Approaches to Lang Testing by Dr Ram

Embed Size (px)

Citation preview

  • 8/6/2019 Approaches to Lang Testing by Dr Ram

    1/15

    +PPROACHES O LANGUAGETESTING- RamuhishGiri'Abstract.

    lhe fundamental question in any language testing theory is 'What does it mean to know alanguage?' In attempting to answer the question. numerous language testing theorists havedeveloped theoretical models of language, and have described tinglage abitty, competence,communicafive competence,and performance. They have adhered that understanding of thesewitteyble. a test developer to des9n, develop and use iarguage tests hat match their descriptions, andtherefore, meet the criteria of validity, authenticit.v and. ipticability of target language use.Out ofthe numerousmodels emergeapproaches.to lan*yqse tginng whi'ch-p*ia" an"unierstanding ifhow to view a languageand languageabilities aia no, best o assesshem.Thepresent arficle discassessome of the most prevalent approaches in thef etd of languagetesting, specially the second or foreign language tesnng, and provides a bnef ciitique for eich-ofthem. Filtally, it presents a case as to what may be,-in o iituonon where there are too manvapproaches to choosefrom, appropriate for testing EFL in counfties like Nepal,

    l. IntroductionThe design for developing language tests is based on and influenced by how a testdeveloper iews anguage. hat s to say,a view of what languages hasdirlct bearingonwhat s testedandhow it is tested.Language ompetencbi tanguage bility s the ,r-ujo,concernn all languageeachingor testingapproaihes ecausehit is, obviously,whai isbeing ested. someapproachesroposeo measuret in isolationor in testconditions;other'sry to see t in the contextof all other actors hat impact t. A structuralist iew oflanguage'or example,s.likely o emphasiseccuracy, ndtherefore xertanapproacholanguageestingwhich divides anguage ystemsnio segmentsn order for developingseparateests or them-A single rait view,on theotherhand, reats anguage sa unitaryabilityand recommendsn integratiVeest structure.A multi-methoa,ulil-trait view oflanguage, owever,.-.views.anguageas a combinationof knowledgeor competenciescoupledwith the ability o implementhat knowtedgen language ppropnateo anygivencontex!,andprovides broadbasisor both hedev-lopmeniunJ r. brunguage ests.' The aim'of this article is to discusssomeof the languageestingapproaches

    particularlysecondor foreign anguageestingapproacheshat hive beendominantlynpracticen the language-testingield in the tastndy years.Thereut. ri* r."iions of thisartiile, eachone devotedo a separateetofupptou.h", representing separatechoolofthought'Eachsectiondiscussesow the concernedchoolortrtinun! "i;, hnguageanclanguagebilitiesconstruct)andwhatapproachesheysuggesto measureuchabilities.

    Mr' Giri is Lecturs at theDepartrnent f EngiishLanguageEducation,Tribhuvanuniversity, Kirtipur. presently,he sa research cholarat the Schoolof Education,Victoria u*versiq,. Melbourne.Australia.Journal ofNELTA L b L 7 , N o . 1 & 2 December,2002

  • 8/6/2019 Approaches to Lang Testing by Dr Ram

    2/15

    Ashish GiriThe approachesiscussedn this articleemerge, s t was ndicatedearlier,out of theand often diverse anguageestingmodels.Discussions f such modelshavebeen excludedas it is beyond he purport of the presentarticle.For af languageesting nodels, eeCanale nd Swain 1980)andMcNamara1996).

    The Approaches1The Discrete-point pproachespproacheso testing anguage ttempt o testknowledgeof a languagenegments.uchapproachesrebased nthe theory hat anguage onsistsdifferentparts(grammar, ocabulary, ounds,'etc.)and differentskills, for example,speaking,eadingandwriting, andtheseare madeup of elementshat can beseparately. hey assumehat a languageest can compriseof a large numberofest items, which, as a group,cover all aspectsof languageability. Languagehus, s testing he ability of handling f differentelements f the language ystemsdevelopment f individualdkills separatelynd ndependentlyn the assumptionhat

    tem givesprecisenformationabout'acandidate'smasteryof that elementor skillhat the otal scoredescribesisor her anguage bilityasawhole(Banerjee 000).language(asystemofsystems)

    Elements& skills (testedseparately /objectively) Aggregated core Languageability interpreted n terms ofthe score)

    Score reflectsone'sAbility in language)Language*rFigure 1:The discrete-point pproacho languageesting(ba$edon the ideas rom Lado 1961)

    heoryof teaching nd estinganguageestson thepremisehat languagea setof habits nvolvesmattersof form, meaning nd distributionof severalevelsofthe testingof language hould eprecise, bjective,elevantand scientific, ndtestingknowledgeof the elements f language t discrete evels is equivalent oknowledge f language. typicaldiscrete-pointest consists f (a) test of a specificf language uch as a syntactic oint or phonological.item r a vocabulary(b) a reading est in which comprehensionf a text is tested hrough esting hef specific tems containedn the text, and (c) a writing test in which as requiredo write onor abouta specifictemor information lready rovided.

    Obviously, he focus is on the achievernentf accuracy hrough reproductionoflanguageelementsndependent f their context rather than on fluency andf meaning.NELTA V o L 7 , N o . 1 & 2 Decanba,2002

  • 8/6/2019 Approaches to Lang Testing by Dr Ram

    3/15

    Approaches to Longuage Testingil3The discretepoint tests, through they are highly reliable,are not valid tests becausethey are irrelevant, does not meet all the requiiementsof testing, are too artificial to beauthentic Weir 1990) and therefore,do not give informationabout how the candidateuseslanguage n real communisative situation (Davies 1990). Knowledge of languageelementshas no practical import4nce without the knowledge of languageuse. So atomistic typesoftesting which test l4nguageelements n isolation has no practical value because t givesonly limited information about the test's languageabiiity (Spolsky 1975), and no attentionto the way languagbelements nteract in b larger context of communication.The discretepoint test is, therefore, ineffectivebecausecrucial propertiesof languageare lost when itselementsare separated Oller 1979).

    2.2 The Integrative ApproachesIntegrative approaches,based mainly on the work of Oller (1979), emergedas anopposition o the discretepoint approaches, nd promoted he notion that language bilityis a unitary factor rather han a divis ibleconstruct.According o the Unitary CompetencetrIypothesis UCH), underlying anguagebehaviours anbe specifiedon the basisof thecandidate'sPragmatic Expectancy Grammar which is a psychological epresentation fthe languageuser'sability to expect utterancesn a given context (Shohamy1996), andwhich is the chief mechanismunderlying the skills of thinking, understanding,speaking,reading and writing. Integrative tests, such as cloze and dictation, are said tooperationalise he candidate's languageability, which includes linguistic as well as extralinguistic knowledge to reconstitute the meaning of a text. That is to say, such testsintegratea numberof languagesub-skillssuch.asgrammatical nowledge, he ability toextract nformation, lexical knowledgeand the ability to construct a discourse.

    A cloze test, for example, tests grammatical knowledge and ability to extract andpredict meaning rom a text.^A dictation tests a candidate's istening ability, ability todecode n writing, and therefore, their ability of writing skills and at the same imq theirability n receptivephonology.A writing test cantest their knowledgeof grammar,abilityto constructa discourseand ability to manage heir language esources Banerjee2000).Similarly,a reading est can test a candidate's ornprehension,bility to extract,construct,establishandguessmeaning,knowledgeof grammar,and exicalknowledge.The integrativeapproacheso languageestingarebasedon the principleof componentfactor. analysis in that scores from tests or sectors such as voca6ulary, grammar,phonology, reading, dictation cloze and compositionare analysed and conelated todetermine he candidate's anguageproficiency. ntegrative ests are commonlyused inprbficiencyestsbecausehey rEflect eal anguage semoreclosely Banerjee 00b)

    .t lCommunicative ticuspreludeSub-test:Writing

    Figure2: The Designof the ntegrative est adaptedromCarroll andHall 198S: )Journal ofNELTA V o L 7 , N o . 1 & 2 December.2002

  • 8/6/2019 Approaches to Lang Testing by Dr Ram

    4/15

    ests,which emphasisehe mastery f language se n the total socialcontext,measure he total communicative spectof language. t recogniseshatperformance involves the individual'scapability of mobilising his or hercompetencendperfonnancebilitiesn an ntegrated ay(Howard1980:275).Oller'sUCH doesnot explicitly mpiy what is traditionallyermedas competenceandHowever,an analysis f the ftatures of UCH revealshat theyshare omef 'performance'Ollerwrites hat theobjectof interest s language s t is used orurposes for getting and giving informationabout facts or statesofand or expressingttitudes owardshose actsor states f affairs. Oller 1979:16).rocesseslearlyeflectHyme's 1972)concept f ability or usedefined s ulesofcNamara 1996) writes that there are two aspects f Oiler'smodel hato performance.i) Naturalness nd real time processing f language lementsnwith the discourse, nd (ii) relating he sequencing f language lementsolinguistic ontext ia pragmaticmapping.Oller'sPragmatic ExpectancyGrammar, which is knowledggof the languages constructed nd modified n the courseof language cquisition. Languageconsistsof suchexpectancy enerating ystem. Language esting, hen, isa task of assessinghe efficiencyof the Pragmatic ExpectancyGramrnar ins in the process f constructing.A languageest n order o be valid nof the theoretical onstructwill have o involveandchallengehe efficiencyof theevelopingrammarOller1979).Corder 1973:351) alls hisphenomenonthecompetence" constantlyevolving,dynamicability in the target language.of the est, her4s to showhow ar the student'sanguageasmovedowards-an

    the nativespeakers'ystern.Oiler'sUCH hasbeenquestionedor the fact that the data or his researchame romacademicestswhich ocuson conscious nd explicit eatures f languageather han heand sociolinguisticeatures, nd that the data s not conclusiveo support heexistence f unitary competenceBachman nd Paimer1996).Oller's approach mplieson the unitarycompetenoe, languageearnershouldbe able o use he targetlanguageequallywell in ali its manifestations. owever, it is not usually the case.practitionerseverywhereexperiencehat non-Englishspeakingbackground(NESB) earners regoodat one skill or aspect ut havedifficulties n someother aspects(Bachman1989).Also, Oller's echnique f componentactor analysiso explore hestructureof his data is inappropriate Chalhoub-Deville 997).One otherproblemof integrativeests s thatthe results rereportedasa single corehough he asksin the tests combinedifferentcomponents f language bility. It is thereforedifficult tointerpret hescoreas o what t actuallymeans.

    NELTA t b l . 7 , N o .& 2 Decenber.2002

  • 8/6/2019 Approaches to Lang Testing by Dr Ram

    5/15

    Approaches to Language Testing/l52.3TheProficiency pproachesThe erm second r for_eignanguage roficiency's used n two differentways.Firstly, tis. elated irectlyo'ability'and s defined s he degrbe f competence"orapabilityn agiven anguage emonstratedy an individual n a liven point tf time independent f aspecificcourseor textbooksor teachingmethods.S""ondly, t relates o th; extentandadequacy f an individual's ontrolor masteryof target anguagen all kindsof socialorinteractive ituationsncludingwork settings sdemoistratedn tests.The ormermeaningof proficiency enotesompetenceherealthe ateronespecifies erformpnceWhat makesa languageest communicativer non-communicatives the problems,tasksand he manipulationsf languagekills Davies1990).To Davies,he purpose ndcontextof testingare hemostreleyanteatures f communicativeanguageesting.He. ina similar ashiono Hymes's ichotomy f knowledgeandability for use,distinguishestwo aspects f communicativeompetence 'knowl elge that' whioh efers o t

  • 8/6/2019 Approaches to Lang Testing by Dr Ram

    6/15

    Ashieh Glrioller (lg7g), the pragmatic approach suggestedby Daviqs (19'68),.the theoreticaladvocated by-cumminj 1t99+;, rating-scale'approachetu:h. as ACTEFLand ASLp& and assessmentrameworks developedby Hilgfotis, Bailey and(19g1, quoted in Chalhoub-Deville1997))are approacheso proficiencyassessment.differen..i in theseapproachesie in the ways they view languageand the focus thejwhile developing-anguage tests. A brief comparison and discussionof thes given n the final sectionof the articie'

    The PerformanceAPProachesperformance-based pproacheso language esting developedout of the uncertaintiesexisted n the tanguageesting field in 1980smainlybecauseof the non-existenceof atheorelicJ model, in spite of the fact that they have a long history and ausage n other fields. So, pragmaticconsiderations nd operationaldefinitionsthe devJlopment nd use of language ests Shohamy1996).As a result, anguage

    were 'theory-free', performance-baied, ask-driven, and were considered o beunctional,direct andauthentic Shohamy1996.145).The performance-based pproacheso language esting,thus, developed n response oneeds in specific-pu.pot" contexts, diawing on the theory and practice ofurr.ru..nt in vocational fields (McNamara 1996)' However, the approacha particular interest in communicative language testing mainly after theadvancemade n the fieldby Hymes 1972)Language estson this approachwere developed ndusedon thebasisof (a) purposeof(b) testin! context,(c) samplingof tesi task, (d) simulationof situation,and (e)of languige contentand of languageabilitiesand skills.

    RaterScale YPerformanc.l*rtn, (score)-)Instrument A

    . CandidateFigure4: Theperformancepproacho languageesting adaptedromMcNamata1996)

    testperformance,s he figure4 aboveshoys, s analysedy means f a ratingscale,provides score nda description.Thereasons hy performanceased pproachesecameerypopula\-?fehat(a) theyefree from compliiations f theoreticai-complicatederms, b) describinganguagenerms was easyto iomprehend, c) rating scalesdescribedn communicable ay, (d) rating scaleswere generic,descriptiveand. heriteria in them and were taken from everyday ife situations,apd.(e)based tests have high face validity high wash back and high ciient

    ofNELTA Vol. 7, No. I & 2 Decenber,2002

  • 8/6/2019 Approaches to Lang Testing by Dr Ram

    7/15

    Approaches to Language Testing/l?The very simplistic approach o tests becomesa weaknessof the perfbrmance-basedapproaches o language esting. As it was said earlier, performancebased ests focus ontasks rather than constructs. Constructswhich refer to theoriesof competence, nowledge,communication, and skills are overlooked causing, herefore, a lack of construct validitybecause t is the constructs hat should drive the development,scoring, and interpretationof performanceassessment. he performanceapproacheso languageassessment,hus, are

    simplistic, narrow and are not based on theory of construct, but on a niurow view ofcommunication Shahomy 996).A theoretical model of communicativecompetencebased on a more comprehensiveunderstandingof the.factors underlying ability for use is neededwhich should guide theselectionand constructionof.relwant test tasksas well as the rationale or scoring criteriaand rubrics.

    2.5 TheCommunicative pproachesThecommunicativepproacho languageesting s basedon the assumptionhat languageis learned o expressdifferentkinds of functions, and emphasiseshe needof testinglanguage y usingprocessesf communicationuchas using anguage ppropriatelyndifferent ypesof situationso interactwith peopleandperformon a widerangeof tasks,Such estsemphasiseuthenticityof test methods,authenticityof test materialsandauthenticity of responses o that the test outcomesgive a real picture of a test'scommunicative ompetence.Thus, the focus is on qualitative evaluation of thecandidate'sanguage nowledgeather hanrthe. uantftativeassessmentf linguisticfeaturesSavignon 000) ,

    Communicativeanguageestingapproachesecamb commonpracticen late 1980'sandearly 1990's, articularly fter hemuch alkedabbutworksof Canale ndSwain1980.andBachman1990)who respondedo the concern f 1970shat,languageroduced nnon-communicativeests was artificial and non-representativeof the test's actualcommunicativeompetenceShahomy 996).So, herewasa call for direct ests n whichboth he est ormatandprocedure uplicatedscloselyaspossiblehe real ife situations.Bachmanand Palmer's 1996) multi-method rnulti-trait approach has beenparticularly. opulbramongESLIEFL practitionersn the US 'andAuitralia becausehisapproachrovides n elaboratendcomplehensiveepresgntationf languageroficiency.Bachman nd Palmerusethe term "framework" o referto what otherscall 'model'(McNamara1996).Bachman'sheoreticalrameworkof communicativeanguageability(CLA) consistsboth knowledge competence) nd the capaclty or implementingoiexecutinghat competencen appropriate,ontexrualisedommunicativeanguage se. tshouldbe noted hat thatthere s anguage bility hat s non-communicative.he anguageability hat is non-communicatives not of muchuseexceptperhapsor research urposes.Every anguages learnedor cosmunication. herefore, ommunicativebility s what is

    Jownalof NELTA VoL7, No. l&,2 Decenbq,2002

  • 8/6/2019 Approaches to Lang Testing by Dr Ram

    8/15

    tuhish Giridistinctions between inguistic, commuiicative and pragmatic

    The CLA framework includes hree components:Language competencecomprisesessentiallyof a set of specificknowledgecomponents hat are utilised in communicationvia language-

    Strategic competence: Strategic competence characterises he mental capacity forimplementing the components of language competence in contextualisedcommunicative language use. Strategic competence thus provides the means forrelating language cornpetenciesto features of the context of situation in whichlanguage use takes place and to the language user's knowledge structures (socio-culturalknowledge,'realworld' knowledge),andniechanisms:Psycho-physiologicalmechanismsrefer to theneurological and psychological processesnvolved in the actual executionof languageas a physicalphenomenonsound ight).

    Svnt. Phor/Graph Cohesion. Rhet. Org I deat tr{arip Heu Imag Semit Smit Semit Cultual. Fuc. Fuc. Func. Func. todial toReg. toNsr Rdtefig.or wiety of Socech

    Figure 5: Frameworkof CommunicativeamguageAbilify (BachmanandPalmer1996)of languageand the knowledge.of the world contribute o the

    strategic bilities eadingo psychologicalndphysiological rocessingnmake utterancesappropriate or a give context of situation.The languagds the figureaboveshows,canbe brokendownto other sub-competencies.covers hreebasicareas, nowledge f language r languageompetence,aspectof ability or useof strategic ompetence,ndmodalities f performancepsycho hysiologicalmechanism.Language ompetenceonsists f two mainaspectsa) organisational ompetenceknowledge of creating or recognisinggrammaticallycorrect utterancesand

    PRAGMATICCOMPE

    V o L 7 , N a l & 2 December.2002

  • 8/6/2019 Approaches to Lang Testing by Dr Ram

    9/15

    Approeches to Language Testing/l9comprehending heir prepositional content (grammatical competence) or of organisingthem to form text (textual competence)and (b) pragmatic competence-knowledgi of thipragmatic conventions for performing felicitous language functions (illocutionarycompetence)and knowledge of sociolinguistic rules of appropriateness sociolinguisti-ccompetence). Strategic competence is a general ability of assessingand planning todetermine and execute means of achieving communicative goals. ttris is ability, notknowledge.Pragmaticcompetencenvolves an understandingof the social rules governingthe use of language.The speaker/writermust know what is socio-linguisticallyacceptableor appropriate n a given context, and be sensitive o varieties of language, o registers,naturalnessand to cultural references. t is not easyto assesspragmaiic -ompetencenotonly because t is difficult to match utteranceswith contexts,but ilso because t is moredifficult to elicit than to measure t. Organisationalcompetence, hus, has tended o be thelocus most anguageestsbecauset is quantifiable.

    The framework seems'rich enough to conceptualisehe issues nvolved in secondlanguageperfbrmancesituation and comprehensive o incorporate every major aspect ofthe languageability; and it appears o be a still-evolving frimework oi second anguageability which provides a map rather than a prescription of language ability, setting outthoseareasof ability that need o be taken into account n second anguageurr.rrrn.nt.There are problems with this approach. Firstly it is not easy to identify testperformancewith trait or ability becausea performancecan be associatedwith ability if thetest task is 'direct'. However, given the fact that language ests, like all mental measures,are indirect, he task of identifyingperformancewith trait is problematic Davies 1995).Secondlyhe approachs primarilyconcernedwith construct nanAity and disregardsace,

    context, and predictive validity The approachsuggests overage n tenhs o? languagecomponents, killsand abilities,and recommends ontentof communicativeanguageeststo be motivating, substantive, integrated and interactive; and topics, opinions,- deas,readingpassagesexts, dictation and composition o be appropriateand 6e basedon acommon heme Bachman 990:320)The last recommendationgiven above may ensurecoverageand authenticitybut it stilldoesnot ensure hat the test will have content validity, whiih is a necessary uality of atest' In order for sucha test to havecontentvalidity the test constnrctionshouldbe basedon a thorough analysisof the target domain upon which the test content s to be based(Davies1995)

    2.6 The Rating-scalesApproaches- ACTFL and ASLpRRating scalesas a basis or assessinganguageproficiencyare becomingwidespread n therecent imesbecause f people's ncreasingnterest n transparencyn educational ystems.Rating scalesare useful because hey describe what the attainment of a given level oflanguage proficiency means in practice CNorth 2000). Rating scale alpp.oaches oproficiencyassessmentre becomingmore and more popularalsobecaus. uiing scales a)Journal ofNELTA L b L 7 . N o . I & 2 December, 2002

  • 8/6/2019 Approaches to Lang Testing by Dr Ram

    10/15

    Giria basis or the learner o comparehim/herse[ 0) establisha frameworkofo describe chievementn termswhich aremeaningful nd user-friendlyor allnd (c) increase he reliabilityof subjectivelyudgedratings,especially fanguageskills, and (d) provide a common bands and meaning or sughRating scalesalso set learning'goalsand descriptionof proficiency ornd allow the results o bemeasuredgainst he targets.Theyalsooffer a profileability in the target language,which becomea basis for selection"or a remedialprogrammeor the perspective mployment rgani'sationsror educationnstitutes

    Today, he mostpopular atingscales re he AustralianSecondanguage roficiencyASLPR), he Interagency anguageRoundtableILR) scales, ndthe AmericanTeaching f ForeignLanguage roficiencyACTFL) Guidelines, hichare allrom the sarnesource, he rating scale developed y the US ForeignService(FSI).The ILR scales re for the US government mployees nly andnot for af English.So, t is excludedromthe discussionere.The AmericanCouncil of Teaching f ForeignLanguagesACTFL f9S2)

    ACTFL Guidelines1982)describe test's languagg roficiencyon a scale angingo superior.The descriptorsepresent egrees f proficiencyn real worlddescribe ow languageearnersypically unctionalong he rangeofhe trisenctional rameworkof the scale see igure6 below) consists f (a)which tells what a test can do ,at what level, (b) content,which sayswhator topics can serve as a basis or assessmentt the level in question,and (c)which describeswhat qualityor feature he candidate'sanguagewill containatwhat evel.LanguageFunctionS Accuracy/'DescriptionofLearnerLaneuaee TexVContent0 - 0 +LowNovicel - l +Intermediate2 - 2 +AdvancedJ - J TSuoerior

    Figure6 Summary ersionof ACTFL Scale adaptedrom ACTFL Guidelines 982))problemswith ACTFL Guidelines re hat they (a) take a generic,educated ativeas a criterion againstwhich to measurehe proficiency, f NESB learners,b)artificial context where the interactional skills and the set of roles that aan play are limited, and (c) do not correspondo findings n L2 acquisition1995).Furthermore,he datacollectionand analysis rocedures ndhow they

    V o L 7 , N o . 1 & 2 Decentbs,2002

  • 8/6/2019 Approaches to Lang Testing by Dr Ram

    11/15

    . Approaqhes to L,anguage Testing/2l

    were initially turned into performancedescriptorsare questionable,and so is the validity ofthe guidelines bgcause liere is no empirical evidencebn how the scale descriptorshavebeen calibrated. Also, the guidelinesallocate 'key-tasks' to levels without a principledbasis North 2000):2.6.2 Australian SecondLanguageProficiencyRatings (ASLPR' 1984)ASLPR is an elaborateprocedure o measuregeneralproficiency in a language earnedas asecond or. foreign language. It is descriptive scale in the sense that it consists ofdescriptions of languagebehaviour at nine levels, from zero to nine. In other words,ASLP& as it is shown in figure 7 below, consistsof general descriptionsof observablebehaviourwith 'examplesof-specifictasks' that exemplifu he sorts of behaviour that areobservable Ingram iSSSl. ASLPR has been developed in English, French, Italian,Japanese nd Spanish,and ts usehasexpanded apidly n Australia.

    Proficiencyevels Skill-wiseGeneialDescription Example ofSoecificTasks Comments

    0Zero or Initial proficiencvl - l +ElementaryorMinimumSurvivd proficiencYJ I

    Minimum SocialProficiency3 Minimum VocationalProficiencv4Vocational Proficiency5Native-likeProfi iencv

    Figure7:A Summary ersionof ASLPR adaptedrom ngram1984)

    ASLPR describes anguagebehaviours n general proficiency levels from zero to native -like. Each macro-skill s described eparately nda test's proficiency s stated n a profile..Each profile has a numerical value which are accompaniedby descviptions. Thedescriptions re statements f what the candidate's ehaviour s in terms of the taskshecan carry out andhow he can carry them out. The numerical scoresare interpreteddirectlyfrom the descriptionsso that the readerknows what sortsof things the candidatecan do inthe target anguage.

    However, ASLPR is not a test instrumentbut a rating scalewhich is basedon, as t issaidearlieron IRL oral interview.The problemsare he scalesuggestshat the assessorandevelophis or her own test asksbut the scaleprovidesno guidance or that,nor does t sayhow the new test askscanbe validated.

    Journal ofNELTA V o L 7 , N o . 1 & 2 December,2002

  • 8/6/2019 Approaches to Lang Testing by Dr Ram

    12/15

    22lRarn Ashish GiriSumApproaches Viewson LanguageProficiencv Considerationsor TestDeveloDment CommentsDiscrete ointApproachesLado1961

    FormalknowledgegrammaticalcompetenceLanguagea s_vstemfhabitsofcommunicative

    Indirect eferenceoperformanceOverallcontrolCross ulturalunderstandins

    No testof interactiveabilityNo considerationfauthenticityof materialslmaterialsesponsePragmaticApproachesDavies 1998)

    ComponentialanalysisOverallcontrolLearningapproachlinguistic)Performance pproach(language t work)A combination f skillsandwork sample ornains

    ComponentialnalysisReliable, alidComponentialnalysis ftestperformanceinteractiveApproachesa)Oller 1979)

    Unitary biliwAbstract.eneral Focus n academiconsciousandexplicitfeatures flanguageLacks unctional ocialaspectComponentialnalysis ftestperformanceInteractive

    Approacliesb)Carroll 1961,72\. 1985

    ComponentialAbilities Domains. killsIntegrationhrough estspecification nd estmodesComponentialnalysis ftestperformance

    CommunicativeApproachesBachman 990CornponentialAbilities Real ife approachIntegration pproachContexualisationf languageTestmethods houldmirroraspectsfperformanceunder estDomains ndskills

    Tooelaborate/comprehensiveresultingn large ests

    Performance-basedApproachesMcNamara 996

    Capabilitiesnvolvedin language seLanguagenowledgeunderlyingpedormance

    Specific urposeestingDirect estsAnalysisof skillsandcomponentsfperformanceWorksampieapproachQualityof executive f

    Focus n task ulfilmentandqualitl'of taskexecution

    UALP/tsICSCummins 984 Cognitiveandacademicailyaspects flanguageproficienry,CommunicativefluenryAbstract, general

    SurfaceluenryAcademicandcommunicativeperfonnanceAcliveandcognitiveparticipationn a taskcontexlembedded ctivitv

    Theextent o whichmeasure f languageproficienryshouldrelateacademicachievemenl ncilanguage roficiency

    AssessmentFrameworkHinofotis,BaileyandStem1981;ContefiSoecific

    GeneralanguageproficienryComponentialbilityFocus n consciousndexplicitGatures flangr:ageLanguage roficiencyDeliverv

    Context pecific

    oINELTA lbl. 7, No . I & 2 December,2002

  • 8/6/2019 Approaches to Lang Testing by Dr Ram

    13/15

  • 8/6/2019 Approaches to Lang Testing by Dr Ram

    14/15

    AshishGtriA rating scalewas providedto the ratefs during the secondround of viewing the tapes.Jf t.gt.rrion analysis revealed that all main categories and sub-catdgoriesto the evaluation.The FTA's categoriescanlbe related to the correspondingof Bachman's 1990) CLA frampwor\. For example, he grammarcomponentne communicationcategoriescan relate to grammatical competenceand illocutionaryof the CLA frameworkrespectively, hough in much curtailedforms.Similarly,Chaulhoub-DevilleIgg7) adapted he ProfidiencyAssessment rameworkstudents' oral proficiency for everyday communication. He too

    he conclusion f Hinbfotis et al's (1981) study hat there s a colrespondencefcategories with the components of CLA framework. Thus, the assessment(Hinofotis el at, and Chaulhoub-Deville) provide empirical justification forspecific aspects of the more general theoretical models to reflect thef specifi contexts.

    To sumup, basedon the discussions bove, t canbe concluded hat (a) most anguageapproachesfavour componential approach to language proficiency though theegarding the nature of the components emains debatable o this date, (b) manyhave strong theoretical foundation but scanty empirical support, (c) sameare based on empirical data but suffers from poor statistical support, (d) some'are to pbstract and generic or too elaborate and comprehensive o be ofpractical value, and finally (e) adaptationof an approachor approacheso suitthe specific characteristicsand purpose without altering:the theoretical foundation isbestoption or a practicalpurpose'

    References:..C. Clapham.D. Wall .199'5. anguageTestConstuctionandEvaluation.Cambridge:CU P.J, C., K.j. Krahnke and C. Stansfield 1987. Reviews f English languageProficie'ncvTests.WashingtonDC. TESOL.L. 1989.An Examinationof SomeLanguageProficienc.v ests rom a Communicative iewPoint.ERIC.2001.L. and. A. Palmer 1996.LanguageTesting n Practice.Designingand DevelopingUsefulLanguageZesls.Oford: OUP.L. f. tSge. The Test of English as a ForeignLanguageas a Measureof CommunicativeCompetence.Technolog,, nd Language Testing:A Coltection of Papers rom the Annual, Colloquium n Lartgua[eTestingResearch. . Stansfield. rinceton.New Jersey, LTS.7: 69-88.. F. i990. Fundamental onsidqrationsn LanguageTesting.Oxford:OI-,,?.. F. 1999.Designing ndDevelopingUsefirllanguageTests.Studiesn Language esting.A- A.B. C.Davies. dler,N. Iwashita. . McNamara. ambridge: UP:109-118'J. 2000. IntegratedTests. n Byram, M. (ed) .2000.RoutledgeEncyclopaediaof LanguageTeaching ndLearnirg.London:RoutledgeG. 1995. anguage lsessmentn ,4ction.Sydney:NationalCentre or EnglishLangrrageeachingandResearch. acquarieUniversity.M. 1994.On SomeTheoreticalFramervorkor LanguageProficienry.Language roficiency andAcademicAchievement. .Rivera:Clevedon, von.Multilingual Matters.M. andM. Swain.1980. heoretical asesor Communicativepproacheso Second anguageTeaching ndTesting. ppliedLingttistics/l : l'11.B. andP.Hall .L985.tlake YourOwn anguageesrs. Oxford:Pergamon ress.

    ofNELTA VoL 7. No. I & 2 December,2002

  • 8/6/2019 Approaches to Lang Testing by Dr Ram

    15/15

    Approaches to l-rnguage Testing/25Chalhoub-Deville,M. 1997.TheoreticalModels,Assess*"nt frameworksand TestConstructionLanguageTesting 4 l 3-22.Coombe,C. 1998.CurrentTrends n EnglishLanguage esting.ERIC.Corder,S.P.1973. ntroduci g,Appl ed Linguistics. ondon:Penguin.Crynmtns,J. 1994. Wanted: A Theoretical Frameworkfor Relating Proficiency o AchievementamongBilingual Students. anguageProficiencyandAcademicAchievemenl. .Rivera.Clwedon.AvonMultilingualMatters:2-19.Davies,A. (ed.)1968. anguageTestingSymposium. Psycholinguisticpproac&.Oxford:OUP.Davies,A. 1990.Principlesof LanguageTesting.Oford: BlackBasil.Davies,A. 1991. Issues n LanguageTesting.ESLDILLS Project Seminar,Melbourne,UniversityofMelbourne.Davies,A. 1994. TestingCommunicativeLanguageor TestingLanguageCommunicatively;What?How?MelbournePapersn LanguageTesting /L: l-20.Davies,A. 1995.TestingCommunicativeLanlrage or Testing l-angnge Communicatively:What?How?MelbournePapersn LanguageTesting.4l: l-20.Davies,A. andA. Brown 1990.DesigningnstrumentsoMeasureLanguage roficiency.IWG/Assessmentof ProficiencyLevels n JFL, Canberra.Davies,A., et al. 1984.Surveyof English LonguageTeachingn Nepal.Kathmandu:British CounciUODA

    and HMGAI, Ministry of EducationandCulture, HMG[{.Departmentof Immigrationand Ethnic Affairs .1984.Reporton the ormal trialting on theASLpR.Canbera:AustralianGovernmentPublishingService.Dhakal, T. P. 2000. Effectivenessof DiscretePoint Testsand IntegrativeTestsas Measuresof EnglishLanguageProficiency:A ComparativeSudy. Departmentof EngtishLanguageEducation,Faiuttyof Education Kathinandu:TribhuvanUniversity.Giri, R. A. 1996.ClozeTestsn the ForeignLanguageClassroonn.ournal of NELTA /l:26-31.Howard, F. '1980. Testing CommunicativeProficiency in French as a Slcond Language:A Search orProcedures.CanadianModernLanguagbRevicw36/2:272-2g0.Hymeg,D. Y?, On CommunicatireCompetence.n Prideand Holmes ds.), Sociotinguistics:269-293.Ingram,P, Y_81Using Proficienly Rating Scaleswith High School oreignlanguagJlearners. ERIC ED24976',7Ingranr,D. E. 1985.AssessingProficiency:An Overviewof Some Aspectsof Testing.Modelting andAssessing ebond_Languagecquisition.K. a. P.Hyltenstar4M. Clwedon,Engl-and, ultifnggalMatters. 8z 215-276.Lado,R. 1961. anguageTesting. ondon: Longman.McNamara, .F. 1996.'Me-asuringecondLanguage'performance.ondon:-LongmanNorth, 8'2000. TheDevelopmentof a CommonFrameworkScaleof Language-Profciency.Oxford: peierLang.oller. J. 1979 LangyageTestsat schools:A pragmaticApproach.London:Longman.Savignon, . 2000.Communicative-Languagedching. InByram,M. (ed.) 2006.RouttedgencyclopaediaShohamy;E. 1994.The Rgle o-fLanguageTests n The ConstructionandValidationof SecondLanguageAcquisitionThgrrieg,Methodolog,,n Second anguage cquisition.E. S. G. a. A. C. C. Tion-e.Cambridge: UP: 133-141.Shohamy,E. 1996.Competencend Performancen.LanguageTesting. n G. Brown,K. MalmkjaerandJ.Williams (ds.) Performanceand Cohjcetencei Seionaunluage Acquisition.Cambriige: ClIp:I 38-15.Singh,G. B. 1996.SLCbamination inNepa,l.Kathrnandu:ResearchCentre or EducationalnnovationandDevelopmentSpolslcy'_B'1975.LanguageTesting: Art or Science?Key Note Addressat Association nternatronalede_ LinguistiqueAppliquec'eWorld Congress, rance,lpolrt y, B. 1993.Testing he Englishof Foreignstudentsq 1930,ERIC.2002.!-rylttry, B. 1995'Measured |lordg: TlqeDevelopment f ObjectiveLanguageTesting.Oxford:OUp.weir, c. J. 1990.communicativeLanguageTesting.Newybrk: prenticeHall.JournalofNEL[A