英语请求言语行为测试整体性和分析性评分的对比研究.pdfVIP

ABSTRACT Inthefieldoflanguagetesting,anumberofresearcheshavebeenconducteddealingwiththe argument between holistic and analytic rating rubrics for performance assessment, especially writing andspeakingtests.However,the same studycenteringon speechactassessmentisrarely found.Forthatreason,thecurrentstudyattemptstoexploreempiricallywhethertheapplicationof thesealternativescalesmaycauseanydifferenceinthescoresallocatedbythesameratersandthe way the raters rate the speech act texts. In this study, two university teacher raters (one female teacherandonemale)wereinvitedtoratetheWDCTtestwhichwasadministeredto30(13male and17female)ChineseEFLuniversitystudentsagedfrom19to22. Data analysis was handled in the present study from two aspects. Qualitative study was carried out through the analysis of the results of two questionnaires, which were developed to elicittheratersattitudesandperceptionstowardsholisticapproachandanalyticratingapproaches, rater training andthe selection ofrating scales. Statistical analysis andFACETS analysis, which presents estimatesofrater severityonanequal-intervalscaleaswell asfitstatistics,wereutilized to investigate rater reliability and severityregards to the two rating scales.From qualitative and quantitativeanalysis,thefollowingfindingscanberevealed:First,bothratingscaleswerereliable andcouldyieldreliable results,with analyticrating scaleslightlymorereliable thanholistic one. Secondly, when it comes to rater severity, the differences in severity estimates were more remarkable in the analytic ratings although both two approaches provided consistency and significantdifferencesin severitybetweenthetworaters.Thirdly,theraterswere foundtoexhibit certainbiasacrossthefourtraitsandcertainpreferenceto