A Construct Validation Study Of Grammar Gap-filling Task In NMET Guangdong

This thesis is about the validation study of grammar gap-filling(G/F)task in NMET(National Matriculation English Test)Guangdong.The grammar G/F test has been put into use in NMET since 2014,but as a matter of fact,it is in NMET Guangdong that grammar G/F task has made its first appearance.In 2007,grammar G/F task appeared in NMET Guangdong for the first time to substitute the traditional multiple-choice(MC)grammar test.Since the emergence of traditional MC grammar test in NMET,its shortcomings become increasingly evident,i.e.many students answer the items resorting to test-wiseness;students learn grammatical knowledge just by reciting the exemplified sentence or rules,students fail to apply the grammatical knowledge into the specific context.Whereas,the grammar G/F task offers test takers a passage with a number of words deleted.Test takers are required to fill the gaps by themselves without any option to choose,the word they fill must be grammatically correct and make the text complete and meaningful.So the intention of test designers designing the grammatical G/F task is to measure students' grammatical competence including form and meaning at the sentential or discourse level.It is expected that the grammatical G/F task can activate students' grammatical knowledge and encourage them to learn to use the grammatical rules in context.But in the actual implementation,does the task can really measure what it is supposed to measure?This issue needs further research.At the same time,since the task has been put into use not only in Guangdong province,but in the whole nation.So the validation study of this task is also quite necessary.The present study combines quantitative and qualitative approach to validate the validity of grammar G/F task in NMET Guangdong.On one hand,the author conducted an interview with one test designer of the task,finding out the construct of grammar G/F task.On the basis of the interview,questionnaires for high school English teachers are constructed to investigate the high school English teachers' opinion about what the grammatical G/F has measured.On the other hand,this studyinvestigates the test takers' answering process by adopting retrospective interview,after coding and matching participants' answering performance with the expected performance,it aims to find out whether test takers' answering process can reflect the abilities the test designer expects to measure.The results of the questionnaire demonstrate that,overall speaking,high school English teachers agree with the statements in the questionnaire concerning the construct of the grammar G/F task.As for participants' answering process,it is found out that,in most cases,participants answer the item correctly due to the grasp of the grammatical knowledge,while fail to answer the item because of lacking of the underlying abilities.So participants' answering process is consistent with the expected answering process.Generally speaking,the high school English teachers' opinion about the construct of grammar G/F task agree with the test designers' intention.At the same time,participants' answering process also reveal that they do use the abilities supposed to be measured by the grammar G/F task.Therefore,we can conclude that grammar G/F task has a good level of construct validity.