This document discusses various approaches to language testing, including discrete-point testing, integrative testing, communicative testing, and performance-based testing. Each approach has strengths and weaknesses. Item analysis is also discussed as a method to evaluate test items and identify issues. Overall, the document examines different theories and approaches to language testing throughout history to better understand how to develop and implement effective language assessments.
2. Introduction
Testingiscloselyalliedtoteachinginseveral ways.Firstof all,manyclassroomactivities
commonlylabeled'teaching'are actually''testing''whenateachercallsonstudentsto
demonstrate bytheirperformance inthe classroomthattheyhave masteredcertain
material orthat theyhave internalizedacertainrule,the teacher, inasense, teststhem
Paulston(insectionfour) mentionstwokindsof language drills –forteachingandtesting
.she saysthat
What is involvedisthe differencebetween drillsthatserve primarilytohelpthe students
memorize apattern with,virtuallynopossibilityformistake anddrillswhichtestorreinforce
the learningof the pattern.
Testingisa universal feature of social life.Throughouthistory peoplehave beenputtothe
testto prove theircapabilitiesorto establishtheircredentials;thisisthe stuff of Homeric
epic,of Arthurian legend.Inmodernsocietiessuchtestshave proliferatedrapidly.Testing
for purposesof detectionortoestablishidentityhas become anacceptedpartof sport
(drugstesting),the law(DNA tests), paternitytests,liedetectiontests),medicine (blood
tests,cancerscreeningtests,hearing,andeye tests),andotherfields.
What istrue of testingingeneral istrue alsoof language testing, nota topiclikelytoquicken
the pulse or excite muchimmediate interest.If itevokesanyreaction,itwillprobablytake
the form of negative associations.Formany,language testsmayconjure up animage of an
examinationroom,atestpaperwithquestions desperate scribblingagainstthe clock,Ora
chair outside the interviewroomanda nervousvictimwaiting withrehearsed phrasestobe
calledintoaninquisitional conversationwiththe examiners. Butthere ismore tolanguage
testingthanthis
To beginwith,the verynature of testinghaschangedquite radically, overthe years to
become lesspositional,more humanisticconceivednotsomuchto catch people outon
whattheydo not knowbutas a more neutral assessmentof whattheydo.
Understandinglanguage testing
There are manyreasonsfordevelopingacritical understandingof the principlesandpractice
of language assessment.Obviously youwill needtodoso if youare actuallyresponsible for
language testdevelopmentandclaimexpertiseinthisfield.Burmanyother people working
inthe fieldof language studymore generallywill wanttobe able to participate asnecessary
inthe discourse of this field,foranumberof reasons
First,language testsplaya powerful role inmanypeople'slive actingasgatewaysat
importanttransitional momentsineducation,inemployment,andinmovingfromone
country to anotherSince language testsare devicesforthe institutionalcontrol of
individuals,itisclearlyimportantthattheyshouldbe under stood,andsubjectedtoscrutiny.
Secondly,youmaybe working withlanguage testsinyourprofessionallifeasa teacheror
3. Administrator,teachingtoa test,administeringtests,orrelying oninformationfromtests
to make decisionsonthe placementof studentsonparticularcourses.
Finally,if youare conductingresearchinlanguage studyyou mayneedtohave measuresof
the language proficiencyof your subjects.Forthisyouneedeithertochoose anappropriate
existinglanguagetestordesignyourown
Thus, an understandingof language testingisrelevantbothfor those actuallyinvolvedin
creatinglanguage tests,andalsomore generallyforthose involvedinusingtestsorthe
informationthey provide,inpractical andresearchcontexts.
Theoretical background
Historical Perspective
As testingispartof the language teaching/learningprocess,ithas alwaysreflectedideas,
proceduresandtheoriesaboutthisprocess.Thiscanbe betterexplainedprovidingan
outline of the general trendsinlanguage teaching.If we take alookat the historical
developmentof language teaching,frombefore WorldWarI to our days,itcan be seenthat
it hasalwaysbeen a matterof interdisciplinaryinterest.
Since at leastthree elementsprovidethe basisforlanguage teaching - the language,the
learner,the objectives - anumberof disciplinesmustplaydifferentroles.Linguistics
provides alinguistictheory.Psychologyisgoingtotake care of the learneras an individual,
whereas sociologystudiesthe learnerwhile ahumanbeinginsocial interaction.Pedagogy
contributes forthe processinthe classroom, throughmethodologyandtechniques.Asa
consequence of concentratingonthe same subject,these areashave resultedinnew
disciplines,suchasapplied linguistics,psycholinguisticsandsociolinguistics.
For the sake of brevityandinorderto have a didacticview of the teachingof Englishas a
secondor foreignlanguage,letusconsiderthree basicstages.Itisimportanttokeepinmind
that these stageshave notbeendevelopedinextreme andseparate ways;theirfeatures
sometimes overlap.Thesethree stageshave beennamedindifferentwaysinthe
developmentof language testinginthiscentury.Spolsky(1975) hasusedthe labels ‘pre-
scientificpsychometric-structuralistpsycholinguistic-sociolinguisticMorrow (1 979) has
characterizedthese stagesasthe ‘Gardenof Eden’,the ‘Vale of Tears’and the ‘Promised
Land’.
i) TRADITIONAL
Linguistictheory –traditionalist-
No psycho-sociological theoryinvolved-
Pedagogical implications - grammar-translationmethod-
ii) ANALYTICAL
4. Linguistictheories - structuralism/TCG-
Psychological theories –behaviorism-
Pedagogicimplications audio-lingualmethod/cognitive approach-
iii) INTEGRATIVE
Linguistictheories - semantics,pragmatics,register,speechacttheory,discourse analysis-
Psycho-sociolinguisticaspect- communication,needsanalysis-
-Pedagogical implications - notional/functional syllabuses
APPROACHES IN LANGUAGETESTING
.Discrete-pointTestingApproach
Discrete Pointtestsare constructedonthe assumptionthatlanguage canbe dividedintoits
componentsparts,andthose parts can be testedsuccessfully.The componentsare the skills
of listening,speaking,reading,writing,andvariousunitof language of phonology,
morphology,lexicon,andsyntax.Discrete pointtestsaimtoachieve ahighreliabilityfactor
by testingalarge numberof discrete items,buteachquestion testsonlyone linguisticpoint.
Integrative TestingApproach
Thisapproach involvesthe testingof language incontextandisthusconcernedprimarily
withmeaningandthe total communicative effectof discourse.Thisapproachstatedthat
communicative competenceissoglobal thatit requiresthe integrationof all linguistic
abilities.AccordingtoOller(1983),if discrete itemstake language skill apart,integrative
testsput itback together;whereasdiscreteitemsattempttotestknowledge of language a
bitat a time,integrative testsattempttoassessalearner’scapacitytouse manybitsall at
the same time
The fact that discrete pointandintegrative testingonlyprovidedameasure of the
candidate’scompetence ratherthanmeasuringthe candidate’sperformance broughtabout
the needforcommunicative languagetesting(Weir1990).By the mid-1980s,the language
testingfieldhadabandonedargumentsaboutthe unitarycompetence andhadbegunto
focuson designingcommunicative languagetesting(Brown,2004)
Communicative TestingApproach
Communicative testingapproachlaysmore emphasisonthe notionandfunction,like
agreeing,persuading,orinviting,thatlanguage meansincommunication.Communicative
language testingapproachisusedtomeasure language learners’abilitytouse the target
5. Language inauthenticsituations.The approachbeliefsthatsomeone/ astudentis
consideredsuccessful inlearningthe targetlanguage if she/he cancommunicate oruse
knowledge andskillsbywayof authenticlistening,speaking,readingandwriting.
Communicative language testshave tobe asaccurate a reflectionof thatsituationas
possible.The exampleof communicative language testisrole play.The teacherasks
studentstodo a role playsuch as pretendingthatthe studentscome tothe doctor,
pretendingthatthe studentsare inthe market
The principlesof testinginthe communicative language testingcanbe describe asthe
following(Anon,1990):
-Tasksin the testshouldresembleasfaras possible tothe onesaswouldbe foundinreal
life intermsof communicative use of language
There is a call for testitemscontextualization-
-There isa needto make testitemsthataddressa definite audienceforapurposeful
communicative intent(goal)tobe envisioned(mighthappen)
-Testinstructionsandscoringplansshouldtouchoneffective,communicationof meaning
rather thanon grammatical accuracy
Performance testingapproach
Anyassessmentcanbe consideredatype of performance whenastudentisplacedinsome
contextandaskedto showwhattheyknow or can doin that context.Performance-based
assessmentbelievesthatthe studentswilllearnbestwhentheyare givenachance to
performandshowwhat theyknowaccordingto theirownplan,collectdata,inferpattern,
draw conclusion,take astandor deliverpresentation.Accordingto Brown(2004), in
developingperformance-basedassessment,we asteachershouldconsiderthe following
principle:
State the overall goal of the performance-
Specifythe objectives(criteria) of the performance indetails-
Prepare studentsforperformance instepwiseprogressions-
Use a reliable evaluationform, checklistorratingsheet-
-Treat performancesasopportunitiesforgivingfeedbackandprovide thatfeedback
systematically
Strengths and Weaknesses of Language Testing Approach
6. .Discrete-pointTestingApproach
Strength
-The testof thisapproach can covera wide range of scope of materialstobe put inthe
tests
The testallowsquantificationonthe students’responses-
-In the termof scoring,the testisalsoreliable because of itsobjectivity;the scoringis
efficient,evenitcanbe performbymachine
Weaknesses
Constructingdiscrete pointtestitemsispotentiallyenergyandtime consuming-
The testdo notinclude social contextwhere verbal communicationnormallytake place-
-Successindoingthe testis notreadilyinferable tothe abilityof the testtakerto
communicate inreal life circumstances
Integrative TestingApproach
Strength
-The approach to meaningandthe total communicative effectof discourse willbe very
useful forpupilsintesting
This approachcan viewpupils’proficiencywithaglobal view-
-The strengthof the testsuch as dictation,writing,andcloze testisthatrelativelycheap
and easyto make
Weaknesses
-Evenif measuringintegratedskillsare betterbutsometimesteachershouldconsiderthe
importance of measuringskillsbasedonparticularneed,suchaswritingonly,speakingonly
-The scoringis not efficientandnotreliable
Communicative TestingApproach
Strength
7. -The testsare more realistictoevaluate the students’language use,asthe studentsina
role as thoughtheywere tocommunicate inthe real world/ dailylives
-It increasesstudents’motivationsincetheycansee the use of language theylearntin
classin the real world
Weaknesses
Not efficient(timeandenergyconsuming)-
-Problemof extrapolation(Weir,1990) (we cannotguarantee thatthe studentswho
successfullyaccomplishthe taskinclasswill alsobe successfulinthe communicationinreal
life)
Performance TestingApproach
Strengths
-Increasinglearningmotivation(The studentstendtobe more motivatedandinvolvedwhen
theyare allowedtoperformaccordingtotheirownplan,collectdata,inferapattern,draw
conclusions,take stand,ordeliverapresentation)
-Meaningful (itismeaningful assessmentsince we require studentstoshow whattheycan
do throughproject,performance,orobservation.Itwill givethemlearningexperience more
than justpaperand pencil test)
-Authentic(since the materialsandtopicswe use inclassis authentic,the studentscansee
the relationof whattheylearnwiththe realityintheirdailylives)
-Challengehighorderthinkingof students(Inordertoprepare forthe bestperformance,the
studentswill trytheirbest toanalyze the problemdeeperandfindmanylearningsourcesby
themselves)
Weaknesses
-Time consuming(forstudents:theyneedtoprepare the performancee.g.Download
informationforthe Internetorpreparingthe costume andpropertyforrole play,for
teacher:Teachersneedtoprovide guidance ineverystage theyare goingtobe through.For
example,inassessingthe studentstomake portfolioof essay,we needtocheckeverysingle
paperof the studentsone byone everyweek,andwhenithasbeenrevised,we have to
checkit again
8. -Expensive (Students:the studentsshouldprovide extramoneytoprepare the performance
such as costumesforrole play)
-Challengethe teachertomatchperformance assessmenttoclassroomgoalsandlearning
objectives
Implementationonlearning
(ByusingItemAnalysistocheckthe performance of students)
ITEM ANALYSIS
Itemanalysisisa methodthatisused in educationtoevaluate testitem.Thiscanensure
that questionsare inappropriate standardandmeasure the effectivenessof individualtest
item.Itemanalysisispurposedtoimprove testitemsandidentifyunfairorbiaseditem.Item
analysisisconductedtoavoidthe chance of teststhat are too difficult(andhave an
insufficientfloor) tendtoleadtofrustrationandleadtodeflatedscores,andteststhatare
too easy(andhave an insufficientceiling) facilitate adecline inmotivationandleadto
inflatedscores.
Itemanalysiscanbe analyzedbycomputing:DifficultyIndex,DiscriminationIndex,Validity
Coefficient,andEffectivenessof Distraction.
A. DifficultyIndex
Accordingto Wilson(2005), itemdifficultyisthe mostessential componentof itemanalysis.
Itemdifficultyisdeterminedbythe numberof peoplewhoansweraparticulartestitem
correctly. It isimportantfora testto containitemsof variousdifficultylevelsinorderto
distinguishbetweenstudentswhoare notpreparedat all,studentswhoare fairlyprepared,
and studentswhoare well prepared.
To compute level of difficultywe use the formula:
DifficultyIndex (p) =C/T
p = DifficultyIndex
C = the numberof studentswhoansweritemXcorrectly
T = the numberof total studentswhoansweritemX
B. Discrimination Index
Discriminationgoesbeyonddeterminingthe proportionof people whoanswercorrectlyand
looksmore specificallyatwhoanswerscorrectly.Inotherwords,itemdiscrimination
determineswhetherthose whodid wellonthe entire testdidwellona particularitem.An
itemshouldinfactbe able todiscriminate betweenupperandlowerscoringgroups.
9. To compute discriminationindex,the firstthingwe have todoisdividingtest-takersinto
twogroups,uppergroup andlowergroup.Thenusingthisformula
Discriminationindex(D) =Pu – Pl
D = Discriminationindex
Pu = Level difficultyof itemXfromthe uppergroup
Pl = Level difficultyof itemXfromlowergroup
C. CoefficientValidity
Coefficientvaliditycan be computedbyusingcorrelation.There are twotechnique of
correlationwhichispopulartobe used.Theyare Point-biserial technique andBiserial
technique
D. The EffectivenessofDistraction
Multiple choice testshave one questionandseveraloptions of answer.Amongthe options,
there isonlyone answerwhichiscorrect,and the other optionsare wronganswers.Those
wronganswersinmultiple choice testsare calleddistraction.A distractioniscalledeffective
whenthere are a lot of studentschoosingit. AccordingtoFernandez(1984) a distraction
can be calledasa good distractionwhenthere are 2% of testtakerschoosingit.
My opinion
Teachersmust take care of some problemstheymay
fall inwhentheyare dealingwithtestssuchas(validity,scope,efficiencyandtestinginthe
contextof instruction)
Conclusion
Testsmay have adverse effectsonlearning –particularlyexternal tests.sometimestests
measure objectivesthatare differentfrominstructional objectives,andstudentsshape their
learningeffortsinaccordance withanticipatedtestmeasurements.The nature of testsoften
influencesteachers'behavioralso.They"teachfortest"some sayinthe form of a complaint.
On the otherhand,good testsmayhave beneficialeffectsonlearning.