A Study On Content Validity And Response Validity Of Proofreading Tasks In NMET From 2016 To 2019

In recent years,most studies on proofreading task are conducted based on test-taking,while few are analyzed based on language testing.Content validity and response validity are two important aspects to evaluate the quality of testing,but few studies research the proofreading task from the two aspects.Therefore,it is necessary to research the test design of proofreading task covering both of them.The present study will analyze the proofreading tasks in NMET from 2016 to 2019.Firstly,the characteristics of text input are analyzed based on Bachman and Palmer' s Test Method Facet.Secondly,the testing-point validity is analyzed based on Li Xiaoju' s testing-point validity theory.Thirdly,the response validity is analyzed based on proofreading test and retrospective study.The present study aims to answer the following three questions:(1)What are the characteristics of text input of proofreading tasks in NMET from2016 to 2019?(2)How is the testing-point validity of proofreading tasks in NMET from 2016 to2019?(3)How is the response validity of proofreading tasks in NMET from 2016 to 2019?Data analysis indicates that the validity of the proofreading tasks in NMET from2016 to 2019 is moderate,but there still exist some problems.Firstly,in terms of text input,the length fluctuates around 100 words required by the test syllabus.Among them,2016? and 2017 ? differ greatly from the required 100 words,and the three proofreading tasks in 2018 meet the requirement of the test syllabus best.There is a lack of variety in topic and genre,too.The topics only involve People and Self,and the genres only involve argumentation and narration.In terms of the three influencing factors of difficulty,the overall readability is at the simple level,the number of new words is small,and the simple sentences are more than complex compound sentences.So the overall difficulty is low.Therefore,for the text input,the content validity of the proofreading tasks in NMET from2016 to 2019 need to be improved.Secondly,in terms of testing-point validity,testing points of proofreading tasks are abundant and extensive,which contain lexical,syntactic and discourse testing points.However,the testing-point levels are unevenly distributed.The sentence level testing points account for 63.33%,which are the most frequently examined testing point.The phrase level testing points are next,accounting for 25%.The discourse level testing points are the least,accounting for 11.47%.In terms of testing-point factor,the grammar factors are far more than meaning factors.Among them,the distribution of testing points is more reasonable in 2016 ?,and it tests more at discourse testing-point level and meaning factor,so it has higher testing-point validity.However,the testing-point validity of the proofreading tasks in NMET from 2016 to 2019 need to be improved.Thirdly,in terms of response validity,the results of the proofreading tests show that the situation that the high school students correct the errors from the expected testing-point level accounts for 81.33%,while the situation that subjects correct the errors from the unexpected testing-point level accounts for 18.67%.According to the analysis of retrospective verbal reports,the errors that corrected from unexpected testing-point level mainly involve frequently examined testing points,such as articles,singular or plural form of noun,or attributive clause.For these errors,candidates can correct them by guessing or doing lots of practice,which might mislead the candidates to develop the bad answering habit and influence the response validity.Basing on the result mentioned above,this study puts forward some suggestions for test designers and teachers.Firstly,the test designers should consider the diversity of the topic and genre,and control the difficulty in order to ensure the discrimination.Secondly,the discourse level testing points and examination of meaning should be increased in order to improve the testing-point validity.Thirdly,test designers should avoid repeatedly examining a certain testing point and improve the comprehensiveness of testing point in order to improve both content validity and response validity.For the teachers,the teaching of proofreading tasks should focus on cultivating discourse ability rather than only the answering strategies or test wiseness.