A Comparison of Multiple-Choice and Constructed-Response Tests for the Assessment of Translation Ability

• Main Authors: Chih-wei Tseng, 曾志偉
• Other Authors: Chun-mei Shieh
• Format: Others
• Language: en_US
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/01105585208437183973

碩士 === 國立屏東商業技術學院 === 應用外語系 === 96 === Recently, translation has been receiving a great attention in institutions of higher education. Such institutions are reorganizing to offer required and elective courses designed both to improve the students’ language proficiency and to enhance professional training for translation. As a result, translation is now a test item for the entrance examination of both universities and graduate schools. However, translation ability is currently evaluated by multiple-choice questions rather than constructed-response questions in some college or graduate school entrance exams. Much research has shown that multiple-choice test scores are comparatively higher than constructed-response test scores. It is skeptical that the examinees’ translation ability can be accurately evaluated simply by a few multiple-choice questions for which there are limited and fixed answers. The purpose of this study aims at not only exploring whether the average score of the multiple-choice test is significantly different from the average score of the equivalent constructed-response test, but also investigating the difference between the strategies used in the two forms of tests. A total of 411 participants of Department of Foreign Language were recruited in this study. The results of the study show that there is a significant difference between the two forms of tests. Meanwhile, the strategies used in answering translation question are significantly different. Consequently, the form of the test will have a significant influence on the subject’s translation performance. In other words, a learner’s ability to translate from English to Chinese could not be measured equally by multiple-choice test and constructed-response test. It could be inferred from the study that a learner’s translation capability could not be effectively assessed and verified simply using multiple-choice tests.