英语测试学复习题.docx

英语测试学复习题.docx

《英语测试学复习题.docx》由会员分享，可在线阅读，更多相关《英语测试学复习题.docx（14页珍藏版）》请在冰豆网上搜索。

英语测试学复习题

基础英语测试学复习题

（一）英语测试的发展阶段

1Pre-scientificstage

Thisstagereferstotheperiodthatlastedto1950s.Atthattimetheforeignlanguageteachingputemphasisonthewrittenformratherthanonthecommunicativeaspectofthelanguage.Nospecialskillorexpertiseintestingwasrequired:

thesubjectivejudgmentoftheteacherwasconsideredtobeofparamountimportance.Testusuallyconsistedofessaywriting,translation,grammaticalanalysis（oftenintheformofcommentsaboutthelanguagebeinglearned）andsometimesoraltests.Spolskycallsthetraditionalapproachpre-scientificbecauseitpaidnoattentiontosuchmattersasreliabilityandobjectivityanddidn’tusestatisticalmethods.Itis,therefore,bettertocallthemmethodsofexaminingratherthantests.Spolskyalsocharacterizesthisstageasbelongingtothetraditionalthinking:

elitistandauthoritarian.Klein—Braleypointsoutthatthehighfacevalidityofthesetypesoftestsstilldoesnotprovethemtobetrulyvalid.Thesubjectivecharacteroftheassessmentmakesitdifficulttocomparetheresults.Thisraisesthedoubtsaboutthereliability,and,consequently,aboutthevalidityofthesetests.

2Psychometric--structurestage

Thedominanttestingforminthe1950sand1960swaspsychometric--structuralapproach.Undertheinfluencesofbehaviorismandstructuralism,languagelearningisinterpretedasastimulus--responseprocessandformationofhabitsthroughpositivereinforcements,andtheformsandstructurescouldbedividedintolayersandelements.

Americanlinguist,RobertLado,oneofthefirsttestingscholar,stressedontwopointswithanaimtoestablishanobjectiveaspectofmeasuringhumanlanguageproficiency:

First,testshouldtestlanguageusagesandnotknowledgeaboutlanguage;Second,thestructurestobetestedshouldbevalidstructuresincolloquiallanguageuse.Thestructuralsentence-basedviewoflanguagefittedinquitewellwiththepsychometricquestforsamplesofindividualelementstobetested.Thisresultedinstandardizedtestswithanemphasisondiscretepointitems,whichcanbefurtherdividedintosubtests.Thetestsarecharacterizedbytheconvictionthattestingcanbeobjective,precise,reliableandscientific.Themultiple-choiceformatwasregardedasasatisfactoryinstrumentinmeasuringtheproficiencyinthecomprehensionofwrittenEnglish（Spolsky1995;215）.Itisviewedasthefirststepintheestablishmentofmodernlanguagetestingprofession.Theconceptofreliability,validity,thedegreeofdifficultyanddifferentiationcouldthenbeappliedtothecompilation,practiceandanalysisoftests.

.3Psycholinguistic—sociolinguisticstage

Thisstageisaperiodthatcoversthe1960sand1970swhenNoamChomsky’slinguisticstheoryswayedthebasisofstructuralism（Bachman,1990:

296）.Thisinvolvesthetestingoflanguageincontextandisthusconcernedprimarilywithmeaningandthetotalcommunicativeeffectofthediscourse.Asearlyas1961,JohnB.Carrollemphasizedthat,inadditiontodiscretepointtests,integrativetestsshouldbeused,thesearetestswhicharenotsomuchaimedattestingseparateelementsbutmeasuringthethetotalcommunicativeeffectofanutterance.Theintegrativetestdoesnotseektoseparatethelanguageskillsintoneatdivisionstoimprovetestreliability,instead,theyareoftendesignedtoassessthelearner’sabilitytousetwoormoreskillssimultaneously.Carroll,Morrow,Olleralladvocatedtheintegrativetest.

Integrativetestisbestcharacterizedbytheuseofclosetestingandofdictation.Oralinterview,translationandessaywritingarealsoincludedinintegrativetests.

.4Communicativepragmaticstage

Thecommunicativepragmaticperiodbegantooccurcorrespondingwiththeglobalshiftsince1970stothecommunicativeaspectoflanguageintheinternationaleducationalcircle.Theriseofthesociolinguisticshasbroughtaboutashifttotheinterestfromlinguisticcompetence（knowledgeoftherulesofalanguage）tocommunicativecompetence（knowledgeoftherulesoflanguageuse）.TheshiftcanalsobeobservedintheFLT（foreignlanguageteaching）,wheremoreandmoreemphasisisputonactualperformanceinordinarysituations.

In1971,Candlinputforwardtheterminology“communicativelanguageteaching”withtheemphasisonthegoalofforeignlanguageteachingbeingenablingalearnertobeabletocommunicateinaforeignlanguage.Theaimofthelanguageteachingshouldbestatedintermsoffunctionratherthanform.Communicativeteachingasksforcommunicativetesting.Developmentsincommunicativetestingcan,therefore,beviewedasattemptstocometotermswiththecommunicativelanguageteachingmovement.Testing,ontheotherhand,requiresanexplicationofthetheoryoflanguage,oflanguageuseandoflanguagelearning.Attheendof1960s,Spolsky,Upshur,Jakobovitsputforwardamethodtomeasurethelearners’communicativecompetence.From1970s,manyappliedlinguistbegantostudythetheoryofcommunicativetestingandhavepresentedmanymodels.

Thecommunicativetestaimstomeasurecandidates’communicativecompetency.Communicativecompetencyisassociatednotonlywiththelinguisticcompetence,butalsowiththecommunicativecompetence.Canale&Swain（1980）arguethatcommunicativelanguagetesting,whilebeingconcernedwiththelearners’knowledgeabouttheformofthelanguageandhowtouseitappropriatelyincontextsofuse,mustdealwiththeextenttowhichthelearnerisactuallyabletodemonstratethisknowledgeinameaningfulcommunicationsituation.

Thecommunicativetestischaracterizedbythestrongemphasisonactualuseoflanguageinreal-lifesettingswhichhavetobebasedonthecommunicativeneedsofthelearners.Aresultofthisfortestingisthatthereisnoneedforoneoveralltestoflanguageproficiency,butratherforacollectionoftasksatdifferentlevels,whichhavetoresembleauthenticsituationsascloselyaspossible.

（二）测试的效度和信度

1Validity

Validityisthedegreetowhichatestmeasureswhatitissupposedtomeasure,orcanbeusedsuccessfullyforthepurposesforwhichitisintended.Everytest,whetheritbeashort,informalclassroomorapublicexamination,shouldbeasvalidastheconstructorcanmakeit.Anumberoftypesofvalidationareappliedtotests.Inthisbriefsurvey,weshalllimitourconsiderationtojustafewofthemostcommoncategories.

A:

Contentvalidity

Thiskindofvaliditydependsonacarefulanalysisofthelanguagebeingtestedandoftheparticularcourseobjectives.Thistestshouldbesoconstructedastocontainarepresentativesampleofthecourse,therelationshipbetweenthetestitemsandthecourseobjectivesalwaysbeingapparent.Whenembarkingontheconstructionofatest,thetestwritershouldfirstdrawupatableoftestspecificationsdescribinginveryclearandprecisetermstheparticularlanguageskillsandareastobeincludedinthetest.Thetestwritershouldattempttoqualifyandbalancethetestcomponents,assigningacertainvaluetoindicatetheimportanceofeachcomponentinrelationtotheothercomponentsinthetest.Inthisway,thetestcouldachievecontentvalidityandreflectthecomponentskillsandareaswhichthetestwriterwishestoincludeintheassessment.

B:

Constructvalidity

Ifatesthasconstructvalidity,itiscapableofmeasuringcertainspecificcharacteristicsinaccordancewithatheoryoflanguagebehaviorandlearning.Thistypeofvalidityassumestheexistenceofcertainlearningtheoriesorconstructsunderlyingtheacquisitionofabilitiesandskills.Forexample,itcanbearguedthataspeedreadingtestbasedonashortcomprehensionpassagesisaninadequatemeasureofreadingability（andthushasalowconstructvalidity）unlessitisbelievedthatthespeedreadingofshortpassagesrelatestotheabilitytoreadabookquicklyandefficientlyandisaprovenfactorinreadingability.

C:

Empiricalvalidity

Thebestwaytocheckontheactualeffectivenessofatestistodeterminehowtestscoresarerelatedtosomeindependent,outsidecriterionsuchasmarksgivenattheendofacourseorassessors’ratings.Iftheevidenceshowsthatthereisahighco-relationbetweentestscoresandatrustworthyexternalcriterion,thenthetestcanbeclaimedtohaveempiricalvalidity.Empiricalvalidityisoftwogeneralkinds,predicativeandconcurrentvalidity,dependingonthewhethertestscoresareco-relatedwithsubsequentorconcurrentcriterionmeasures.Empiricalvalidityreliesinlargepartonthereliabilityofbothtestandcriterionmeasure.

D:

Facevalidity

Facevalidityheresimplyreferstothewaythetestlooks—totheexaminers,testadministers,educatorsandthelike.Obviously,thisisnotvalidityinthetechnicalsense,yetitsimportanceshouldnotbeunderestimated,forifthecontentofatestappearsirrelevant,sillyorinappropriate,testadministerswillhesitatetoadoptitandexamineeswilllackthepropermotivation.Forexample,ifatestofreadingcomprehensioncontainsmanydialectwordswhichmightbeunknowntothestudents,thetestmaybesaidtolackfacevalidity.

GCommunicativeapproach

2Reliability

Reliabilityisameasureofthedegreetowhichatestgivesconsistentresults.Atestissaidtobereliableifitgivesthesameresultswhenitisgivenondifferentoccasionsorwhenitisusedbydifferentpeople.Atestcan’tmeasureanythingwellunlessitmeasuresconsistently.Twodifferenttypesofconsistencyorreliabilityareinvolved:

reliabilityofthetestitselfandreliabilityofthescoringofthetest.Testreliabilityisaffectedbyanumberoffactors,chieflyamongthembeingtheadequacyofthe