This study investigates the cognitive and metacognitive processes used by EFL test-takers in completing a compound dictation test through analyses of their verbal protocols obtained immediately after the test and thro...This study investigates the cognitive and metacognitive processes used by EFL test-takers in completing a compound dictation test through analyses of their verbal protocols obtained immediately after the test and through semi-structured retrospective interviews. The study explores relationships between the test-takers＇ actual performance and the instructions for the compound dictation test, the language abilities measured, any major construct-irrelevant factors affecting the test-taking processes, and the performance patterns of performance across test-takers of different overall proficiency levels. Findings are as follows. （1） Test-takers＇ actual performance does not apparently relate to the instructions given in the compound dictation test, which may raise doubts over the effectiveness of the instructions, thereby posing a potential threat to test validity. （2） Test-takers may pay more attention to the pronunciation of the words to be used to fill the blanks than to the actual meaning of the words, their difficulties in spelling and sentential expression possibly interfering with test performance even though they appear to use a variety of cognitive and metacognitive strategies throughout the test. （3） A number of construct-irrelevant factors posing a potential threat to test validity were found in the test-taking processes. Some of these factors are related to the test-takers, for example, memory capacity, attention failure and psychological factors, while others relate to the test, including speech rate and time constraints. （4） Test-takers in the study at all three proficiency levels appeared not to follow the instructions given in the compound dictation test. However, higher proficiency test-takers appeared to make more frequent use of cognitive and metacognitive strategies, and they appeared to be less influenced by construct-irrelevant factors.展开更多
基金part of the achievements funded by the Fundamental Research Funds for the Central Universities in China (No. 0205005201030)the National Scholarship Council of China (No. 2010850066)granted to the first author to support a one-year visiting professorship
文摘This study investigates the cognitive and metacognitive processes used by EFL test-takers in completing a compound dictation test through analyses of their verbal protocols obtained immediately after the test and through semi-structured retrospective interviews. The study explores relationships between the test-takers＇ actual performance and the instructions for the compound dictation test, the language abilities measured, any major construct-irrelevant factors affecting the test-taking processes, and the performance patterns of performance across test-takers of different overall proficiency levels. Findings are as follows. （1） Test-takers＇ actual performance does not apparently relate to the instructions given in the compound dictation test, which may raise doubts over the effectiveness of the instructions, thereby posing a potential threat to test validity. （2） Test-takers may pay more attention to the pronunciation of the words to be used to fill the blanks than to the actual meaning of the words, their difficulties in spelling and sentential expression possibly interfering with test performance even though they appear to use a variety of cognitive and metacognitive strategies throughout the test. （3） A number of construct-irrelevant factors posing a potential threat to test validity were found in the test-taking processes. Some of these factors are related to the test-takers, for example, memory capacity, attention failure and psychological factors, while others relate to the test, including speech rate and time constraints. （4） Test-takers in the study at all three proficiency levels appeared not to follow the instructions given in the compound dictation test. However, higher proficiency test-takers appeared to make more frequent use of cognitive and metacognitive strategies, and they appeared to be less influenced by construct-irrelevant factors.
