Introduction
ThisisanactionresearchstudyonanintroductoryEFLcourseforJapaneseuniversity students.Themajor targetskillarea wasreading comprehension,and participantsread English-languageessaysofjournalisticstyle.Thecoursewasbasicallymeaning-focused,and studentsstrovetounderstandwrittenEnglishandenrichtheirvocabularythroughreading comprehensionactivities.However,theteacher,whowastheresearcherhimself,arrangedto integrate some form-focused tasks into the task cycle because opportunities to produce outputarebelieved to facilitateL2acquisition (Swain,1985,1993,1995).Theparticipants engaged in oralcloze and sentence-levelwriting tasks,which can be regarded as an intermediary between focus-on-form and focus-on-formS treatments.These form-focused tasks were chosen,instead ofa communicative speaking task,for the reason thatthe participantshad chosen thebeginning-levelreading course,nota communicativespeaking course.
Thelanguage-enhancementtaskswereadministeredalternatively atthebeginning and attheendofaclasssession.Targetlinguisticformswerelexicalphrases:i.e.,multi-word unitsthatarerememberedandretrievedasunanalyzedwholes(Nattinger,1980;Nattinger & DeCarrico,1992;Lewis,1993,1997).They arealsoknown asformulaicsequences(Wray, 2000,2002:Schmitt& Carter,2004)ormultiple-worditems(Moon,1997).Nativespeakersof a languagehavea largestock oflexicalphrasesthatfacilitatetheirfluentand accurate delivery ofspeech (Ellis,2005;Wray,2002),and theacquisition ofsuch formulaicphrases andsentencestructuresislikelytofacilitateL2learners・speechorwriting,aswellastheir 学苑英語コミュニケーション紀要 No.882 15～31（20144）
TheRol
eofLanguage-enhancementTasksi
n
anEFL Readi
ngComprehensi
onCourse
Yoshi
masaOgawa
Abstract
Thisisa pseudo-experimentalstudy to evaluatetheeffectsofform-focused l anguage-enhancementtasksonJapaneseEFLstudents・acquisitionoflexicalphrases.Participantswere 81 non-major students at a private Japanese university. Oral cloze and sentence-level compositiontaskswereadministeredalternatelybeforeandafterreadingcomprehensiontasks. Theresultsoftwo-way repeated-measuresANOVAsindicated thatthetasksfacilitated the participants・acquisitionoflexicalphrasesoverthesemester,butthetimingforadministering theform-focusedtasks(i.e.,atthebeginning,orattheend,ofaclasssession)didnothave anyimpactontheiracquisitionoftargetforms.
understanding oforalor written messages,in thetargetlanguage.Participantsin this study wereguided to learn severallexicalphrasesfrom each textbook unit.Themajor researchgoalsaretodetermine:(a)whetherornotthelanguage-enhancementtasksmight facilitatetheparticipants・acquisitionoflexicalphrasesand(b)whereinthetaskcycl eform-focused tasks can be best administered,i.e.,before or after meaning-focused reading activities.
Theparticipantswereenrolled in aprestigiousJapaneseuniversity,buttheirEnglish proficiencieswereclassifiedasthebeginninglevel(notbythenationalstandardsbutbythe university・sstandards)based on theplacementtestresults.Thereading materialsin the main textbook covered social,economic,religious,military,and politicalissues,and each unitcontained a numberofwordsthatearliergroupsofstudentshad found unfamiliar. Consequently,theacquisitionofsomeofthelexicalitemsparticularlylexicalphraseswas likelytoposeachallengetotheparticipantsinthisstudy.Ontheotherhand,whereasthey maynotbestronglymotivatedtolearnEnglishasaforeignlanguage,majoringindifferent academicsubjects,theirgeneralacademicproficienciesareveryhigh,andtheymayemploy theirownlearningstrategiesthataredifferent,forexample,from ESL learnerswhohave environmental support and have been involved in various communicative activities. Uncovering the exactprocess by which Japanese university students acquire words and lexicalphrasesisbeyondthereachandscopeofthisstudy,butthestudyaimstodetermine: (a)thedegreetowhich form-focused language-enhancementtasksfacilitatethisparticular studentgroup・sacquisitionoflexicalphrasesand(b)thebesttimingforadministeringthe language-enhancementtasksduringaclasssession.
Backgrounds Form-focusedTreatment
Form-focused instruction hasbeen differentiated from meaning-focused instruction in which learnersarenotguided to attend to any grammaticalruleorlinguisticitem.The targetunitsforform-focusedlearningcaneitherbethegrammaticalsystem (e.g.,syntactic rules)orexemplars(e.g.,pronunciations,vocabularyitems,morphologyendings,andlexical phrases).Themajorassumptionisthattheyareproblematicitemsthatpresentachallenge tolanguagelearners.
Form-focused instruction tends to be dichotomously divided into focus-on-form and focus-on-formS (Long,1991;Doughty & Williams,1998),butEllis(2001)categorized form-focusedinstruction intothreetypes:focus-on-formS,plannedfocus-on-form,andincidental focus-on-form.Moreover,thereisa continuum between focus-on-form and focus-on-formS becauseanexplicitanddirectpresentationofaphonological,lexical,orsyntacticform can beintegratedintoacommunicativeorinteractionallanguageactivityordesignedtoinvolve somelevelofcognitiveprocessing.
Theoralclozeand sentence-levelcomposition tasksadministered in thisstudy can be regardedasanintermediarybetween(planned)focus-on-form andfocus-on-formS.Theywere intendedtodraw learners・attentiontotargetformsinthereconstructiveprocesses,andthe language-focused tasks were connected to the meaning-focused reading comprehensive activitiesineachclasssession.However,theformswereprescribedandsomewhatintensively enhancednotspontaneously provided,noticed,orresponded to during interactionsand participantsstroveto achieveformalaccuracy by repeating targetlexicalphrasesseveral times or recycling the phrases that they had just been exposed to,which entailed a characteristicoffocus-on-formSinstructi on.Therefore,thesetasksarereferredtoasform-focusedtreatmentsinthispaper.
TimingforAdministeringForm-focusedTasks
NattingerandDeCarrico(1992),whoadvocatedinstructionoflexicalphrases,suggested thateventhefocus-on-formSinstruction,whichisoftenconsideredtobeineffectivebecause of its failure to activate cognitive processing, can facilitate language acquisition by kinestheticallyfamiliarizinglearnerswithnew grammaticalforms.Theyobservedthat:
There is nothing wrong with memorizing some essentialchunks,especially at the beginningstagesoflanguagelearning［...］Thechallengefortheteacherwouldbeto usesuchdrillstoallow confidenceandfluency,yetnotoverdothem tothepointthat theybecomemindlessexercise.(p.116)
In theirperspective,form-focusedinstruction (even ifitisfocus-on-formS-oriented)can be effectivelyutilizedtopreparelearnersformoredifficultcommunicativeormeaning-focused tasksinvolvingthetargetforms.
Ontheotherhand,Nunan(2004)andWillisandWillis(2007)arguedthatform-focused tasksmightcontributetolanguageacquisitionwhenlearnershavealreadybeenexposedto targetformsthrough meaning-focused activities.My position leanstoward Nunan・sand WillisandWillis・observation,whichsupportstheideaofpolishingtheformsthathavebeen atleastpartially learned through meaning-focused orcommunicativeactivities.However, eitherapproachcanbemoreeffectivedependingonparticularlearningcontextsorindividual learners・backgrounds,characteristics,orpreferences.Thepresentstudyfocusesonagroup ofnon-majorEFL studentsataJapaneseuniversity,whosegeneralacademicproficiencies are very high buttheir English proficiencies are comparatively lower by the pertinent university・sstandards.
ExampleCasesofForm-focusedTasksBeforeorAfterCommunicativeActivities
The present study involves form-focused language-enhancement tasks that are not genuinely communicative.However,theeffectsofform-focused tasksadministered either
beforeoraftermoremeaning-focused (orcommunicative)taskscan beillustrated by the followingthreestudies.
JonesandHaywood・sstudy(2004)exemplifiestheprocedurethatengageslearnersinthe meaning-focusedactivitiesfirstandtheninform-focusedactivitiesafterward.Theyexplored an effective way to teach formulaic sequences to undergraduate and graduate students enrolled in an English-for-academic-purposes course at the University of Nottingham (treatment group,n＝10;controlgroup,n＝11).Participants first enrolled in meani ng-focused reading activities.Then,afterbeing familiarized with thetextthrough meani ng-focusedreading,thetreatmentgroupreceivedafocus-on-form instruction:theoriginaltext waspresentedagainwithformulaicsequencestypographicallyenhancedinbolditalics,and theparticipantsengagedingap-fillingexercisesandessaywriting.Theresultsshowedthat theirawarenessofformulaicsequencesimproved,butnoevidencewasproducedthattheir abilitytoretrieveandusethetargetphraseshadimproved.
Asan exampleofaprogram thatintroduced form-focused activitiesfirst,Hagerand Lyman-Hager (2004) reported on a two-week in situ internship program that The PennsylvaniaStateUniversityandtheUniversityofArtoishadjointlysetupforAmerican engineering studentspreparing to engagein bi-nationalbusinessprojects.In thisspecial program inFrance,theparticipantswerepresentedwithusefulexpressionsfirst( focus-on-formS)and then engaged in outside-of-classtasksto practiceusing them (i.e.,meani ng-focusedcommunicativetasks)next.(Iacknowledgethattheprogram inFrancediffersfrom the presentstudy in thatform-focused tasks were followed by communicative speaking activities,notreading comprehension activities.Nonetheless,themajorpointisthateach training session began with form-focused tasks.) First, the participants learned and rehearsed using survival expressions in the classroom: two instructors predicted the expressionsthattheywouldhavetouse(e.g.,findingoutabouttransportation,purchasing foodortoiletries)andtailoredthelessonstotheirneeds.Then,theparticipantsmovedina van to different locations and practiced using useful expressions in various real-life communicative situations. It must be noticed, however, that, back in the van, the participantsdebriefed theexpressionsthey had used orthecommunicativepurposesthey hadfailedtoaccomplish,andtheinstructorsorotherstudentsofferedfeedback,whichwas a form-focused activity.In thissense,thetask cyclecan also beregarded asalternation betweenform-focusedandmeaning-focusedactivities.
Macias(2004)reportedonaSpanish-for-academic-purposesprogram atHartnellCollege inSalinas,California,whichwasmoreclearlyintendedasacyclicprocess.Theparticipants wereinvolvedin form-focusedandmeaning-focusedtasksalternately andrepeatedly.They startedwithasimplecommunicativetask(notform-focused)usingexpressionsoftheirown choice.Then,theywereprovidedwithbasicsentencesandsyntacticrulesandpracticedusing theforms.Atthefollowingclasssession,learnersreviewedthelearnedforms,forexample,
by introducing themselvesby using basicsentencestructures,readbasicreading materials forchildren,anddiscussedtheissuescoveredinreadingactivities.Theyfinishedthelesson with communicative activities,and,at the following session,the vocabulary items and grammaticalrules were reviewed before a new concept or theme was introduced.The instructionalcycle began with communicative,meaning-focused tasks,but,subsequently, form-focusedandmeaning-focusedtaskswereadministeredalternately.
ResearchQuestions
Thefollowingtworesearchquestionsguidedthepresentstudy.
Research Question 1:Doform-focused language-enhancementtasksfacilitateparticipants・ acquisitionofusefullexicalphrasesinthecontextofreadingcomprehensionlessons? Research Question 2:Do participants learn more lexical phrases from the l
anguage-enhancementtasksadministeredbeforemeaning-focusedreadingcomprehensionactivities, ordotheybenefitmorefrom thesametasksadministeredafterreadingactivities?
Method Participants
ParticipantswereenrolledinoneofthemostprestigiousprivateuniversitiesinJapan, andtheirgeneralacademicproficiencieswereveryhighasevidencedbythefactthatthey had won admission to the school.On the other hand,the pertinent EFL course was classified asan introductory (orbeginning-level)EFL coursebased on theplacementtest thatthedepartmentadministered priorto courseregistration.Theplacementtestscores they attained wereequivalentto 365-555TOEIC scores.TheEFL coursethatthey took, labeled as Basic-LevelEnglish:News Reading,was offered to two different groups of students,scheduledfortwodifferentclassperiodsonthesamedayofweek.Oneclasshad 39students,andtheother42;thetotalnumberofparticipantswas81.Theresearchstudy wasconductedon thesetwointactclasses.Theentireparticipantgroupincluded53fi rst-yearstudents,8second-yearstudents,10third-yearstudents,and10fourth-yearstudents. Sixty-nineofthem weremale,and12werefemale.
TheparticipantshadstudiedEnglishatJapanesejuniorhighschoolsandhighschools foratotalofsixyears;somehadspentoneortwoadditionalyearstakinghigh-school-level English coursesatpreparatory schoolsforentranceexaminations.TheirEFL learning in middleschools(and preparatory schools)had been heavily oriented toward theGrammar-TranslationMethod,designedmainlytoimprovetheirreadingcomprehensi onandgrammar-analysisskillsand toexpand theirvocabulary ofliterary English.Very few had received trainingstospeakorwriteEnglishforcommunicativepurposes.Allpartici pantswerenon-English-majors, majoring, instead, in Japanese literature, pedagogical sciences, social
sciences,culturalsciences,biology,mathematics,orhistoryandgeography. Materials
ThemaintextbookusedinthespringwasNew IssuesforGlobalChangeandChallenges (Hirota& Balsamo,2006).Theessaysin thetextbook,journalisticin style,werecarefully edited forEFL learners・useand included abalanced combination oftopics.On theother hand,the essays were based on news reports in the past and tended to be outdated. Therefore,additionalnewspaperarticlesonrecentsocialeventsweredistributedintheform ofXeroxedmaterials.However,theanalysisoflexical-phraseacquisitionwasbasedonlyon their reading ofthe textbook units thatwere regulated in terms oflength,style,and difficulty.Participantswereinstructedtoreadthetextbookbeforecomingtotheclass,but theteacher・sobservationinthepastyearsindicatedthatfew readthetextbookunitsvery carefullybeforecomingtoclass.Consequently,whenform-focusedinstructionwasconducted atthebeginningofaclass,formanyparticipants,itwastheirfirstoccasiontopayclose attentiontothetargetlexicalphrases.
InstructionalTreatments
Themajorportion ofclasstimewasspenton meaning-focusedreading comprehension activities.Theteacherprovidedthesameinstruction tothetwogroupsofstudentsusing exactly thesamematerials.Themeaning-focused reading comprehension practiceinvolved skimmingovertheassignedessay,ornewsarticle,forthedayandansweringcomprehension questions.Theteachercalledonsomeparticipantstointerpretthesemanticandpragmatic meaningsofcertain partsofan assigned reading and,whenevernecessary,explained the lexicalitemsorsyntacticstructuresthattheparticipantshadtroubleunderstandi ng.Word-by-wordtranslationofindividualsentencesorparagraphswasavoidedasmuchaspossible. In addition to thereading comprehension activities,theparticipantsengaged in two types of form-focused language-enhancement tasks,which were intended to draw their attentiontousefullexicalphrasesandguidethem torememberandusesomeofthem.The firstlanguage-enhancementtaskwasanoralclozeusingPowerPoint,modeledonblackboard reproduction (Nation, 1974) or progressive deletion (Willis & Willis, 2007). A set of PowerPointslideswereprepared.Thefirstslidepresentedasentencethatconveyedamajor pointinthetextbookunitthattheyreadandwasembeddedwithlexicalphrases.Thetarget lexicalphrases were highlighted in bold,and difficult syntactic or lexicalitems were explainedifnecessary.On thesecondandthirdslides,partsofthetargetlexicalphrases weredeleted,and,on thefourth slide,theentirelexicalphrasesand sometimesa few additionalcontentwordsinthesamesentenceweredeleted.Theparticipantsendeavoredto reconstructtheoriginalsentence.Thefollowingisanexampleofasetofslidesusedforthis task.
Slide1:English isnow thelinguafrancaoftheworld.Ithasbecomethelanguageof businessandtravelandthekeytojobadvancementandsecurity.
Slide2:English isnow thelinguafrancaoftheworld.Ithasbecomethelanguageof businessandtravelandthe( )( )jobadvancementandsecurity.
Slide3:English isnow thelingua ( ) oftheworld.Ithasbecomethelanguageof businessandtravelandthe( )( )jobadvancementandsecurity.
Slide4:Englishisnow the( )( )oftheworld.Ithasbecomethelanguageofbusiness andtraveland( )( )( )job( )andsecurity.
In this sentence,the key to was the major target lexicalphrase,underlined and highlightedin bold.Lingua franca andjobadvancementwereadditionalphrases,partsof which werealso deleted on thethird and fourth slidesfora remember-and-retrievetask. Many ofthesentencesdisplayed on thePowerPointscreen contained onelexicalphrase whereassomecontainedtwoormore.
Thesecondform-focusedtaskwastotranslateaJapanesesentenceintoEnglish.Itwas apartial-translationtask:i.e.,partsofeachtargetEnglishsentencewereprovidedashints. Participantsneeded to useoneofthelexicalphrasesthey had learned,orwould learn, during theday・slesson tocompletethistranslation task,andthey wereaskedtowriteit on a supplied sheetofpaperand submititto theteacher.Thecomposition task guided participantstomapatargetform onitssemanticfunction.
Theprocedureforthepartial-translationtaskwasasfollows.Theteacherdisplayed,on thebig screen,aJapanesesentenceandan equivalentEnglish sentencemissing thetarget lexicalphraseandseveralotherwords.Theparticipantsfilledintheblanksandreproduced theEnglishsentenceonasuppliedanswersheet.Then,theyexchangedtheiranswersheets withtheirneighborsandcorrectedeachother・smistakesreferringtothefeedbackthatthe teachersubsequently displayed on thescreen.Itwasannounced atthebeginning ofthe semesterthatthisin-classwriting waspartoftheirpracticetolearn lexicalphrases,and thescoresonthistaskwouldnotaffecttheirfinalgrades.However,thesubmissionofthe on-the-spotcompositionwasbelievedtopressurethem totakeonthetaskseriously.
Thefirstgroup engaged in theform-focused tasksattheend ofodd-numbered class sessionsandatthebeginningofeven-numberedsessions.Thesecondgroupengagedinthe sameform-focusedtasksatthebeginningofodd-numberedsessi onsandattheendofeven-numbered sessions.The contents ofthe lessons were thus counterbalanced.When form-focusedtaskswereadministeredatthebeginning,theparticipants(i.e.,exceptforthefew diligentoneswhohad carefully read thetextbook athome)had notbeen exposed tothe target lexical phrases. At the class sessions in which the form-focused tasks were administeredattheend,allparticipantshadbeen exposedtothephrasesthrough reading activities.
Assessment
A pretest, three mid-term quizzes, and a final test were conducted to measure participants・knowledge of lexicalphrases.The pretest measured the participants・base knowledgeoflexicalphrasesbeforereceiving any instruction.Thethreemid-term quizzes assessed their learning ofdifferentlexicalphrases atdifferenttime periods during the semester;themeansofthesetestswerebelieved to reflecttheirshort-term retention of targetlexicalphrases.Thefinaltestwasadministeredtwoweeksafterthelastmid-term quiz,andtheresultsmightreflecttheirlong-term retentionofphrases,althoughthelapse oftimeafterthelastclasssessionwasnotverylong.Thepretestandthefinaltestincluded thesamequestionitems,andthemid-term quizzesweresmallpartsofthepretestorfinal test.However,theorderofquestionswaschangedforeachtest,andthefinaltestcontained afew non-targetitemstopreventany testing effect.Each ofthethreetestsincludedtwo sections:(a)a multiple-choicesection thatrequired participantsto chooseoneoutofthe threealternativesaspartofthetargetlexicalphraseand(b)asentencecompletionsection thatrequiredparticipantstofillintheblankswithanappropriatelexicalphrasebasedon an equivalentJapanesephraseprovided asahint.Thesecond section wasintended tobe moredifficult.
Again,thetwointactclasses(enrolledin twocoursesadministeredduring consecutive classperiods)studiedthesametextbookandmaterialsandwereexposedtothesamelexical items,buttheyengagedintheform-focusedlanguage-enhancementtaskatthebeginningor attheendofaclassalternately.Theeffectivenessofthetwotreatmentswasevaluatedby administering a two-way repeated-measures ANOVA. Within-subjects comparison was conducted,insteadofbetween-subjectscomparison,inordertoadjusttheEnglishabilitiesof thetwointactclassesandalsototreatallstudentsfairlyfrom aneducationalandethical pointofview.
Additionally,theparticipantswereaskedtoindicatetheirpersonalpreferenceforeither treatmentintheform ofaninformalsurveyattheendofthesemester.Theiranswerswere counted,andachi-squaretestwasconductedtomeasurethedegreeofstatisticalsignificance.
Thelevelofsignificanceforallstatisticaltestswassetat・＝0.05. Results
Multiple-choiceTestResults
Priortotheadministrationofatwo-wayrepeated-measuresANOVA,theparticipants・ multiple-choicescoresonthepretest,mid-term quizzes,andfinaltestweretransformedinto Raschdigits,usingthedichotomousRaschmodel(Bond& Fox,2007).Raschmeasuresare moreusefulforaccuratestatisticalmeasurementthan raw scoresbecausethey areequal -intervalmeasures that are derived from the probabilistic relationships between person abilitiesanditem difficulties.
The Rasch analyses indicated that the person separation was 0.77,and the person reliability was0.37;both wereratherlow,which may beattributabletothefactthatthe participants,enrolledinthebeginning-levelEFLcourseatthesameuniversity,weresimilar in English proficiency.Theitem separation was4.28,which wasmuch higher than the criterionpointof2.0,andtheitem reliabilitywas0.95,alsoabovethecriterionpointof0.90.
TheRaschpersonmeasureswereconvertedtoresponseprobabilityunits(CHIPS).This lineartransformationmeantthattheaveragepersonmeasurewassetat50,andthehighest and lowest possible scores were respectively 80 and 20.The finalN-size for statistical analyseswas65.Table1displaysthedescriptivestatisticsforthemultiple-choicetests.
Then,atwo-waywithin-subjectsANOVA wasconductedusingtheseconvertedscoresto evaluatetheeffectsoftreatmentandteston theparticipants・acquisition oftargetlexical phrases.Thewithin-subjectsfactorsweretreatmentwith twolevels(language-enhancement atthebeginning orattheend)andtestwith threelevels(pretest,mid-term quizzes,and finaltest).Thedependentvariablesweretheparticipants・scoreson thethreetestsafter
Table1.DescriptiveStatisticsforMultiple-choiceTests
Beginning End
Pretest M 49.54 49.98 95％CI LowerBound 48.55 48.97 UpperBound 50.53 50.98 SD 3.99 4.06 Skewness －0.21 －0.45 SES 0.30 0.30 Kurtosis －0.28 0.37 SEK 0.59 0.59
Mid-term Quizzes M 54.22 54.73 95％CI LowerBound 53.19 53.40 UpperBound 55.25 56.06 SD 4.15 5.38 Skewness 0.84 0.39 SES 0.30 0.30 Kurtosis 1.39 －0.47 SEK 0.59 0.59 FinalTest M 60.88 61.38 95％CI LowerBound 59.42 60.03 UpperBound 62.33 62.73 SD 5.86 5.46 Skewness －0.36 －0.44 SES 0.30 0.30 Kurtosis －0.98 －0.89 SEK 0.59 0.59 Note.N＝65.
eachtreatment.
Mauchly・s Sphericity Testindicated thatthe assumption ofsphericity was notmet eitherforthetestfactor,W＝0.87,p＝0.01,orforthetreatmentxtestinteraction,W＝0.90, p＝0.04.Consequently,theGreenhouse-Geisserstatisticswasreferredtointheinterpretation ofANOVA results.Thesphericity assumption didnotapply tothetreatmentfactorthat hadonlytwolevels.
Thetimemaineffectandthetimextreatmentinteractioneffectweretestedusingthe multivariatecriterionofWilks・slambda(Λ).Thetestmaineffectwassignificant,Λ＝0.20, F(2,63)＝127.54,p＝0.001,・2＝0.80,butthe testx treatmentinteraction effect was not
significant,Λ＝1.00,F(2,63)＝0.01,p＝0.99,・2＝0.00.
Likewise,theresultsoftheunivariatetest(seeTable2)showedthatthetestmaineffect wassignificant,F(1.76,112.91)＝172.02,p＝0.001,・2＝0.73,butthetreatmentmaineffectwas
not significant, F(1, 64)＝1.57, p＝0.22, ・2＝0.02, suggesting that the timing for the
administrationoflanguage-enhancementtasks(i.e.,eitheratthebeginningorattheendof alesson)didnothaveany significantinfluenceon learners・acquisition oflexicalphrases. Thetreatmentxtimeinteractioneffectwasnotsignificant,F(1.82,116.31)＝0.01,p＝0.99, ・2＝0.00.
Inordertofollow upthesignificanttestmaineffect,themeansforthethreetestswere computed,and threepaired-samplest-testswereconducted.Holm・ssequentialBonferroni approach wasused tocontrolforfamilywiseerrorrateacrossthesetests.Themean for Test3(M＝61.13,SD＝0.64)wassignificantly higherthan thatforTest1(M＝49.76,SD ＝0.40),t(64)＝16.07,p＝0.001,d＝22.18,the mean for Test 2 (M＝54.48,SD＝0.48)was significantly higherthan thatforTest1,t(64)＝9.25,p＝0.001,d＝9.18,and themean for Test3wassignificantly higherthan thatforTest2,t(64)＝10.82,p＝0.001,d＝13.54.The d-values indicated thatthe effectsizes were very large.Thatis,the scores,for either treatment,improvedprogressivelyfrom thepretesttothemid-term quizzestothefinaltest.
Table2.UnivariateTestResultsoftheTwo-wayRepeated-measuresANOVA (Multiple-choiceTests) Effect df SS MS F p ・2 Treatment 1 22.75 22.75 1.57 0.22 0.02 Residual 64 928.94 14.51 Test 1.76 8483.96 4809.10 172.02 0.001 0.73 Residual 112.91 3156.46 27.96 TreatmentxTest 1.82 0.11 0.06 0.01 0.99 0.00 Residual 116.31 1528.47 13.14 Note.・＝0.05.
SentenceCompletionTestResults
Theparticipants・sentencecompletiontestscoresforthepre-test,mid-term quizzes,and thefinaltestwerefirsttransformedintoRaschdigits,usingthepartial-creditRaschmodel (Bond & Fox,2007).Theperson separation was1.41,and theperson reliability was0.67; neitherreachedthedesirablelevelof2.00or0.90.Theitem separationwas9.89,andtheitem reliability was0.99,which wereboth very high.Asin theanalysisofmultiple-choicetest scores,theRasch digitswereconvertedtoresponseprobability units.ThefinalN-sizefor statisticalanalyses was 66.Table 3 displays the descriptive statistics for the sentence completiontests.
Themeansforthetwotreatments(i.e.,language-enhancementeitheratthebeginning orattheendofalesson)wereverysimilarateachofthethreetests.Ontheotherhand, themeanforeachtreatmentincreasedfrom thepretesttothemid-term quizzestothefinal testtoanoticeabledegree.
Then,a two-way within-subjects ANOVA was conducted to evaluate the effects of
Table3.DescriptiveStatisticsforSentenceCompletionTests
Beginning End
Pretest M 43.08 43.08 95％CI LowerBound 41.46 41.51 UpperBound 44.70 44.65 SD 6.59 6.39 Skewness －0.22 －0.70 SES 0.29 0.29 Kurtosis －0.05 0.98 SEK 0.58 0.58 Mid-Term M 50.24 51.24 95％CI LowerBound 48.30 49.51 UpperBound 52.18 52.98 SD 7.88 7.06 Skewness －0.84 －0.15 SES 0.29 0.29 Kurtosis 1.39 0.84 SEK 0.58 0.58 FinalTest M 62.94 62.84 95％CI LowerBound 61.31 61.28 UpperBound 64.56 64.40 SD 6.62 6.35 Skewness －0.10 0.18 SES 0.29 0.29 Kurtosis －0.66 －1.17 SEK 0.58 0.58 Note.N＝66.
treatmentand teston theparticipants・acquisition oftargetlexicalphrases.Thewithi n-subjectsfactorsweretreatmentwithtwolevels(language-enhancementatthebeginningor at the end)and test with three levels (pretest,mid-term quizzes,and finaltest).The dependentvariablesweretheparticipants・scoresonthethreetestsforeachtreatment.
Mauchly・sSphericityTestresultsshowedthattheassumptionofsphericitywasmetfor both the testfactor (W＝0.99,p＝0.79)and the textx treatmentinteraction (W＝0.93, p＝0.11).
Themultivariatetestresultsshowedthatthetestmaineffectwassignificant,Λ＝0.09, F(2,64)＝310.49,p＝0.001,・2＝0.91.Ontheotherhand,thetreatmentxtestinteractionwas
notsignificant,Λ＝0.98,F(2,64)＝0.56,p＝0.57,・2＝0.02.
Theunivariatetestresults(seeTable4)also showed thatthetestmain effectwas significant, F(2, 130)＝296.28, p＝0.001, ・2＝0.82. The treatment main effect was not
significant,F(1,65)＝0.34,p＝0.56,・2＝0.01,and thetreatmentx testinteraction wasnot
significant,F(2,130)＝0.49,p＝0.62,・2＝0.01,either.
Inordertofollow upthemaintesteffect,themeansforthethreetestswerecomputed, andthreepaired-samplest-testswereconducted.Holm・ssequentialBonferroniapproachwas usedtocontrolforfamilywiseerrorrateacrossthesetests.ThemeanforTest3(M＝62.89, SD＝0.72)wassignificantlyhigherthanthemeanforTest1(M＝43.03,SD＝0.60),t(65)＝25.04, p＝0.001,d＝29.88,themean forTest2(M＝50.74,SD＝0.82)wassignificantly higherthan themeanforTest1,t(65)＝9.36,p＝0.001,d＝10.66,andthemeanforTest3wassignificantly higherthanthemeanforTest2,t(65)＝14.27,p＝0.001,d＝15.73.Theparticipants・scoreson thesentencecompletiontestimprovedsignificantlyfrom thepretesttothemid-term quizzes tothefinaltest.
Chi-squareTestResults
A questionitem inthequestionnairesurveyconductedattheendofthesemesterasked theparticipantswhethertheypreferredtoengageintheform-focusedlanguage-enhancement tasksatthebeginningofaclasssession(i.e.,beforereadingcomprehensionactivities)orat
Table4.UnivariateTestResultsoftheTwo-wayRepeated-measuresANOVA (SentenceCompletionTests) Effect df SS MS F p ・2 Test 2 26329.62 13164.81 296.28 0.001 0.82 Residual 130 5776.31 44.43 Treatment 1 9.03 9.03 0.34 0.56 0.01 Residual 65 1748.91 26.91 TestxTreatment 2 24.49 12.25 0.49 0.62 0.01 Residual 130 3270.07 25.15 Note.・＝0.05.
theend.Outofthe81students,30indicatedthattheypreferredtohavelanguage-enhancement atthebeginning,and33preferredthetreatmentattheend;interestingly,theirpreferences werealmostevenlydivided.Seventeengavenoanswer,eitherhavingnopreferenceorsimply being obliviousto thequestion,whereasoneparticipantexpressly indicated thathe/she preferredtohaveitalternatelyatthebeginningandattheendofasession.
A one-waychi-squaretestwasconductedtoassessthedegreeofstatisticalsignificance. Theresultofthetestwassignificant,χ2(3,N＝81)＝31.54,p＝0.001.Inordertofollow upthe
significanteffect,sixpairwisecomparisonswereconducted;theresultsareshowninTable 5.Itwasnotsurprising thatthereweresignificantdifferencesbetween alternation (the observedfrequencyofwhichwasonlyone),ontheonehandandbefore(30),after(33),and no response(17)on theother.However,therewasno significantdifferencebetween the portionsofpeoplewho preferred thetreatmentatthebeginning and thosepreferred the treatmentattheend,χ2(1,N＝63)＝0.14,p＝0.71.
Discussion
Thefirstresearchquestionwas:Doform-focusedlanguage-enhancementtasksfacilitate participants・acquisition ofusefullexicalphrasesin thecontextofreading comprehension lessons? Thestatisticalresultsindicated thattherewasa significantprogressfrom the pretesttothemid-term quizzesandtothefinaltest.Thus,itissafetoassumethattheoral clozeand simplecomposition taskscontributed to theparticipants・acquisition oftarget lexicalphrases.Asacknowledgedatthebeginning ofthispaper,thesetasksarenotofa communicativetype,although designedtoinducesomecognitiveprocessing.Thetasksare simply designed to help participantsrememberand retrieveusefullexicalchunks,butit mustbenoted again thatalargestock ofusefullexicalphraseshelpsEFL/ESL learners carryoutcommunicativeinteractions(Ellis,2005;Wray,2002).
In thepertinentresearch context,thefinal-testresultsshould beinterpreted asthe accumulationoflexicalknowledgeduetotheirrepeatedexposuretothetargetformsover time,notacarry-overeffectofthetreatmentaftertheinstructionaltreatmentended.As partofthestatisticalmeasurement,an arrangementwasmadetopreventtesting effects.
Table5.ResultsoftheFollow-upChi-squareTests χ2 p Beforevs.After 0.14 0.71 Beforevs.NoResponse 3.6 0.06 NoResponsevs.After 5.12 0.2 Beforevs.Alternation 27.13 0.001 NoResponsevs.Alternation 14.22 0.001 Aftervs.Alternation 30.12 0.001 Note.df＝1;・＝0.05.
However,beingexposedtothesameformsthroughtasksandtests,theparticipantsmight haveconsolidated theirmemory oftargetphrasestoward theend ofthesemester.Their mid-term quizzeswerereturned to theparticipantsbeforethefinaltest,and they had a chancetoreview thelexicalphrasesthatthey had learned during thesemester.Itmight havenotbeenanidealconditionforstatisticalassessment,butitisanecessarypartofthe educationaltreatment.Theparticipantshad therighttohavetheearlierquizzesortests back as feedback,and they had been informed atthe beginning ofthe semester that, whereasthequizzeswerepartoftheirlearningprocedureandwouldnotaffecttheirgrades, theirfinaltestscoreswoulddo.
Thesecondresearchquestionwas:Doparticipantslearnmorelexicalphrasesfrom the language-enhancement tasks administered before meaning-focused reading comprehension activities,ordotheybenefitmorefrom thesametasksadministeredafterreadingactivities? Thestatisticaltestresultssuggestedthatthetimingfortheadministrationofform-focused instruction did not influence their learning of lexical phrases. At each test, their performance was practically the same whether they engaged in the oral cloze and composition tasks before or after reading activities.Furthermore,practically the equal proportions of students indicated their preference to receive the language-enhancement treatmentatthebeginning ortheendofalesson.Neithertask sequenceappealedtothe majorityofstudents.
Oneinterpretationisthatthelanguage-enhancementtaskswerenotsubstantialenough to influenceparticipants・learning oflexicalphrasesor thatthesetasks,which did not involveinterpersonalinteractions,higher-levelcognitiveprocessing,orcreativelanguageuse, werenotadequateor appropriate.However,another possibleinterpretation isthatboth treatments (i.e.,language-enhancement before reading or after reading)had their own positiveeffectsand,when measured statistically,canceled outthegainsand losses.The latterpositioncanbe,atleastpartially,supportedbythefactthatsomeparticipantsgave positivecommentsononetreatmentandothersreactedpositivelytotheother.Forexample, four of the participants who indicated their preference for the language-enhancement treatmentatthebeginningspecificallystatedthatitwashelpfultoknow whatthemajor targetlinguisticpointswerebeforeengaginginthemeaning-focusedactivities.Threeothers voluntarily offeredpositivecommentsforthesametreatment.On theotherhand,fourof theparticipantswhopreferredthetreatmentattheendobservedthattheycouldunderstand themeaningsofthelexicalphrasesbetterafterbeingexposedtothem incontext,andsix indicatedthattheycouldmemorizetargetlexicalphrasesmoreeasilyafterfindingthem in context.Their observations resonate with Nunan・s (2004)and Willis and Willis・s (2007) position. Five others voluntarily offered positive comments for the latter treatment. Furthermore,as mentioned in the results section,one participant explicitly wrote that he/shepreferredtohaveitalternatelyatthebeginningandattheendofasessiondespite
thefactthatsuch an answer had notbeen provided asa prescribed alternativein the questionnaire.
Ifthesecondinterpretation (i.e.,both treatmentshavepositiveeffects)holdstrue,the safestpedagogicalarrangementistousethetwotreatmentsalternately or,asHagerand Lyman-Hager・s(2004)andMacias・s(2004)reportssuggested,tointegratetheform-focused and meaning-focused (or communicative)tasks cyclically within the same EFL course. However,there is no doubt that the same issue needs to be further probed through replicationstudieswithdifferentparticipantsengagedindifferentlanguage-enhancementor communicativeform-focusedactivities.
Conclusion
Theresultsofthepresentstudy suggestedthatformed-focusedlanguage-enhancement taskswouldfacilitateJapaneseEFLuniversitystudents・acquisitionoflexicalphrasesinthe context of reading comprehension lessons.However,the effectiveness of the l anguage-enhancementtasksused in thisstudyoralclozeand sentence-levelcompositiondid not differdepending on whetherthey wereadministered beforeorafterthemeaning-focused readingcomprehensionactivities.Nodecisiveevidencewasfoundtoindi catethattheseform-focusedtasks,somewhatfocus-on-formS-oriented,canbetterbeutilizedtoprepareJapanese non-majorEFL studentsformoremeaning-focused activitiesorforfine-tuning theforms withwhichtheyhavealreadybeenfamiliarized.
Itisacknowledgedthattheform-focusedtasksadministeredandevaluatedwerelimited tothememorization andretrievalofphrasesin sentence-orparagraph-levelcontexts.For futurestudies,itwillbeworthwhileto observeand analyzetheperformanceoflearners engagedin morecommunicativetasksin which they needtochoosetheirown wordsand expressionstoexpresstheirideasortodealwithreal-lifelinguisticsituations.Analysisof theirlanguageuseininterpersonalcommunicationwouldbemoredifficultandcomplicated, butitispedagogicallymoremeaningfulinthatparticipantshavechancestonoticethegaps betweentheiroutputandnative-liketargetformsmoreclearly.
Regarding thematerialsforreading tasks,replication studiesshould beconducted on students・reading ofup-to-datenewspaperarticlesdistributedin class,insteadofessaysin thetextbook published severalyearsbefore.Ifthenewsareup-to-dateand moreclosely relatedtotheirdaily concerns,theirmotivation tounderstandthereadingsislikely tobe higher,whichisanimportantconditionformeaning-focusedreadingactivities.Theymaybe motivatedtolearnlexicalphrases,oranyothertargetlinguisticforms,thatareneededto describetheissues,ideas,oreventsoftheirconcern.
Evaluating the relationships between learners・individualproficiencies or l anguage-learningexperiencesandtheiracquisitionoflexicalphrasesisyetanotherissuetobedealt with in futurestudies.Thisstudy only observed a caseatoneJapaneseuniversity,but
comparison between different student groups with clearly different proficiencies and characteristicsmightshedlightontheprocessbywhichJapaneseEFLlearnersmayacquire lexicalphrasesorlexicalitemsingeneral.
References
Bond,T.G.,& Fox,C.M.(2007).ApplyingtheRaschmodel:Fundamentalmeasurementinthehuman sciences.Mahwah,NJ:LawrenceErlbaum.
Doughty,C.,& Williams,J.(1998).Issuesandterminology.InC.Doughty,& J.Williams(Eds.),Focus on form in classroom second languageacquisition (pp.111).Cambridge:CambridgeUniversity Press.
Ellis,R.(2001).Introduction:Investigating form-focused instruction.In R.Ellis(Ed.),Form-focused instructionandsecondlanguagelearning(pp.146).Malden,MA:Blackwell.
Ellis,R.(2005).Principlesofinstructedlanguagelearning.System,33,209224.
Hager,W.R.,& Lyman-Hager,M.A.(2004).Bridging thegap between thesciencesand humanities: French forengineersandothertechnicalprofessions.In B.L.Leaver,& J.R.Willis(Eds.) ,Task-basedinstructioninforeignlanguageeducation:Practicesandprograms(pp.161177).Washington D.C.:GeorgetownUniversityPress.
Hirota,N.,& Balsamo,W.M.(2006).New issuesforglobalchangeand challenges.Tokyo:Kinseido Publishing.
Jones,M.,& Haywood,S.(2004).Facilitating theacquisition offormulaicsequences:An exploratory studyinanEAPcontext.InN.Schmitt(Ed.),Formulaicsequences:Acquisition,processing,anduse (pp.269300).Philadelphia:JohnBenjaminsPublishingCompany.
Lewis,M.(1993).Thelexicalapproach.Hampshire,UK:HeinleCengageLearning.
Lewis,M.(1997).Implementing thelexicalapproach:Putting theory into practice.Hampshire,UK: HeinleCengageLearning.
Long,M.H.(1991).Focusonform:A designfeatureinlanguageteachingmethodology.InK.DeBot, R.B.Ginsberg,& C.Kramsch (Eds.),Foreignlanguageresearchincross-culturalperspective(pp.39 52).Philadelphia:JohnBenjaminsPublishingCompany.
Macias,C.(2004).Task-basedinstructionforteachingSpanishtoprofessionals.InB.L.Leaver,& J.R. Willis(Eds.),Task-basedinstructioninforeignlanguageeducation:Practicesandprograms(pp.142 160).WashingtonD.C.:GeorgetownUniversityPress.
Moon,R.(1997).Vocabularyconnections:Multi-worditemsinEnglish.InN.Schmitt,& M.McCarthy (Eds.),Vocabulary:Description,acquisition and pedagogy(pp.4063).Cambridge,UK:Cambridge UniversityPress.
Nation,I.S.P.(1974).Motivation,repetitionandlanguageteachingtechniques.ELTJournal,29(2),115 120.
Nattinger,J.R.(1980).A lexicalphrasegrammarforESL.TESOL Quarterly,14(3),337344.
Nattinger,J.R.,& DeCarrico,J.S.(1992).Lexicalphrasesand languageteaching.Oxford:Oxford UniversityPress.
Nunan,D.(2004).Task-basedlanguageteaching.Cambridge:CambridgeUniversityPress.
Schmitt,N.,& Carter,R.(2004).Formulaicsequencesin action:An introduction.In N.Schmitt(Ed.), Formulaic sequences:Acquisition,processing,and use (pp.122).Philadelphia:John Benjamins
PublishingCompany.
Swain,M.(1985).Communicativecompetence:Somerolesofcomprehensibleinputandcomprehensible outputin itsdevelopment.In S.Gass,& C.Madden (Eds.),Inputin second languageacquisition. (pp.235253).Rowley,MA:NewburyHouse.
Swain,M.(1993).Theoutputhypothesis:Justspeaking and writing aren・tenough.TheCanadian ModernLanguageReview,50,158164.
Swain,M.(1995).Threefunctionsofoutputin secondlanguagelearning.In G.Cook,& B.Seidlhofer (Eds.).Principlesandpracticeinappliedlinguistics:StudiesinhonourofH.G.Widdowson (pp.125 144).Oxford:OxfordUniversityPress.
Willis,D.,& Willis,J.(2007).Doingtask-basedteaching.Oxford:OxfordUniversityPress.
Wray,A.(2000).Formulaicsequencesin second languageteaching:Principleand practice.Applied Linguistics,21(4),463489.
Wray,A.(2002).Formulaiclanguageandthelexicon.Cambridge:CambridgeUniversityPress.
