The case study as a type of qualitative research

28 JOURNALOFCONTEMPORARYEDUCATIONALSTUDIES1/2013 A.B.Starman

Adrijana Biba Starman

Thecasestudyasatypeofqualitativeresearch

Abstract: Thisarticlepresentsthecasestudyasatypeofqualitativeresearch.Itsaimistogiveadetaileddescriptionofacasestudy–itsdefinition,someclassifications,andseveraladvantagesanddisadvantages–inordertoprovideabetterunderstandingofthiswidelyusedtypeofqualitativeapproach.Incomparisontoothertypesofqualitativeresearch,casestudieshavebeenlittle understoodbothfromamethodologicalpointofview,wheredisagreementsexistaboutwhethercasestudiesshouldbeconsideredaresearchmethodoraresearchtype,andfromacontentpointofview,wherethereareambiguitiesregardingwhatshouldbeconsideredacaseorresearchsubject.Agreatemphasisisplacedonthedisadvantagesofcasestudies,wherewetrytorefutesomeofthecriticismsconcerningcasestudies,particularlyincomparisontoquantitativeresearchapproaches.

Keywords: casestudy,qualitativeresearch,qualitativemethods

UDC:37.012

Scientificarticle

Adrijana Biba Starman, Master of library science, Bergantova 13, SI-1215 Medvode, Slovenia;e-mail for correspondence: [email protected]

JOURNALOFCONTEMPORARYEDUCATIONALSTUDIES1/2013,28–43

Thecasestudyasatypeofqualitativeresearch 29

Introduction

Casestudieswereoneofthefirsttypesofresearchtobeusedinthefieldofqualitativemethodology.1Today,theyaccountforalargeproportionoftheresearchpresentedinbooksandarticlesinpsychology,history,education,andmedicine,tolistjustafewofthefundamentalsciences.Muchofwhatweknowtodayabouttheempiricalworldhasbeenproducedbycasestudyresearch,andmanyofthemosttreasuredclassicsineachdisciplinearecasestudies2(Flyvbjerg2011,p.302).

Casestudieshavebeenlargelyusedinthesocialsciencesandhavebeenfoundtobeespeciallyvaluableinpractice-orientedfields(suchaseducation,management,publicadministration,andsocialwork).Butdespitethislonghis-toryandwidespreaduse,casestudyresearchhasreceivedlittleattentionamongthevariousmethodologiesinsocialscienceresearch.AccordingtotheauthorsoftheEncyclopedia of Case Study Research(Millsetal.2010),onlyafewtextsdealdirectlywithcasestudiesasacentralsubjectandnoencyclopaedicreferenceprovidesathoroughoverviewofthedesignandmethodsincasestudyresearchasaguidanceforstudents,researchers,andprofessionalswhoaretryingtoin-corporatecasestudiesintoarigorousresearchprojectorprogram(ibid.,p.xxxi).D.A.deVaus(inThomas2011,p.511)stated,“Mostresearchmethodstextseitherignorecasestudiesorconfuseitwithothertypesofsocialresearch.”Fromthis,wecanconcludethatinspiteoftheirwidespreaduseandpopularity,casestudiesarecharacterizedbyambiguitiesandinconsistenciesinunderstandingtheirdefinition,subjectsofinvestigation,andmethodologicalchoice(Verschuren2003,p.121).Casestudiesarethereforemisunderstoodasatype,aswellasamethod,ofqualitativeresearch(Gerring2004,p.341).

1Casestudies,inthefieldofpsychology,forexample,datebacktothemiddleofthe19thcentury.Insocialwork,theyhavebeeninusesince1920,referredtoascaseworks(Millsetal.,2010,p.109);basedonthegroundbreakingworkofS.B.MerriaminCase Study Research in Education(Merriam1988),therehasbeensignificantprogressinthefieldofqualitativeresearchingeneral,andthusadvanceshavealsobeenmadeinthestandardizationofcasestudiesinthefieldofeducation.

2ThemostfamouscasestudiesinpsychologyarethoseofPiaget,Freud,Money,andotherfamouspsychologists(Casestudyinpsychology…n.d.).TheuseofcasestudiesinthefieldofeducationisdescribedintheJournal of Case Studies in Education.


Are case studies a qualitative research type or a qualitative research method?

Beforedelvingfurther intothis investigation, it is importanttomakeadistinctioninhowcasestudiesareviewed;someauthorsseethemasaqualita-tiveresearchtype(BaxterandJack2008;Flyvbjerg2006,2011;Sagadin2004;Simons2009;Stake2005;Sturman1997;Verschuren2003),whileothersperceivethemtobeaqualitativeresearchmethod(GeorgeandBennett2005;Gerring2004).Inthisarticle,wewilldemonstratethatcasestudiesaremorethanjustamethodologicalchoice;therefore,wechoosetodefinecasestudiesasaqualitativeresearchtype.

Althoughcasestudieshaveoftenbeenconsideredtobepartofqualitativeresearchandmethodology,theymayalsobequantitativeorcontainacombinationofqualitativeandquantitativeapproaches.Qualitativeresearchischaracterizedbyaninterpretativeparadigm,whichemphasizessubjectiveexperiencesandthemeaningstheyhaveforanindividual.Therefore,thesubjectiveviewsofaresearcheronaparticularsituationplayavitalpartinthestudyresults.Anothercharacteristicofqualitativeresearchisitsidiographicapproach3(Vogrinc2008,p.14),whichemphasizesanindividual’sperspectiveontheinvestigativesituation,process,relations,etc.(ibid.,p.19).Theinterpretativeparadigm,phenomenologicalapproach,andconstructivism4asaparadigmaticbasisofqualitativeresearcharecloselylinkedtothedefinitionandcharacteristicsofcasestudies.Acasestudyisthereforemorequalitativethanquantitativeinnature,butnotexclusively,foritcanbequalitative,quantitative,oracombinationofbothapproaches(withbothrepresentedequallyoroneapproachprevailingandtheothersupplementing).Qualitativeandquantitativeresultsshouldcomplementeachothertocreateameaningfulwholeaccordingtotheobjectandpurposeoftheinvestigation(Sa-gadin2004,p.89).

Weshouldalsoclarifysomeotherterms,suchas“comparativemethods,”“casestudymethods,”and“qualitativemethods.”Comparativemethods(comparingasmallamountofcasesandexploringfacts,relations,orprocessesinordertofinddifferencesorsimilarities)differfromcasestudiesinthatacasestudycoversinvestigationwithinindividualcases,whilethecomparativemethoddoesnot.Qualitativemethodsarecloselylinkedtocasestudies.Acasestudyisconsideredbysomeresearcherstobeapartofqualitativeresearch–atypeand,sometimes,amethodorscientificapproach.Inthisarticle,casestudiesareplacedwithinthequalitativefieldandviewedasaqualitativeresearchtype,althoughthefactthattheycancontainsomequantitativeelements,especiallyregardingresearchquestionsandgoals,isalsotakenintoaccount.

3Theexaminationofindividualcases.4Theinterpretiveparadigm,thephenomenologicalapproach,andconstructivismareparticularly

interestedinindividualexperiencesofreality.Objectiverealityandtruth,accordingtoconstructivistsandphenomenologists,doesnotexist,butisratheralwaysaconstruct.Itisthereforetheideathatpeoplehaveofrealitythatisimportantforresearchers,nottherealityitself(Vogrinc2008,p.27).


Inthispaper,wewillfirstprovidevariousdefinitionsofcasestudies,rangingfromverygeneraltomorespecific,andwilldescribethemainadvantagesanddifferentclassificationsofcasestudies.Lateron,wewillfocusonadetailedde-scriptionofcasestudies’disadvantagesandcriticismsinordertoachieveabetterunderstandingofthistypeofqualitativeresearchandtocreateaclearerpictureofwhatcasestudyis,whenitisapplicabletoresearch,andwhataresearchershouldpayattentiontowhenconductingasurveyusingacasestudy.

Definitions and classifications of a case study

Gerring(2004)notesthattheeffortsofmanyauthorstoclarifytheconceptofacasestudyhaveoftenleadtoadefinitionaljumblebecauseeverytimesomeonetriestoclarifytheconfusionusingdefinitions,itonlymakesitmoreconfusing(ibid.,p.342).Flyvbjerg(2011)thereforebelievesthatifadefinitionofacasestudyisneeded,itisbetterthatitismoregeneralanddoesnotcontainaplethoraofmeticulousdescriptions(ibid.,p.302).However,wecannotsaythatthedefinitionofacasestudyisunnecessarybecauseitisthedefinitionthatplacesthecasestudywithinitsownspaceandgivesititsowncharacteristicsincomparisontoothertypesofqualitativeresearch.Severalresearchershaveprovidedgeneraldefinitionsofcasestudies.

AccordingtoSturman(1997),“[a]casestudyisageneraltermfortheexplo-rationofanindividual,grouporphenomenon”(ibid.,p.61).Therefore,acasestudyisacomprehensivedescriptionofanindividualcaseanditsanalysis;i.e.,thecharacterizationofthecaseandtheevents,aswellasadescriptionofthediscoveryprocessofthesefeaturesthatistheprocessofresearchitself(Mesec1998,p.45).Mesecoffersadefinitionofacasestudywithinthefieldofsocialwork,butitcouldalsobeappliedtothefieldofeducation:Acasestudy“isadescrip-tionandanalysisofanindividualmatterorcase[…]withthepurposetoidentifyvariables,structures,formsandordersofinteractionbetweentheparticipantsinthesituation(theoreticalpurpose),or,inordertoassesstheperformanceofworkorprogressindevelopment(practicalpurpose)”(ibid.,p.383).Headdsthatonecasestudycouldservebothpurposesatthesametime(ibid.).

Formoredetaileddefinitionsofacasestudy,Sagadin(1991)statesthata“casestudyisusedwhenweanalyseanddescribe,forexampleeachpersonindividually(hisorheractivity,specialneeds,lifesituation,lifehistory,etc.),agroupofpeople(aschooldepartment,agroupofstudentswithspecialneeds,teachingstaff,etc.),individualinstitutionsoraproblem(orseveralproblems),process,phenomenonoreventinaparticularinstitution,etc.indetail.Ifweremaininsuchanalysesonthedescriptivelevel,thenacasestudyisconsideredasaformofdescriptivemethod,butifweclimbtothecausallevel,casestudyproceedstowardscausal-experimentalmethod”(ibid.,p.31).Further,casestudieshighlightadevelopmentalfactor,whichmeansthatthecasesaregeneratedandevolveovertime,oftenasaseriesofspecificandinterrelatedeventsthatoccurin“thatparticulartimeandthatparticularplace.”Holisticallyspeaking,thisconstitutes the case.Finally,


casestudiesfocusontheenvironment;i.e.,thecontext.Outliningthebordersofindividualunitswithinthesurveyestablisheswhatcountsasacaseandwhatbecomesitscontext5(ibid.,p.301).

Simons(2009)createdthefollowingdefinitionofacasestudybasedonacriticalreviewthatsoughtcommonalitiesofvariouscasestudydefinitions:“Casestudyisanin-depthexplorationfrommultipleperspectivesofthecomplexityanduniquenessofaparticularproject,policy, institution,programorsystemina‘reallife’”(ibid.,p.21).Shealsoemphasizedthatacasestudyshouldnotbeseenasamethodinandofitself.Rather,itisadesignframethatmayincorporateanumberofmethods.Stakeagrees;hestatedthatacasestudyisnotamethodo-logicalchoice,butratherachoiceofwhatistobestudied–bywhatevermethodswechoosetostudythe case.Insodoing,wecanstudyitanalytically,holistically,hermeneutically,culturally,andbymixedmethods,butweconcentrate,atleastforthetimebeing,onthecase(Stake2005,p.443).Choiceofmethod,then,doesnotdefineacasestudy.Itistheanalyticaleclecticismthatisessential6(Thomas2011,p.512).Flyvbjerg(2011)sharesasimilaropinion,sayingthatifwedecidetouseacasestudyinourresearch,thisdoesnotmeantheselectionofamethod,butratheraselectionofwhatwillbeexplored(ibid.,p.301).Anindividualcasecanbestudiedfromdifferentperspectives–bothqualitativeandquantitative.7

Ifweanalysethesedefinitionsintermsofdifferencesandsimilarities,wewouldreachaconclusioninagreementwithSimons(2009);thatis,theyallsharecommitmenttotheexaminationofcomplexityinavarietyofreal-lifesituationsandtheydonotincludevariousmethodsofdataacquisition,fortheirfocusiselsewhere.Asforthedifferencesamongthedefinitions,wecanattributethemtothedifferentepistemologicalbasesthatvariousresearchersleanupon(Thomas2011,p.512),suchasthepurpose(Mesec1998),level(Sagadin2004),timeframe(Flyvbjerg2011),orcontext(ibid.;Thomas2011)oftheresearch.

Whendiscussingadefinitionofacasestudy,manyresearchersfocusontheindividualcase(ormultiplecases)athandratherthanoncasestudiesasatypeofresearch.AccordingtoVerschuren(2001,p.137),thisisexactlythereasonforthemanydefinitionsofcasestudiesfromamethodologicalpointofview,since,inhisopinion,disagreementsappearamongdefinitionswithatendencytoclassifycasestudiesasastudyofone(ormore)casesinsteadofviewingcasestudiesasaresearchapproach.

Various classifications: Case study types and categories

Casestudydefinitionsareusuallyrathergeneralanddonotcontainvariousclassificationsortypesofcasestudies,asitisalmostimpossibletocoveralltypesofstudiesinonedefinition.Furthermore,authorsofferavarietyofcomplementaryordistinguishingclassificationsaccordingtoclassificationtype.Sagadin(2004)hasalreadymadeatransparentandcomprehensivecontributionregardingthe

5Alsocalledthe object,moreincontinuation.6Forexample,takingoverandmergingdifferentsystems,views,findings.7Seep.2.


variousclassificationsofcasestudies8thatwillbesupplementedinthispaperwithsomeadditionalclassifications.

Classificationaccordingtothetimedimension

Inacasestudy,oneormorecasescanbeinvestigated.Whenexaminingonecase,werefertoasingularcasestudy,andamultipleorpluralcasestudyisusedtodescribeastudyexaminingseveralcases.Inmultiplecasestudies,eachcaseisstudiedasifitisasingularstudyandisthencomparedtoothercases.Theanalysisofeachfollowingcaseisbuiltontheknowledgeobtainedintheanalysisofpreviouscases(Mesec1998,p.384).Forsingularandmultiplecasestudies,Thomassuggestsanadditionalclassification,accordingtothetypeoftimedimen-sion.Thetypesofsingularcasestudies,regardingtimedimension,areasfollows(Thomas2011,p.517):

– Retrospectivecasestudies:Thesimplesttypeofstudy;itinvolvesthecollec-tionofdatarelatingtoapastphenomenonofanykind.Theresearcherislookingbackonaphenomenon,situation,person,oreventandstudyingitinitshistoricalintegrity.

– Snapshotstudies:Thecaseisbeingexaminedinoneparticularperiodoftime,suchasacurrentevent,adayinthelifeofaperson,adiary,etc.Whetheramonth,aweek,aday,orevenaperiodasshortasanhour,theanalysisisaidedbythetemporaljuxtapositionofevents.Asthesnapshotdevelops,thepicturepresentsitselfasaGestaltoveratighttimeframe.

– Diachronic studies:Changeover timeandaresimilar to longitudinalstudies.

Examplesofmultiplecasestudiesareasfollows(ibid.):

– Nestedstudies:Involvethecomparisonofelementswithinonecase(nested elements).Withnestedstudies,thebreakdowniswithin theprincipalunitofanalysis.Anestedstudyisdistinctfromastraightforwardmultiplestudyinthatitgainsitsintegrity–itswholeness–fromthewidercase.Forexample,aresearchermightobservethreewardswithinonehospital.Theonlysig-nificanceaboutthemistheirphysicalhousingatthehospital.Suchacasewouldnotbeconsideredtobenested,astheelementsarenestedonlyinasensethattheyformanintegralpartofabroaderpicture.Inthiscase,thatmeansthewardsareobservedinordertoprovideabroaderpictureof,forexample,howtheyaffectthepatients’well-being,whatthehospital’sagendaislike,andtherelationshipsandattitudesamongthewards,patients,staff,etc.

8Fortheconceptualandmethodologicalcriteria,seeBogdanandBiklen(1982)andStenhouse(1985);forthepurpose,seeStake(1994);forthethesegmentation/integrity,seeCreswell(inSagadin2004).


– Parallelstudies:Thecasesareallhappeningandbeingstudiedconcur-rently.

– Sequentialstudies:Thecaseshappenconsecutively,andthereisanassump-tionthatwhathashappenedatonetimepointorinaninterveningperiodwillinsomewayaffectthenextincident.

Classificationaccordingtothetheoryformation

GeorgeandBennett(2005)presentsixtypesofcasestudiesclassifiedaccordingtowhethertheycontributetotheorybuilding(ibid.,pp.75–76):

– Atheoretical/configurative idiographic casestudies: Illustrativecasestudiesthatdonotaccumulateorcontributedirectlytotheory.

– Disciplinedconfigurative casestudies:Useestablishedtheoriestoexplainthecase.

– Heuristiccasestudies:Identifynew,unexpectedpaths; forsuchstudies,marginal,deviant,oroutliercasesmaybeparticularlyuseful.

– Theory-testingcasestudies:Studiesthatassessthevalidityandscopecondi-tionsofsingleorcompetingtheories.

– Plausibility probes:Preliminarystudiesusedtodeterminewhetherfurtherexaminationiswarranted.

– “Building Block”studies:Studiesofparticulartypesorsubtypesofapheno-menon,that,whenputtogether,contributetoamorecomprehensivetheory.

Inthisarticle,weonlylistafewclassificationsinordertopresentvarioustypesofcasestudiesaccordingtowhatwewanttoexplore,forwhatpurpose,whatwewanttoachieve,andhow.Someclassificationsaremutuallycomplementary.Themoreclassificationswearefamiliarwith,thebetterandeasierwecancatego-rizeourowncasestudy.However,atsomepoint,wehavetodrawaline,becausebeingtoofocusedondetailwhenclassifyingacasestudycouldconcealthegeneralimportanceofthestudy.Therefore,classificationscanbehelpfulwhenplacingourcasestudywithinacontext,withinwhichwewillexploreacertaintopic.

Keydifferenceswithinthesedefinitionsrefertothecriteriafromwhichthevariousclassificationsweredrawn.Whilesomeauthorsdividecasestudiesaccordingtothenumberofcasesthatwillbestudied,towhichdegreeanindividualcasewillbeanalysed,orwhetheritwillbecoveredholistically,othersrefertothecriteriaaccordingtothepurpose,timedimension,ortheory/nontheorybuilding.

Afterlookingatthevariousdefinitionsofcasestudies,wedrawtheconclu-sionsthatthedifferentclassificationsarederivedfromdifferenttheoreticalposi-tions.Choosingaparticulartypeorkindofacasestudydependsontheresearchpurpose.Forexample,dowewanttodescribeaparticularcaseandthusremainatthedescriptivelevelordowewanttoexploreitonthecausallevelaswell?Dowewanttocompareseveralcases?Whatcountsasa“case”inacasestudyandhowcanitbeproperlyselected?Dowewanttocheckanexistinghypothesisordowewanttodiscovernewones?Doweperhapsevenwanttodevelopanew


theory?Inthefollowingsection,wewillpresentamoredetaileddiscussionoftheseplatforms.

The case (subject), research field (object), and case selection

Tobeabletodebateacasestudy,ithastobedefinedwithinananalyticalframeworkorobjectintheconstitutionofthestudy(Thomas2011,p.512),orasGeorgeandBennettputit(2005,p.69),theinvestigatorshouldclearlyidentifytheresearchfield;thatis,the“class”or“subclass”ofeventswithinwhichasinglecaseorseveralcasesareinstancestobestudied.Thesubjectofthestudyisthusaninstanceofsomephenomenon,andthephenomenoncomprisestheanalyticalframe;thatis,theobject(ibid.).

Casestudiesasaresearchtypemightappearabitvague.Theirloosenessandemphasisonthecase(subject)maybewhyresearchers,students,etc.(especiallythosewhoareinexperienced),neglecttheimportanceofdefininganobjectintheirexploration.Identifyingonlyasubjectleadstoashortageofabroaderdescriptionandinterpretationandinsteadonlyoffersasimplifieddescriptionofaresearchpiece.Therefore,theobjectconsistsofananalyticalframeworkwithinwhichthecase(subject)isunderstoodandillustrated.Itisnotnecessary,however,fortheobjecttobedefinedatthebeginningofthestudy;thisoftenoccurslaterintheexplorationprocess(Thomas2011,p.515).

Acasestudyisaboutdeterminingwhattheinvestigatedcasemaybe; itisnotaboutdefiningpopulationsandselectingappropriatesamples(Sagadin1991,p.34).Acasestudyisusuallyastudyofasinglecaseorasmallnumberofcases.Theideaofrepresentativesamplingandstatisticalgeneralizationstoawiderpopulationshouldberejected,andanalyticalinductionshouldbechoseninstead.Someauthorsbelievethatthecaseinacasestudycountsasaresearchunit,whileothersdisagree.Theuseoftheterm“unit”cancauseconfusion.Someauthorsbelievethatitrelatestothecaseorresearchsubject(e.g.,WieviorkainThomas2011,p.513;Mesec1998),whileothersuseittodescribetheobjectwiththeunderstandingthattheunit(object)andthecaseinfluenceeachothermutu-ally(VanWynsbergheandKhaninThomas2011,p.513).Inthisarticle,thetermunitisassociatedwiththecase(subject).

Mesecsuggestsselectingsuchcaseforaresearchunit(anindividual,familyorothergroup,organization,orcommunity)whereapracticalproblemthatweareinterestedinexists.Wemayalsoexamineseveralindividualcasesthatareselectedinsuchawaythattheiranalysisprovidesuswiththemostdiverseinformationthatweareablecollect.Weshouldselect interestingcases(e.g.,contrasting,extreme,exceptionalcases)insteadoftypical,averagecases(Mesec1998,p.55).Thesubject(thecase)isnotselectedbaseduponarepresentativesample,butratherisselectedbecauseitisinteresting,unusual,striking,andmaycausechangesinthecharacteristicsandspecificitiesoftheobject(Thomas2011,p.514).SimilartoMesec,Thomasalsosuggestschoosinganatypicalcase,wherethesubjectandobjectinteractinadynamicrelationship.


Ontheotherhand,Yinrecommendsselectingarepresentativeortypicalcase(Yin2009,p.48)becauseinsodoing,wemayfindnewhypothesesanddeeperlayersthatprevioustheoryhasmissed.Eachcasehasitsadvantagesanddisadvantages,buttheselectionofcasesandshouldmostlydependupontheresearchproblem.

Caseselectionhasalsotargetedbysomecasestudycritics.Theircriticismmainlyfocusesonpossiblesubjectivecaseselection,theso-calledselectionbias(i.e.,theimpactofaresearcher’spriorknowledgeaboutthecaseandhispossiblefavouritismtowardcertainhypotheses)thatcanimpactthecaseselection(GeorgeandBennett2005,p.24).However,theselectionofacasebasedonpriorknowledgeleadstoabetterresearchplan.Casesselectedonthebasisofpriorknowledgearemostlikelycrucialforenablingthedevelopmentofastrongtheoreticalbasefortheresearch,whichmakestheprocedureoftheorytestingmorerigorous.Inaddition,thereareseveralmethodologicalprovisionstoprotectastudyfromtheinfluenceofresearcherbias,suchasdiligenceandconsistencyinthetrackingprocess(ibid.).Thisincludesanaccurateandcomprehensivedescriptionofthedatacollectionproceduresanddocumentationofeverypieceofinformationinordertoachievereliabilityofacasestudy(ibid.,p.10).

Case study advantages

Casestudiesaregenerallystrongpreciselywherequantitativestudiesareweaker(ibid.,p.19).GeorgeandBennetthaveidentifiedfouradvantagesofcasestudiesincomparisontoquantitativemethods9:Theirpotentialtoachievehighconceptualvalidity,strongproceduresforfosteringnewhypotheses,usefulnessforcloselyexaminingthehypothesizedroleofcausalmechanismsinthecontextofindividualcases,andtheircapacityforaddressingcausalcomplexity(ibid.).

Conceptual validity

Conceptualvalidityreferstotheidentificationandmeasurementoftheindicatorsthatbestpresentthetheoreticalconceptsthataresearcherwantstomeasure.Manyofthevariablesthatsocialscientistsareinterestedin,suchasdemocracy,power,etc.,aredifficulttomeasure,sotheresearcherhastocarryouta“contextualizedcomparison,”whichautomaticallysearchesforanalyticallyequivalentphenomenaeveniftheyareexpressedindifferenttermsandcontexts.Thisrequiresadetailedconsiderationofcontextualfactors,whichisextremelydifficulttodoinquantitativeresearchbutisverycommonincasestudies.Whereasquantitativeresearchrunstheriskof“conceptualstretching”bythrowingtogetherdissimilarcasestogetalargersample,casestudiesallowforconceptualrefine-mentswithahighervalidityleveloverfewernumberofcases(ibid.,p.19).

9Theauthorsdefinecasestudyasamethodofqualitativeresearch;thisiswhytheycompareittoquantitativemethods.


Deriving new hypotheses

Casestudiesareverysuitableforservingtheheuristicpurposeofinductivelyidentifyingadditionalvariablesandnewhypotheses.Quantitativestudieslackproceduresforinductivelygeneratingnewhypotheses.Moreover,casestudiescananalysequalitativelycomplexeventsandtakeintoaccountnumerousvari-ablespreciselybecausetheydonotrequiremanycasesoralimitednumberofvariables.Casestudyresearchersarenotlimitedtoreadilyquantifiedvariablesorpre-existing,well-defineddatasets(ibid.,p.45).

Quantitativeresearchcanbeusedtoidentifydeviantcasesthatmayleadtonewhypothesesbut,inandofthemselves,lackanyclearmeansofactuallyidentifyingnewhypotheses.Withoutadditionalexamination,suchasopen-endedinterviews,itisnotpossibletofindinductivemeansofidentifyingomittedvari-ables(ibid.,p.21).

Exploring causal mechanisms10

Casestudiesexaminetheoperationofcausalmechanismsinindividualcasesindetail.Withinasinglecase,theylookatalargenumberofinterveningvariablesandinductivelyobserveanyunexpectedaspectoftheoperationofaparticularcausalmechanismorhelpidentifywhatconditionsarepresentinacasethatactivatethecausalmechanism,whilequantitativestudiesintheircorrelationslacksuchcausality(ibid.,p.21).However,onemustkeepinmindthatitisnotentirelytruethatquantitativeresearchdoesnotincludeanycausality.Wearereferringtoquantitativeresearch’sinabilitytotakeintoaccountcontextualfactorsotherthanthosethatarecodifiedwithinthevariablesbeingmeasured;inthissituation,manyadditionalvariablesthatmightalsobecontextuallyimportantaremissed.

Modelling and assessing complex causal relations

Casestudiesareabletoaccommodatecomplexcausalrelations,suchasequifinality,11complexinteractioneffects,andpathdependency.12Thisadvantageisrelativeratherthanabsolute.Casestudiescanallowforequifinalitybypro-ducinggeneralizationsthatarenarrowerandmorecontingent.Notwithstandingthisadvantage(moreaboutgeneralizationincontinuation),otherswhopreferquantitativemethodsappreciatetheoriesthataremoregeneralevenifthismeansthattheyaremorevagueandmorepronetocounterexamples(ibid.,p.22).

10Causalmechanism:“YhappenedbecauseofA,inspiteofB,”whereasAmeansasetofparticipa-tivecausesandBmeansapotentiallyemptyspaceofoppositecauses(Acannotbeempty;otherwise,itwouldnotbeabletoexplainY).Forexample,thecardroveofftheroadduetoinappropriatespeedandsandontheroaddespitegoodroadvisibilityandthedriver’salertness(SalmoninGeorgeandBennett2005,p.145).

11Equifinalitymeansthatthesameendresultcanbeobtainedindifferentways(InstituteoftheSlovenianLanguage…n.d.)

12Historicalheritageessentiallydefinesthedevelopmentalpossibilitiesoffutureevolution(e.g.,ofeachnation)(Vehovar2005,p.309).


Theuseofcasestudieshassomeadditionaladvantagesaswell.Theconnect-ednesstoeverydaylifeandcasestudies’abundanceofindividualelementsanddetailsareimportantforresearchersfromtwoviewpoints.First,acasestudyisimportantfordevelopingdifferentviewsofreality,includingtheawarenessthathumanbehaviourcannotbeunderstoodmerelyasanactthatisdrivenbyaruleoratheory.Second,casestudiescancontributetotheprofessionaldevelopmentofaresearcher,ascasestudiescanprovideconcrete,context-dependentexperiencethatincreasestheirresearchskills(Flyvbjerg2006,p.223).

Paradox, misunderstandings, and criticism

Intheintroduction,wenotedthatcasestudiesarewidelyusedbutunder-represented.Basedonthesefindings,Gerringhasidentifiedaparadoxinwhichhecorrectlystatesthatacasestudyexistsinastrange,curiousmethodologicallimbo,which,hebelieves,isduetoalackofunderstandingofthismethod(Gerring2004,p.341).Flyvbjerghasthereforesoughttoresolvethisparadoxand,insodoing,toachieveawideracceptanceandapplicationofresearchusingcasestudies.Hehasidentifiedfivemisunderstandingsaboutcasestudiesthatunderminethecredibilityandapplicationofthisresearchtype.Thesemisunderstandingsreferprimarilytothetheory,reliability,andvalidity(Flyvbjerg2006;2011):

General,theoretical(context-independent)knowledgeismorevaluablethan1.concrete,practical(context-dependent)knowledge.Itisimpossibletogeneralizeonthebasisofanindividualcase;therefore,2.thecasestudycannotcontributetoscientificdevelopment.Thecasestudyismostusefulforgeneratinghypotheses(thatis,inthefirst3.stageofatotalresearchprocess),whereasothermethodsaremoresuitableforhypothesestestingandtheorybuilding.Casestudiescontainabiastowardverification;thatis,atendencytoconfirm4.theresearcher’spreconceivednotions.Itisoftendifficulttosummarizeanddevelopgeneralpropositionsandtheo-5.riesonthebasisofspecificcasestudies.

Wewillnowattempttoresolveandclarifythesemisunderstandings.

General, theoretical knowledge is more valuable than concrete, practical case knowledge.

Socialsciencesdonothavemuchtoofferexceptconcretecasesandcontext-dependentknowledgebecauseresearchershavenotsucceededinproducinggeneral,context-independenttheories.Casestudiesareespeciallywell-suitedtoproducingthisexacttypeofknowledge.Thefirstargumentcanthereforeberevisedasthisstatement(Flyvbjerg2006;2011):“Concretecaseknowledgeismorevaluableforsocialsciencesthanthevainsearchforpredictivetheoriesanduniversals.”


Generalization upon the basis of an individual case is not possible; therefore, case studies cannot contribute to scientific development.

Thisisatypicalassumptionaboutcasestudiesamongproponentsofthenaturalscienceidealwithinthesocialsciences,yetevenresearcherswithoutastrongassociationwiththisidealmaysharethisviewpoint.Giddens,forexample,statesthatthetraditionalsmall-scaleresearchcommunityofanthropologyfieldwork,inandofitself,isnotgeneralizingstudies,butcaneasilybecomesoifcarriedoutinsomenumberssothattheirtypicaljudgementscanjustifiablybemade(GiddensinFlyvberg2006,p.225).

Incasestudies,inferenceisbasedonanalyticalinduction(analyticgene-ralization)andnotonstatisticalinduction(enumeration).Instatisticalinduction,oneisnotinterestedinstructuralorfunctionalconnectivitycharacteristicswithinindividualunits,butonlytheirpresenceorabsenceandquantitativesignificance,frequency,differences,andcorrelations.However,inanalyticinduction,weareexaminingaparticularcase–therelationshipsamongindividualcharacteristics,processes,oreventsandhowtheyareconnectedtoeachother(Mesec1998,p.50).Mesecthereforearguesthatiftheconnectionexistseveninjustonesinglecase,itmaybetheoreticallyimportant(ibid.).

Holistics,13inparticular,believethatgeneralizationmaybepossibleevenonthebasisofasinglecasestudy.Diesing,forexample,statesthatscienceencountersregularity(i.e.,thesearchforgeneralprinciplesandrules)andcreativity(i.e.,lookingfornew,originalcases).Iftheprimaryfocusisonregularity,thecreativitywillappear,andifthefocusisoncreativity,thenprincipleseventuallyshowup.Casestudiesincludeboththeparticularandtheuniversalwithoutbeingmutuallyexclusiveandmovebetweentheparticularanduniversalingradedsteps(DiesinginSturman1997,p.63).

Stakeholdsasimilarposition;hestatesthataprocessofnaturalisticgenerali-zationarrivesfromthetacitknowledgeofhowthingsare,whytheyare,howpeoplefeelaboutthem,andhowthesethingsarelikelytobelateronorinotherplacesthispersonisfamiliarwith.Generalizationisthereforepossiblebyrecognizingthesimilaritiesoftheobjectsandissuesindifferentcontextsandbyunderstandingthechangesastheyhappen(Stake1980inSturman1997,p.69).However,forthiskindofgeneralizationtobepossible,itisessentialtoensurethatthesalientfeaturesofthecasearedocumentedsothatnewsituationscanbeilluminatedbyaverythoroughunderstandingofaknowncase(Sturman1997,p.63).

AcasestudyisidealforgeneralizingfindingsusingthetypeoftestthatKarlPopper(inFlyvbjerg2006,p.228;Flyvbjerg2011,p.305)called“falsification”;insocialscience,thistestformspartofcriticalreflexivity.Popperbelievesthateverytruescientifictheoryallowsrefutation(StanfordEncyclopediaofPhilosophy2009).Falsificationthereforestatesthatahypothesisisconsideredtobescientificwhenitsdefenderisabletodeterminetheconditionsunderwhichthehypoth-

13Proponentsoftheholisticapproachintheepistemologyofscienceemphasizethestudyofcom-plexityintermsofthewhole.Inholism,thewholeismoreimportantthanthesumofitsindividualparts.Holismistheoppositeofindividualism,buttheyoftenoccurinpairs–inmacroandmicroperspectivesofobservingsocialreality(Mali2006,p.131).


esiscouldberefuted.14Falsificationisoneofthemostrigorousteststowhichascientificpropositioncanbesubjected–ifjustoneobservationdoesnotfitwiththeproposition,itisconsideredtonotbevalidandmustthereforebeeitherrevisedorrejected(Flyvbjerg2011,p.305).Deviantcasesandthefalsificationstheyentailaremainsourcesoftheorydevelopmentbecausetheypointtothedevelopmentofnewconcepts,variables,andcausalmechanismsthatisnecessaryinordertoaccountforthedeviantcaseandothercaseslikeit(ibid.).

Flyvbjergcorrectsthesecondmisunderstandingasfollows:“Onecanoftengeneralizeonthebasisofasinglecase,andthecasestudymaybecentraltoscientificdevelopmentviageneralizationassupplementoralternativetoothermethods”(ibid.,p.305).

Case studies are useful for generating hypotheses (i.e., at the beginning of the research process), while other methods are more suitable for hypotheses testing and theory building.

Thesourceofthisargumentlaysinthepreviousmisunderstandingthatitisimpossibletogeneralizefromcasestudies.Generalizationisassociatedwithhypothesistestingandisthenextstepofcaseselection.ButGeorgeandBennettareconvincedthatcasestudyisespeciallywell-suitedfortheorydevelopmentbecauseittacklesthefollowingtasksintheresearchprocessevenbetterthanothermethods(foranexample,seeGeorgeandBennett2005,pp.6–9):

– Processtracingthatlinkscausesandoutcomes,– detailedexplorationofhypothesizedcausalmechanisms,– developmentandtestingofhistoricalexplanations,– understandingthesensitivityofconceptstocontext,and– formationofnewhypothesesandnewquestionstostudysparkedbydeviant

cases.

The case study contains a bias toward verification; that is, a tendency to confirm the researcher’s preconceived notions.

Inthesectionaboutcaseselection,wehavealreadydiscussedsomeoftheconcernsregardingaresearcher’sbias.Doubtsandprejudicetowardverificationinscientificinvestigationisgeneral,buttheallegeddeficiencyofthecasestudyandothertypesofqualitativeresearchisthattheyostensiblyallowmoreroomfortheresearcher’ssubjectiveandarbitraryjudgmentthanquantitativeinvestigation(Flyvbjerg2011,p.309;GeorgeandBennett2005;Mesec1998;Thomas2011).

Sturmanbelievesthatacasestudycanachieveitsownformofprecision(Sturman1997,p.65)or,asWilsoncallsit,a“disciplinedsubjectivity”(Wilsoninibid.).Theprincipleofverifiabilityinacasestudy(andinqualitativeresearchingeneral)isrealizedbydescribingtheentireresearchprocessindetail,especially

14InPopper’sfamousexampleof“allswansarewhite,”heproposedthatjustonenotionofasingleblackswan(deviation)wouldfalsifythispropositionandinthiswaywillhavegeneralsignificanceandwillstimulatefurtherinvestigationsandtheorybuilding.Thecasestudyiswell-suitedforidentifying“blackswans”becauseofitsin-depthapproach(Flyvbjerg2011,p.305).


theanalysisprocessinwhichconceptsareshapedandtheregularityandpatternsofbehaviour,interaction,andexperiencearedetermined(Mesec1998,p.45).Toachievethecredibilityinacasestudy,Sturmansuggeststhefollowingstrategies(Sturman1997,p.65):

– Proceduresfordatacollectionshouldbeexplained,– datacollectedshouldbedisplayedandreadyforreanalysis,– negativeinstancesshouldbereported,– biasesshouldbeacknowledged,– fieldworkanalysesneedtobedocumented,– therelationshipbetweenassertionandevidenceshouldbeclarified,– primaryevidenceshouldbedistinguishedfromsecondaryevidenceand

descriptionandinterpretationshouldalsobedistinguished,– diariesorlogsshouldbeusedtotrackwhatwasactuallydoneduringdif-

ferentstagesofthestudy,and– methodsshouldbedevisedtocheckthequalityofdata.

Ingeneral,itisknownthatmoresimilarresultsandconclusionsarepossiblewhenrepeatingacertainstudy,whichleadstoincreasedreliabilityofthestudy.Iftheexperimentisrepeatedseveraltimesandalwayshasthesameresults,thenitsreliabilityis100%.Thisisthesameformeasuring,testing,etc.Casestudiescannotberepeatedbecauseduringrepetition,thecaseisalreadydifferent.Sotheabovedefinitionofreliabilityissomewhatmitigatedwhenitcomestoacasestudy.Therefore,acasestudyismorereliable–asmuchasweareabletoshowthatwecouldcometothesameconclusions–ifweareabletorepeatthesurveyunderanunchangedstateofcircumstances.Thisrequiresaccurateanddetaileddescriptionofdataacquisitionproceduresaswellasdocumentingeverysinglepieceofinformation(Mesec1998,p.148).

Mesecpointsoutthatthefindingsandresultsofacasestudyshouldbethefirst,notthelast,chapterinaparticularresearcharea.Casestudiesshouldthenbefollowedbyothersubsequentcasestudiesinordertoreinforcetheaccuracyofthefirststudy'sfindings.Thisshouldbedonewithrevisionoftheobservationsandfindingsand,mostimportantly,byspreadingthenetworkofnewlydiscoveredconnectionsamongcases.Casestudiesarecertainlymorethanjustanintroduc-tiontoquantitativeresearch.Ifwedonotwanttocount,wedonothavetodosoinordertolearnsomething(ibid.,p.380).

Conclusion

Quiteafewauthorshavealteredtheirviewsaboutcasestudiesasatypeofqualitativeresearchtype(see,forexample,Campbell1975andEysenck1976inFlyvbjerg2006).Inconsiderationaboutchanginghisview,Eysenckwrotefollowing:“Sometimeswesimplyhavetokeepoureyesopenandlookcarefullyatindividualcases–notinthehopeofprovinganything,butratherinthehopeoflearning


something”(Eysenck1976inFlyvbjerg2006,p.224).Becauseofalackof“hard”theory(theorythatcontainsexplanationsandpredictions)insocialsciences,itisdifficulttoattainstrongandcommonlyvalidrules.Butthisdoesnotmeanthatsocialscienceresearchhasnocontributiontoscienceatall–quitetheopposite,infact.Thereisconstantprogressiontowardnewdiscoveriesandcognitions!Acasestudycanbehelpfulwhenweareeagertoanswerthequestionsof“how”and“why,”whenwecannotinfluencethebehaviourofthoseinvolvedinastudy,andwhenwewanttocovercontextualconditionsbecausewebelievetheyarerelevanttothephenomenonunderstudyorwhentheboundariesbetweenthephenomenonandcontextarenotclear(YininBaxter&Jack2008,p.545).

Butwemustalsorecognizethatacasestudyismorethanjustatypeofqualitativeresearch.Itisaticketthatallowsustoenteraresearchfieldinwhichwediscovertheunknownwithinwell-knownborderswhilecontinuallymonitoringourownperformance;scalability;andourown,aswellasgeneral,existingknow-ledge.Wehopethisarticlesupportsandfosterstheviewofcasestudiesasatypeofqualitativeresearch.

References

Baxter,P.andJack,S. (2008).QualitativeCaseStudyMethodology:StudyDesignandImplementationforNoviceResearchers.The Qualitative Report,13,Issue4,pp.544–559.

Case study in psychology(n.d.).Retrievedfromhttp://en.wikipedia.org/wiki/Case_study_in_psychology#Famous_case_studies_in_psychology(Accessedon5.4.2012).

Creswell,J.(1994).Research design: qualitative and quantitative approaches.ThousandOaks,California:SagePublications.

Flyvbjerg,B.(2006).Fivemisunderstandingsaboutcase-studyresearch.Qualitative Inquir y,12,Issue2,pp.219–245.

Flyvbjerg,B.(2011).Casestudy.In:N.K.DenzinandY.S.Lincoln(eds.).The Sage Hand-book of Qualitative Research,4thed.ThousandOaks,CA:Sage,pp.301–316.

George,A.L.andBennett,A.(2005).Case Studies and Theory Development in the Social Science.Cambridge:MITPress.

Gerring,J.(2004).Whatisacasestudyandwhatisitgoodfor?The American Political Science Review,98,Issue2,pp.341–354.

Inštitut za slovenski jezik Frana Ramovša ZRC SAZU, laboratorij za korpus slovenskega jezika (n.d.).Retrievedfromhttp://bos.zrc-sazu.si/c/neva.exe?n=a_si_s&e=C_I_D_SM_MA%2020894(Accessedon4.4.2012).

Mali,F.(2006).Epistemologija družbenih ved: razlaga in razumevanje.Ljubljana:Fakultetazadružbenevede.

Mesec,B.(1998).Uvod v kvalitativno raziskovanje v socialnem delu.Ljubljana:Visokašolazasocialnodelo.

Merriam,S.B.(1988).Case study research in education. SanFrancisco,London:Jossey-BassPublishers.


Mills,A.J.,Durepos,G.andWiebe,E.(2010).Encyclopedia of case study research.London:Sage.

Sagadin,J.(1991).Razprave iz pedagoške metodologije.Ljubljana:ZnanstveniinštitutFilozofskefakultete.

Sagadin,J.(2004).Tipiinvlogaštudijprimerovvpedagoškemraziskovanju.Sodobna pedagogika,55,Issue4,pp.88–99.

Simons,H.(2009).Case study research in practice.London:SAGE.Stake,R.E.(2005).Qualitativecasestudies.In:N.K.DenzinandY.S.Lincoln(eds.).

The SAGE handbook of qualitative research (3rded.).ThousandOaks,CA:Sage,pp.443–466.

Stanford Encyclopedia of philosophy: Karl Popper(2009).Retrievedfromhttp://plato.stan-ford.edu/entries/popper/(Accessedon4.4.2012).

Sturman,A.(1997).Casestudymethods.In:J.P.Keeves(ed.).Educational research, methodology and measurement: an international handbook(2nded.).Oxford:Per-gamon,pp.61–66.

Thomas,G.(2011).ATypologyforthecasestudyinsocialsciencefollowingareviewofdefinition,discourseandstructure.Qualitative Inquiry,17,Issue6,pp.511–521.

Vehovar,U.(2005).FraneAdam,MatejMakarovič,BorutRončević,MatevžTomšič:Thechallengesofsustaineddevelopment:Theroleofsocio-culturalfactorsinEast-centralEurope:recenzija.Družboslovne razprave,21,Issue49–50,pp.309–311.

Verschuren,P.J.M.(2003).Casestudyasaresearchstrategy:someambiguitiesandopportunities.International Journal of Social Research Methodology,6,Issue2,pp.121–139.

Vogrinc,J.(2008).Kvalitativno raziskovanje na pedagoškem področju.Ljubljana:Pedagoškafakulteta.

Yin,R.K.(2009).Case study research: design and methods (4thed.). ThousandOaks,CA:Sage.

Documents

The case study as a type of qualitative research