efficiency ofadministrating test.(4) CBEMAT increases the effectiveness of the rating, as well asimproving the accuracy of marking. It is fairer and more objective during theprocedure of marking.(5) The total score and items’ scoring of inter-rater andintra-rater is relatively high. Thus, with a good consistency between two formats of test in rater reliability, it is ideal in rater reliability of CBEMAT and PBEMAT.(6)Test-takers are little affected by computer familiarity, and computer-based test modedo not have influence on test-takers’ test-taking strategies a lot. Thus, responsevalidity is quite satisfactory.