138 RitsumeikanSocialSciencesReview(Vol53.No.1) OncesuchacodingruleisenteredintoKH Coder,notonlythedocumentscontaining Gilbert butalso thosecontaining

Transcription

1 第 53 巻第 1 号 A Two-StepApproachtoQuantitativeContentAnalysis(PartI)(HIGUCHIKoichi) RitsumeikanSocialSciencesReview 2017 年 6 月 137 ATwo-StepApproachtoQuantitativeContentAnalysis: KH CoderTutorialUsingAnneofGreenGables(PartII) HIGUCHIKoichi ⅰ Abstract:Thisarticleintroducesatwo-stepapproachtoperformingquantitativecontentanalysisoftext data.first,anoutlineoftheapproachisbrieflydescribed.second,theprocedureofusingtheapproach toanalyzethenovelanneofgreengablesisdescribedasatutorial.third,thefeaturesoftheapproach arediscussedwithreferencetotheresultsoftheanalysis. Thetutorialsectionofthisarticlealowsreaderstosimulatethesameanalysisontheirown personalcomputers.weusefreesoftwareandmostofthenecessaryoperationsareilustratedinfigures. ThesubjectoftheanalysisisthepopularnovelAnneofGreenGables.Itispointedoutthattheheroine Anne sfostermothermarilaplaysanessentialroleinthenovelandthatmarilaismoreimportantthan Anne sbestfrienddiana,andgilbertwithwhom Annehasafaintromance.Intheanalysisofthe tutorial,weexaminewhetherthequantitativeanalysisbasedonthetwo-stepapproachalsoilustratesthe importanceofmarila. Thesecondhalfofthisarticleispublishedhere.ThefirsthalfhasbeenpublishedinVolume52, Issue3ofthisbuletin. Keywords: quantitativecontentanalysis,kh Coder,AnneofGreenGables,tutorial 5Step2:FocusingonMarila 5.1ComposingCodingRules InStep1,tablesandfiguresarecreatedinawaylesssusceptibletotheinfluenceoftheuser s prejudicesandpreconceptions.instep2,however,whichisdescribedinthissection,theuser sviewpoint wilbeutilizedandreflectedintheanalysis.nevertheless,whilereflectingtheuser spointofview and interpretation,theprocessofanalysisshouldbekeptopentoverificationandcriticism bythirdparties.in otherwords,theuser spointofview shouldbereflectednotimplicitlybutexplicitlyintheanalysis. Composingcodingrulesisaspecificmethodforperforminganalysisinthatway.Aseriesofsuchcoding rulesissometimescaleda dictionary forcoding. Aftercoding,theusercancounttheappearanceoftheconceptorcategorythathe/shefocuseson insteadofthatofeachword.forexample,gilbert,acharacterinanneofgreengables,issometimes referredtoas Gilbert andsometimesas Gil.Theusercancountbothnamesasanappearanceofa concept CharacterNameGilbert bycomposingthefolowingcodingrule: *Character_Name_Gilbert GilbertorGil ⅰ AssociateProfessor,FacultyofSocialSciences,RitsumeikanUniversity

2 138 RitsumeikanSocialSciencesReview(Vol53.No.1) OncesuchacodingruleisenteredintoKH Coder,notonlythedocumentscontaining Gilbert butalso thosecontaining Gil areassignedthecode *Character_Name_Gilbert,andthenyoucancountthe appearanceofthecode.youcancomposeasmanyrulesasyouneedandcountmultipleconcepts.ifone documentsatisfiesthecriteriaformultiplecodingrules,multiplecodeswilbeatachedtothedocument. CodingofKH Coderisbasedontheideaofextractingconceptsfrom adocumentratherthanclassifyinga documentintoasinglecategory.thisconceptconsidersthatadocumentcancontainnotonlyonebutalso multipleconcepts. Thereareseveralpointstobenotedwhencomposingcodingrules.First,wordsnotcontainedinthe dataandthosenotextractedaswordsbykh Codercannotbecountedeveniftheyaredesignated. Therefore,usersarerecommendedtocheckwhetherawordappearsinthedatabyreferringtotheword list(section4.2)and/orbyusingthewordsearchfunction(goto[tools][words][search]inthemenu) beforeusingthewordinacodingrule. Second,usersshouldconfirm whatdocumentacodeisactualyassignedwhencreatingthecodingrule. ThedocumentsearchfunctionpresentedinFigure12isausefulmeansofconfirmation.Procedure(2)in Figure12showsthatthetextfileatachedtothetutorialnamed code_1.txt hasbeenopened.thistextfile containscodingrules,suchas *ANNE, *Marila,and *Mathew.Inprocedure(3)inFigure12,users candouble-clickanyofthecodestoretrievethedocumentsgiventhatcode.hereuserscanalsodoubleclick #none toretrievethedocumentsgivennocodes. Third,asageneralrule,codingrulesshouldbemadepublictoalowthirdpartiestoverifywhetherthe researcherhasextractedtheconceptsfrom thedatainareasonablemanner.evenifthereisinsuficient spacetolistalthecodingrules,usersshouldpresentsomeofthemainwordsineachcodingrule,and disclosealthecodingrulesifrequestedaftertheresearchispublished. Figure12:Retrievedocumentsassignedaspecificcode

3 A Two-StepApproachtoQuantitativeContentAnalysis(PartI)(HIGUCHIKoichi) CharactersinEachChapter Inthissection,wecounthow manytimesmarilaappearsineachchapterbycomposingcodingrules. Strictlyspeaking,wecounthowmanysentencescontainthename Marila ineachchapter.insection4.4, wherewedividedthestoryintofourparts,wefoundthatmarilaappearsevenlythroughouttheparts.we thenask,howdoessheappearineachofthe38chapters?letuscomparethepresenceofmarilawiththat ofothercharacterschapter-by-chapteringreaterdetail. Figure13showstheprocedureforcrosstabulatingtheresultsofcodingwithchapternumber.As Sentences isselectedasthe CodingUnit inprocedure(3)infigure13,eachsentenceisjudgedifit meetscriteriaofcodingrules.thus,sentencescontainingmarilaandothercharacternamesarecounted. Forexample,theresultpaneinFigure13showsthat23sentencescontainMarilainChapter01,whichis 16.91%ofatotalof136sentencesinthechapter.However,itisdificulttoreadsuchnumbersforalthe38 chapters,sowecancreateabubbleplotasshowninfigure14byprocedure(5)infigure13. Figure13:Crosstabulationofcodingresultsandgenerationofbubbleplots Figure14visualizesthepercentageofthesentencescontainingthenameofeachcharacterforeach chapter.theareaofeachsquareisinproportiontothepercentage.furthermore,thecolordensityindicates thedegreeofdiferencewhencomparedtootherchapters. Figure14showsthatMarilaisliteralypresentthroughouttheentirestory.Marilaappearsinfarmore chaptersthananyothercharacterexceptfortheheroineanneandalmostaswidelyasanne.marilaisthe onlycharacterwhoappearsthroughoutthestoryexceptanne,whichalsosuggeststheimportanceofher role. Chapter35istheonlychapterwhereMariladoesnotappear.InsteadofMarila,Gilbertappears frequently.inthischapter,anneislodginginatowntostudyatqueen sacademyandacademicaly competeswithherrivalgilbert.meanwhile,marilastaysinavonleavilageanddoesnotappearinthis chapter.althoughtheywerephysicalyapart,therewasanemotionalreunioninthefolowingchapter36. Annewonascholarshipforbeinganoutstandingstudentandrejoicedsaying Oh,won tmathew and Marilabepleased! AlsoMarila smiledafectionatelyathergirl andexpressedherlovetoannewho

4 140 RitsumeikanSocialSciencesReview(Vol53.No.1) Figure14:Appearancerateofmaincharactersonachapterbasis returnedtoavonlea.wecanspeculatethatthefactthatmarilaandannewereapartfrom eachotherina chapterhelpsfurtherdeepenthebondbetweenthetwointhefolowingchapters. 5.3Co-occurrenceofCharactersandVerbs Next,letusadvanceouranalysistothedetailoftheroleplayedbyMarila.Byfocusingonthecooccurrenceofthecharactersandverbs,weexplorewhatthemaincharactersincludingMariladointhe story.here,wefocusonfivecommunication-relatedverbs: think, know, tel, look,and feel among theverbslistedintable1.ofcourse say intable1isalsorelatedtocommunication,butitco-occurswith althemaincharacters,andthereforedoesnotmanifestthecharacteristicsofindividualcharacters.so,itis excludedfrom thescopeofthisanalysis 7. Figure15:Createaco-occurrencenetworkofcodes

5 A Two-StepApproachtoQuantitativeContentAnalysis(PartI)(HIGUCHIKoichi) 141 Figure16:Co-occurrencenetworkofmaincharactersandverbs Aco-occurrencenetworkiscreatedincludingthemaincharacterslistedinFigure14andthefiveverbs mentionedabove.figure15showstheprocedureofcreatingit.aftertheco-occurrencenetworkissavedby KH Coder,itistheneditedusingAdobeIlustratortomakeiteasiertoread.Here,charactersareindicated bygraycircles,andverbs,bywhitecircles.also,co-occurrencesofcharactersareindicatedbysolidlines, andothers,bydashedlines.theresultisshowninfigure16.thenumbersinthefigurearejaccard indices,whichrepresentthedegreeofco-occurrence.thelargerthevalue,thestrongerthedegreeofcooccurrence. Focusingonthelinksbetweencharacters,Figure16showsthatGilbertandDiana,whoarearoundthe sameageasanne,areindependentlyconnectedtoanne.incontrast,anne,marila,andmathew form a triad.thisisprobablybecauseanne,herfostermothermarila,andfosterfathermathew aredepictedas agroupofpeoplecloselyconnectedtoeachother,i.e.,afamily. Next,focusingontheverbsco-occurringwithMarila,weobserveinFigure16that feel stronglycooccursonlywithmarila.thismeansthattheverb feel ismorestronglyrelatedtomarilathantothe heroineanne.thisresultmaybeasurpriseformanyreaders,butactualy,itdoesnotnecessarilymean thatmarila sfeelingsaregivenmoreimportancethananne s.ofcoursetherearemanydescriptionsof Marila sfeelings,butthereasonwhy feel stronglyco-occurswithmarilaisalsobecause feel isoften containedinanne swordstogetherwith Marila,suchas Idofeeldreadfulysad,Marila (Chapter21). Inthissentence, feel doesnotco-occurwithanne,butonlyco-occurswithmarila.thesceneswhere

6 142 RitsumeikanSocialSciencesReview(Vol53.No.1) AnneexpressesherfeelingstoMarilaareoftendescribedinadditiontoMarila sownfeelings. Figure16alsoshowsthattheverb look stronglyco-occursonlywithmarilaandanne.then, searchingthesentencescontaining Marila, Anne,and look,wefoundthattherearemanydescriptions ofhowmarilaandannelookateachother.forexample,oneofthereasonsmariladecidedtoadoptanne wasbecausemarilalookedat,andwasmovedby,anne sfacialexpression. MarilalookedatAnneandsoftenedatsightofthechild spalefacewithitslookofmutemisery themisery ofahelplesslitlecreaturewhofindsitselfoncemorecaughtinthetrapfrom whichithadescaped. (Chapter6) Afterward,inthelaterhalfofthestory,thereisascenewhereMarilagentlylooksatAnne. Marilalookedatherwithatendernessthatwouldneverhavebeensuferedtorevealitselfinanyclearerlight thanthatsoftminglingoffireshineandshadow.(chapter30) Also,AnnegivesagentlelooktoMarilasuferingfrom headache. Annelookedatherwitheyeslimpidwithsympathy.(Chapter20) AnneexpressesherfeelingofgratitudetoMarilanotonlybywordsbutalsowithhereyes,suchas looked upearnestlyintoherface (Chapter30). Thus,MarilaandAnneexchangetheirfeelingsnotonlybywords,butalsowiththeireyes,meaning thatacloseandintimaterelationshipisdepictedbetweenthetwo. 5.4ChangeofWordsCo-occurringwithMarila Intheprevioussections,wehavecomparedMarilawithothermaincharacters,butinthissection,we finalyfocusonmarilaherself.wemakealistofthewordsco-occurringwithmarilaforeachoffourparts ofthestory,sothatwewilbeabletoexplorehowmarilaisdescribedandhowthedescriptionchangesas thestoryprogresses. FolowingtheprocedureshowninFigure17,wecancreatealistofthewordsstronglyco-occurring withmarilainpart Tosearchthewordsco-occurringwithMarilainthefolowingpart 08-19, repeatprocedure(3)infigure17andthenclick[*08-19]insteadof[*01-07]inprocedure(4).dothesame foralthefourpartsandlistthetop10wordsforeachparttocreatetable2.intable2,thecels containingthewordstheauthorparticularlynotedarehatched. Table2showsthatthewordmostcharacteristicofMarilainpart is Mathew,meaningthat therearefrequentinteractionsbetweenmarilaandmathew.also,not Anne but child islistedforthe firstpart,suggestingthatanneisnotcaledbyherownname,butlikelytobetreatedasanameless child bymarilain thispart.additionaly,marilaisnotused to treating a child,and thereforean uncomfortable situationoccursasfolows: Marilarealydidnotknowhowtotalktothechild,andheruncomfortableignorancemadehercrispandcurt whenshedidnotmeantobe.(chapter04) Inthefirstpart,Marilaisnotnecessarilykindtothechild,butisratherapersonoffewwords. However,forthefolowingparts and 20-28, say and Anne arelistedasthewords characteristicofmarila(table2),showingthatthe child isupgradedto Anne andimplyingthatitis impossibletobringupachildwithout saying anything. Idon tthinkthereismuchfearofyourdyingofgriefaslongasyoucantalk,anne,saidmarila unsympatheticaly.(chapter17) SuchadescriptiondoesnotsuggestthatMarilaindulgesAnne,butobviouslyshowsthatMarila satitude hasappreciablychangedfrom uncomfortableignorance inthefirstpart.alongwiththat,inthepart 20-

7 28, feel islistedasacharacteristicword,meaningthatthesceneswhereanneexpressesherfeelingsto MarilacometobedescribedaswelasMarila sownfeelings(section5.3). Inthelastpart 29-36,Mathew appearsagain(table2).thisisprobablybecausemathew passes awayalmostattheendofthestory.aftermathew sdeath,marilaandannetalkaboutmathew.here, look appearsforthefirsttimeasacharacteristicword.asdescribedinsection5.3,marila sandanne s eyesoneachotherisdepictedbytheword look.forexample,inthescenewhereannesaysthatshehas decidedtoquitleavinghomeandstaywithmarila,whoisanxiousaboutherhealth,thereisadescription A Two-StepApproachtoQuantitativeContentAnalysis(PartI)(HIGUCHIKoichi) 143 Table2:Changeofwordsco-occurringwithMarila Mathew.042 say.072 say.053 Mathew.040 look.034 think.059 ANNE.040 mare.039 sit.032 ANNE.039 just.040 Cuthbert.038 ANNE.030 cake.036 think.038 table.038 say.028 make.031 brooch.037 dish.031 face.028 minister.030 tel.033 child.026 girl.026 Alan.025 evening.032 bed.024 think.025 feel.024 home.032 say.022 want.024 know.024 set.032 uncomfortable.022 lean.023 time.023 let.032 sorrel *ThevaluesareJaccardindices,whichrepresentthedegreeofco-occurrence. Figure17:Searchrelatedwords

8 144 RitsumeikanSocialSciencesReview(Vol53.No.1) asfolows: NotgoingtoRedmond! Marilaliftedherwornfacefrom herhandsandlookedatanne. Why,whatdoyou mean? (Chapter38) Theirstrongemotionsareexpressedwiththeireyesasaproverbsays Theeyeisthewindowofthemind. Thus,Table2showsthatMarila,whooncewasnotusedtotreatingchildrenand crispandcurt, changesheratitudeasthestoryprogresses.marilagradualychangesherselftomakeacloseandintimate relationship,suchasexchangingfeelingsandemotionaleyecontacts,withanne.marilanotonlyappears frequently,butalsoplaysanessentialroleofgradualymakingadeepandrichrelationshipwithannein thisstory. TheabovequantitativeanalysissupportstheassertionmadebyDoody(1997)thattheeducationof Marilaisthecentralthemeofthestory(Section2.1).Themostimportantkeywordsamongthose suggestedbythequantitativeanalysisinclude child and uncomfortable inearlyparts,and feel and look inlaterparts.identifyingsuchkeywordsthroughquantitativeanalysisisconsideredtobeusefulfor extractingdepictionwhichspecificalydescribesmarila schangeoreducation. 6FeaturesofTwo-StepApproach 6.1AdvantagesofQuantitativeAnalysis Theprimaryadvantageofperformingquantitativeanalysisasintroducedinthisarticleisthatitalows forexploringdata,inotherwords,contributestobeterunderstandingofthedata. Oneaspectofdataexplorationbyquantitativeanalysisisthatwecanobtainoverviewsofentiredata.By automaticalycountingwordsusingacomputer,youmaynotice,forexample,that Marilaappearsmore frequentlythanithought.overviewsofdatacanbeusefulbyitselfandalow ustodiscoverfeaturesof datathathavenotbeenpreviouslyobserved.also,overviewsareusefulwhenatemptingtoanalyzethe meaningofeachwordorphraseindetail,sincetryingtoexaminethemeaningofonesentenceindetail makesitdificulttosimultaneouslyviewthedataasawhole.thisissueiscompensatedbyvisualizingentire dataintoaform ofagraphormapusingquantitativemethods.forexample,ifyouobtainanoverview of thestoryflowbyperformingcorrespondenceanalysis(figure11),itwilbeeasiertoexamineacharacter s roleinaspecificscene. Inadditiontobeingabletoobtainanoverviewofthedata,anotherimportantaspectofdataexploration isthatthedatacanalertresearcherstoimportantsentencesthatshouldbereadindetail.forexample,if youfindthatsentencescontaining Marila oftencontain feel too,youmaynoticesomethingnew about thedatabysearchingforsentencescontainingboth Marila and feel.also,iftheappearanceofaspecific charactersuddenlyincreasesordecreasesamongbubbleplotssuchasshowninfigure14,theremaybe somemovementinthestory.thus,quantitativeanalysissuggestswhichpartofthedataisconsideredtobe importantandwhichpartofthedataistobeinterpretedindetailbyresearchers. Thesecondaryadvantageofquantitativeanalysisisthatitimprovesthereliabilityofanalysis.Actualy, thispointisinextricablyassociatedwiththeadvantagethatitalowsforexploringthedataasdescribed above.thisisbecauseifanoverviewofthedataispresentedbyquantitativeanalysis,italowsthirdparties tocheckwhethertheuserenvisionsabiased,selectivewholeimageconvenientforhis/herhypothesesor theories.also,youwilbeabletoaddresssuchdoubtsas How didyouchoosethesentencetoquoteor interpretfrom thedata? or Didn tyouquoteonlythepartsconvenientforyou? tosomeextent.for example,youwilbeabletoexplainthatyoufocusedpartswhereasuddenchangeoccursinthegraphor

9 A Two-StepApproachtoQuantitativeContentAnalysis(PartI)(HIGUCHIKoichi) 145 yousearchedforsentencescontainingtheword feel thatwascharacteristictoacharacter.asdescribed above,youcanclearlydemonstratehow theconclusionwasderivedfrom thedata,thusimprovingthe reliabilityoftheanalysis.thiswilhelptoaccumulateresearchthatcanwithstandcomparisonand verification. 6.2AcademicBackgroundofTwo-StepApproach Theauthorproposedatwo-stepapproachintroducedinthisarticleintendingtomaketheadvantagesof thequantitativeanalysisdescribedaboveeasiertobeusedbymoreresearchers(higuchi2004,2014).this approachisbasedontheideaofcontentanalysis.inthissection,theacademicbackgroundandphilosophy ofthisapproacharedescribedastheclosingremarkofthisarticle. Computershavebeenactivelyusedforcontentanalysissincethe1960s.Atthattime,thereweretwo conflictingideasregardinganalyticalapproaches(stone1999).thefirstapproachfoundgroupsofwords thatoftenappeartogetherinthesamesentencethroughstatisticalanalysis,andisnowcaled Correlational approach 8.Theotherisananalysismethodforextractingconceptsfrom databyusingcodingrules,andis now caled Dictionary-basedapproach.TheadvocatesoftheCorrelationalapproacharguedthatthe advantageoftheirapproachwasthattheanalyticalresultsarenot contaminated bytheresearcher s theories,hypotheses,orprejudices(iker& Harway1969).Meanwhile,theadvocatesoftheDictionary-based approachthoughtthatcodingwasessentialtoachievethepurposeofanalysis(osgoodetal.1957). Atfirstglance,thetwo-stepapproachintroducedinthisarticlemayseem tobeasimplejoiningof thesetwoapproaches.however,instep1,therearesomediferencesfrom theconventionalcorrelational approach.thatis,whentheconventionalcorrelationalapproachwasusedinactualresearch,wordstobe analyzedwereoftenhand-selected,andwordswithsimilarmeanings,suchas say and talk,were designatedasingularunit.ifyouperform suchmanualdesignationsmanytimes,contrarytotheabove argumentbyiker& Harway(1969),thepreconceptionsoftheresearchermaybeimplicitlyintroduced. Therefore,wedecidednottodosuchmanualworkatthefirststageofthetwo-stepapproachintroducedin thisarticle.bydoingso,youaresparedtheefortofperformingmanualdesignationswhileimprovingthe reliabilityofanalysisresults. Also,withtheconventionalDictionary-basedapproach,composingcodingruleswasneveraneasytask. Composingcodingrulesusedtorequiretimeandefortconsumingtaskssuchasexaminingtheentireset ofdata(saporta& Sebeok1959).Also,itwasdificultforthirdpartiestojudgetheappropriatenessofthe contentofthecodingrules.however,withthetwo-stepapproachintroducedinthisarticle,youcan composecodingrulesreferringtotheoverview ofthedatathatisrevealedinstep1.also,youonlyhave towritecodingrulesfortheconceptsthatneedtobeadditionalyextracted.therefore,itismucheasierto composecodingrulesthanwiththeconventionaldictionary-basedapproach.furthermore,bycomparing theresultofstep1analysiswiththecodingrules,anythirdpartycanconfirm onwhichpartofthedatathe researcherhasfocused.thereliabilityofanalysisisimprovedalsointhisrespect. Thus,thetwo-stepapproachintroducedinthisarticleistojointheexistingtwoapproacheswhile makingmodifications.withthismodificationandjoining,thisapproachhasmadecontentanalysiseasier andmorereliable. Acknowledgments ThisworkwassupportedbyJSPSKAKENHIGrantNumber

10 146 RitsumeikanSocialSciencesReview(Vol53.No.1) Notes 7 Aslongasoperatedexplicitlyasabove,usersarerecommendedtoproactivelyfocusonsomewordsor conceptsinstep2inthismanner. 8 CorrelationalapproachisalsocaledasStatisticalAssociationapproach. References Danowski,J.A.,1993, NetworkAnalysisofMessageContent,W.D.RichardsJr.& G.A.Barneteds.,Progresin CommunicationSciencesIV,Norwood,NJ:Ablex, Doody,M.A.1997, Introduction,W.E.Barry,M.A.DoodyandM.E.D.Joneseds.TheAnnotatedAnneof GreenGables,OxfordUniversityPress,NewYork,9-34. Greenacre,M.J.,2007,CorrespondenceAnalysisinPractice2nded.,BocaRaton,FL:Chapman& Hal/CRC. Higuchi,K.,2004, QuantitativeAnalysisofTextualData:DiferentiationandCoordinationofTwoApproaches, SociologicalTheoryandMethods,19(1):101-15(WriteninJapanese). Higuchi,K.,2014,QuantitativeTextAnalysisforSocialResearchers:AContributiontoContentAnalysis,Nakanishiya Publishing:Kyoto,Japan(WriteninJapanese). Higuchi,K.,2016, ATwo-StepApproachtoQuantitativeContentAnalysis:KH CoderTutorialUsingAnneof GreenGables(PartI),RitsumeikanSocialScienceReview,52(3): Iker,H.P.& N.I.Harway,1969, ComputerSystemsApproachtowardtheRecognitionandAnalysisofContent, G.A.Gerbner,O.R.Holsti,K.Krippendorf,W.J.Paisly& P.J.Stoneeds.,TheAnalysisofCommunication Content:DevelopmentsinScientificTheoriesandComputerTechniques,NewYork:Wiley& Sons, Kawabata,Y.,2008, SurpriseofMarilaCuthbert Katsura,Y.andShirai,S.eds.TheworldofMasterpiecesWe WanttoKnowMore10:AnneofGreenGables,Minerva:Kyoto,Japan,109-19(WriteninJapanese). Matsumoto,Y.,2008,JourneytotheAnneofGreenGables:HiddenLoveandMystery,NHK Publishing:Tokyo,Japan (WriteninJapanese). Osgood,C.E.,1959, TheRepresentationalModelandRelevantResearchMethods, I.d.S.Pooled.,Trendsin ContentAnalysis,Urbana,IL:UniversityofIlinoisPress, Osgood,C.E.,G.J.Suci& P.H.Tennenbaum,1957,TheMeasurementofMeaning,Urbana,IL:Universityof IlinoisPress. Saporta,S.& T.A.Sebeok,1959, LinguisticandContentAnalysis, I.d.S.Pooled.,TrendsinContentAnalysis, Urbana,IL:UniversityofIlinoisPress, Stone,P.J.,1997, ThematicTextAnalysis:New AgendasforAnalyzingTextContent, C.W.Robertsed.,Text AnalysisfortheSocialSciences,Mahwah,NJ:LawrenceErlbaum, Yamamoto,S.2008,From AnneShirleytoJaneEyre:IntroducingEnglishLiteratureinUniversityClases,University oftokyopress:tokyojapan(writeninjapanese).

11 A Two-StepApproachtoQuantitativeContentAnalysis(PartI)(HIGUCHIKoichi) 147 接合アプローチによる量的内容分析の実践 ( 二 ) 赤毛のアン を用いた KH Coder チュートリアル 樋口耕一 ⅰ 本稿では, 量的な内容分析を実践するための方法として筆者が提案している 計量テキスト分析 を, 新たな分析事例とともに紹介する 計量テキスト分析において, データを分析する具体的な手順にはいくつかのバリエーションがあるが (Higuchi2014), 本稿では特に 接合アプローチ と呼ばれる手順をとりあげる 第一に, このアプローチと, その実現のために筆者が開発 公開しているフリーソフトウェア KH Coder について概要を手短に紹介する 第二に, このアプローチにもとづいて小説 赤毛のアン を分析する手順を, 読者が自分の PC で同じ分析を行えるチュートリアルの形で記述する 第三に, 分析の結果を踏まえて, 本アプローチの特徴について議論する 本稿で紹介する接合アプローチとは, 従来の内容分析で利用されてきた2つのアプローチを接合したものである 従来の内容分析では, テキスト型データを計量的に分析するために Correlational アプローチか Dictionary-based アプローチを用いることが多かった Correlational アプローチはクラスター分析のような統計手法を用い, 頻繁に同じ文書の中にあらわれる言葉のグループを見つけだすといった方法で, データ中の主題を探索するアプローチである このアプローチは StatisticalAssociation アプローチと呼ばれることもある それに対して Dictionary-based アプローチでは, 統計手法ではなく, 分析者自身の指定した基準にそって言葉や文書を分類し, 計量的な分析を行なう これら2つは考え方が大きく異なるアプローチでありながら, 実際の分析においては混同されやすい部分もあった そこで混同されやすい部分を峻別した上で, これら2 つを接合したものが, 本稿で紹介する接合アプローチである 本稿のチュートリアルでは, この接合アプローチを用いて, 小説 赤毛のアン の原文を分析する 小説 赤毛のアン では, 主人公である孤児のアンが, マシューとマリラの兄弟に引き取られ, 成長していく様子が描かれている この物語においては養母マリラの果たした役割が非常に大きいという指摘がある 親友のダイアナや, アンとの淡いロマンスが描かれるギルバートよりも, マリラの方が中心的であったという また 赤毛のアン は, マリラが子供を愛することを学び, それによって自分自身も幸せになっていくという, 大人の成熟と生き直しの物語であると指摘されている 本稿の分析では, こうしたマリラの重要性を, 計量的分析からも読み取ることができるのかどうかを確認する なお本稿の後半をここに掲載する 前半については本誌の52 巻 3 号に掲載している キーワード : 量的内容分析,KH Coder, 赤毛のアン, チュートリアル, 計量テキスト分析 ⅰ 立命館大学産業社会学部准教授