An Empirical Study On The Correlation Between Translation Test Items, Scoring Criteria And Translation Competence Testing

In recent years, with the increasing social demand for the English learner's communicative ability and bilingual competence, translation test has been widely employed in current large-scale, nation-wide English tests. To name a few, the former optional translation part of CET-4 has now become a compulsory part; the proportion of translation part of TEM 8 has risen from 15% to 20%; and for the English Test of Postgraduate Entrance Examination, translation has been a retained part. Nevertheless, have the widely used translation tests fulfilled the task of effectively testing the candidates'translation competence? And how to design the translation tests so as to attain the goal of testing translation competence? To figure out these problems means a lot to today's English teaching and translation tests.Based on the comparison and analysis of the translation tests in six current large-scale, nation-wide English tests, from the perspective of test format, test item and scoring criteria, the author finds the following weaknesses in them: (1) Paragraph translation is the test item most frequently used in these tests, which proved excessively integrative and subjective; (2) Scoring criteria for subjective items tend to be vague, and the primary grading method is error deduction marking scheme or the holistic grading, which may inevitably reduce the test's reliability and validity.In light of these weaknesses in translation test, the author studied Bachman's Communicative Language Testing and the latest Translation Competence research findings, and initially verified that the translation competence can be decomposed into sub-competences, including source language (SL) competence, translation strategic competence, and target language (TL) competence, which are in complementary relation and function as a whole in translation process. Then, the author put forward principles of translation test design, and accordingly, designed an experimental translation paper with six test items: idiom translation items, choice of good translations, gap-filling items, paragraph translation, analysis/assessment items and version-correction items, and also the related scoring criteria as well.To testify its reliability and validity so as to further verify the correlation between translation test items, scoring criteria and translation competence testing, the author carried out an experimental translation test on 50 English major seniors of Shandong Normal University, shortly after a pilot test, which helped the author make some revisions. Then the papers were collected, scored and analyzed by both quantitative and qualitative approaches. The analysis of the reliability and validity (including the face, content and construct validity) of the translation test items showed that the experimental translation test had reached a satisfactory effect to test the translation competence and had achieved the aim of TEM8 Syllabus.The effectiveness of the experimental translation test suggests that the principles of translation test design put forward by the author may provide clues to the writing of translation items and scoring criteria, and hopefully promote the future translation testing to be more scientific, standardized and effective.