A Study On The Difficulty Of Chinese-to-english Translation Tests In CATTI

This thesis focuses on the difficulty of Chinese-to-English translation tests in CATTI. Knowing the difficulty of a translation test is important in translation accreditation, research and pedagogy. It is conducive to the selection of test materials, the design of test requirements, researchers’ understanding of difficulty in order to improve measures for translation difficulty. It is also helpful to translation teachers who are constantly in need of scientific and objective ways to examine their teaching achievements and evaluating students’ translation competences. However, very few studies have focused on difficulty of C-E translation, particularly in a translation accreditation. Therefore, it is necessary to explore and compare the difficulty of CATTI Level 2 and Level 3 C-E tests.This study adopts Meshkati’s human mental workload model as its major theoretical basis, and conducts an experiment and survey to explore the following three research questions concerning translation difficulty. First, how is CATTI Level 2 C-E test different from Level 3 in terms of students’ mental workload? Second, how is Level 2 C-E test different from Level 3 in terms of students’ translation performance? Third, what factors may cause the differences in translation difficulty?By conducting simulation translation tests on and distributing mental workload rating scales to 60 postgraduates majoring in Translation, the study attempts to identify students’ mental workloads in completing the two levels of C-E tests of May 2005. Two professional teachers have been invited to grade students’ translations based on a reliable scoring criterion so that the researcher can analyze how the two tests are different in terms of students’ translation quality scores and translation speed. The study also discusses the reasons for the differences by means of readability tool and error analysis.The study has the following major findings. First, based on the results of rating scales, students’ mental workload in completing Level 3 C-E test is significantly higher than that in Level 2, indicating greater overall translation difficulty of Level 3. Second, students’ translation speeds in the two tests show no significant difference, indicating insignificant temporal demands in the two tests. Third, students’ translation quality scores in the two tests are significantly different from each other. Fourth, grammar, syntax, lexical usage and mistranslation are major causes of translation errors in both tests, representing students’ inadequacy in the corresponding translation competences and specific translation difficulty in the tests. Moreover, the fact that some categories of translation errors are more frequent in one test than the other may indicate different focus of the two tests. This is consistent with students’ feedback from post-translation questionnaires and the three indicators of readability analysis(difficult words, simple sentence ratio and content word frequency in logarithmic). Finally, the readability analysis shows that overall text difficulty of Level 3 is greater than that of Level 2. Sixth, in addition to the answers to three research questions, the study identifies inconsistence between the test syllabus and the real test implementation, i.e. the speed requirements in the two tests cannot be implemented in the tests in that E-C and C-E sections are put in one paper, enabling the test takers to allocate time based on their own habits.According to CATTI’s requirement, Level 2 is intended to be more difficult than Level 3. However, the results of this study show the opposite trend. The major reason for the differences is that the difficulty of a translation test is such a complex term that it is not determined by a single factor. The correlations between various indicators and translation difficulty have not be fully recognized. In addition, the difficulty of a translation test may be subject to variables other than text difficulty and translator competences, such as the test environmental, time requirement, etc. Moreover, this study only selects test papers from May 2005 and conducted an experiment on a specific group of postgraduates. The results might be varied when test materials are different. Thus, it is necessary to further explore the construct of translation difficulty and conduct more empirical studies on CATTI.