Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
American Marketing Association is collaborating with JSTOR to digitize, preserve and extend access to
Journal of Marketing Research.
http://www.jstor.org
Since Hippocratesfirstpresented14 classic case stud- premises,given perfectmeasuresthe validityof all find-
ies of disease some 2300 years ago, science has pro- ings dependssolely on the quality of logic employed.
ceeded along two divergentknowledgepaths. One in- However,in practiceprecisionin measurementoften is
volves formulatinga tentativetheoryof a phenomenon obtainedat the expenseof the currency,contextualrich-
"writlarge,"deducingimpliedempiricalconsequences, ness, or "externalvalidity"of the findings.
andcontrollingsituationalevents in orderto observethe The tradeoffof increasedprecisionfor reducedgen-
validityof empiricaldeductions.The second path, less eralityis not a useful one in all situations;many inter-
frequentlyused but equallyvalid, is to reasonfrom in- estingphenomenacannotbe understoodif removedfrom
dividualand naturallyoccurringbut largely uncontroll- their social context. Thus, some researchershave re-
able observationstowardgeneralizableinductiveprin- vived inductive, qualitative,and clinical researchap-
ciples. proachesas an alternativepathto scientificlearning.
Researchin marketing(and more generallyin social Thepurposeof this articleis to advocatecase research
science)hasemphasizedthe role of deduction.In its pur- to marketersas one methodof the inductiveapproach,
est form, deductivereasoningis exemplifiedby the log- and to argue generallyfor more applicationsof quali-
ical syllogism, wherebyone poses premisesand draws tativetechniquesin marketingscience. Caseresearch,as
fromthemsomeconclusion.Becausethe conclusioncan the termis used here, refersto the qualitativeand field-
contain no more informationthan is containedin the basedconstructionof case studies.It is guidedby a pro-
cess model of discoverywhich leads to (1) a set of the-
oreticalgeneralizationsfrom the clinical observations,
*ThomasV. Bonomais AssociateProfessor,GraduateSchool of (2) clinical "constrainttesting"of these generalizations,
BusinessAdministration, HarvardUniversity. andeventually(3) a clinicallyvalidatedtheoryof some
The authoris gratefulto the Editorand threeanonymousJMRre-
viewersfortheircommentson themanuscript. KennethWongof Queens marketingphenomenon.
In the next sectionof the article,discussionof the util-
Universitycontributedmateriallyto a previousversionof the article.
ity and limitationsof case researchis framedwithin a
199
measures,and control are more difficult to achieve. Often, previous time. For instance, without evidence of asso-
the latter kind of study demands a greater use of sub- ciation there is no a priori reason to consider issues of
jective or, at best, clinical analysis. Thus, in Figure 1 cause and effect. Similarly, until a phenomenon has been
such methods as laboratory experiments are positioned described, it cannot be adequately classified nor can op-
in the top left corer (high data integrity, low currency), erational measures be defined. In this sense, studies to-
whereas methods such as field studies and case re- ward the description end of the continuum might be as-
search-which seek to relate phenomena as perceived sociated more frequently with theory building, whereas
by the investigator in natural settings-are oppositely those near the cause-and-effect end are more frequently
placed (low data integrity, high currency). used for theory disconfirmation.
Ideally, a researcher can simultaneously pursue high As researchers move toward the causal end of the re-
levels of data validity and generalizability by adopting search hierarchy, the tensions of knowledge accrual tend
triangulationstrategies which provide replication and/or to direct investigators toward methods high in data in-
corroborationof findings across methods within a single tegrity. High data integrity builds confidence "that an
researchproject. The "multitrait-multimethod"approach observed relationshipbetween variables is causal and that
of Campbell and Fiske (1959) exemplifies this type of the absence of a relationship implies no cause" (Cook
endeavor. In practice, however, within-project triangu- and Campbell 1979, p. 84). Once causation is estab-
lation is rarely used. As Martin (1981) points out, there lished, replications (usually by other high data integrity
often are technological and expense barriers to using methods) may be used to test the limits of (or to cor-
triangulation within any one research project. roborate) observed relationships.
Ordinarily, science implicitly relies on "interproject" When the existing body of knowledge or theory is well
triangulationto help accrue a body of knowledge satis- developed, the use of methods oriented toward the lower-
fying the two desiderata of integrity and currency. Re- right apex of Figure 1 may be inefficient. Though the-
searchers pursuing similar topics may choose different oretical propositions can be confronted with naturalistic
methods and contexts. This is not bad, for when there observation, the disconfirmation of propositions solely
is experimental and field study and case confirmation with these methods is ill-advised. Lower apex methods
that a given phenomenon holds, one may feel more con- on the figure risk measurement bias and error due to re-
fident about the phenomenon even though some studies searchersubjectivity(Gordon 1976) or, as London (1964)
are low in currency and others have data integrity prob- states, "the imposition of value and the fact upon each
lems. The notion of cross-method, cross-project vali- other." It is not surprising that when the focal concerns
dation is similar to Leone and Schultz' (1980) notion of are causal limits, quantification, and the attendantability
research corroboration. to attain measurable internal validity, statistical conclu-
In some research areas, however, such corroboration sion validity, and falsification, the findings of studies in
does not occur because the "feasibility constraint"is nar- the lower-right apex of Figure 1 are considered suspect
rowed unnecessarily by habitually chosen methods. If or even "unscientific" (van Maanen 1979b).
the methods set used is narrow, a balance of inquiry is However, when researchers' interests or phenomenon
less likely to be achieved in a discipline that marries in- requirements dictate theory building rather than verifi-
tegrity with currency across studies. cation or extension, the tasks of description, classifica-
tion, and comparison become relevant. Then methods
Making the Tradeoffs oriented toward the lower-right apex of Figure 1 may be
Though a broad set of sociology of science variables more efficient than others. If properly conducted, re-
affects methods choice (cf. Mitroff 1974), the focus here search by these methods can provide a "deep under-
is on aspects of research itself which might help re- standing" (Geertz 1973), a fuller contextual sense of the
searchersidentify those occasions on which one desider- phenomena under study (Miles 1979), and an explicit
atum should be preserved at the expense of another. Two provocation toward theory building that often is missing
characteristics of research problems are considered, the from both simple descriptive work and most cause-and-
purpose of the research and the nature of the phenom- effect research (van Maanen 1982a). Presumably, re-
enon under study. searchersarmed with understandingso gained may then,
Purpose of the research. Following Simon (1978), under appropriateconditions, seek to move up the hi-
McGrath (1982), and Cook and Campbell (1979), one erarchy of study types toward causal disconfirmation.
may discriminatepoints along a hierarchyof study types. When methods oriented toward the upper-left apex of
The different types of studies appear to engage the re- Figure 1 are used to investigate research topics about
search tensions differently. One common listing of the which theoretical development is scant or uncertain, re-
points along the research continuum might be: descrip- search often is inefficient or misleading. Either the power
tion, classification, comparison, measurement/estima- of deductive methods is underutilized, or theory and/or
tion, establishing association, and determining cause and method are prematurely pressed into service when their
effect. At each sequential step in the research hierarchy underlyingassumptionscannot be met. In the formercase,
one presumes that the preceding types of studies have superior understandingcould be achieved by using tools
been undertaken, either formally or informally, at some that allow development of in-depth information on con-
emergence of qualitative research in the disciplines of CASES, CASE TEACHING, AND CASE RESEARCH
sociology and psychology: "The sources of disenchant- Case studies are most familiar to marketers as a peda-
ment [with quantitative/deductive tools] are many, but
gogical device, or as a means of generating exploratory
deservingof passing note are: the relatively trivial amount insights prior to more "rigorous" investigations. Here,
of explained variance, the abstract and remote character neitherof these uses of cases is viewed as case research;
of key variables, the lack of comparability across stud-
rather, the use of cases as research tools is our focus.
ies, the failure to achieve much predictive validity . . .
and the causal complexity of multivariateanalysis, which, Though examples of case research qua research can be
found (c.f., Bonoma, in press; Corey 1978; Corey and
even when understood, makes change-oriented actions Star 1971), little guidance about how to conduct mar-
difficult to contemplate" (p. 13).
A growing number of researchers in economics (e.g., keting case research is available, except in literaturesnot
often examined by marketing researchers (e.g., Geertz
Piore 1979), medicine (e.g., Feinstein 1977), organi-
1973; van Maanen 1982a). In this section, therefore, we
zational behavior (e.g., Fombrun 1982; van Maanen discuss the nature of a case, then differentiate the use of
1979a), sociology (McGrath, Martin, and Kulka 1982; cases for teaching, prescientific, and research purposes,
Mitroff 1974), and psychiatryhave advocated and helped and set the stage for discussion of a four-stage qualitative
foster rebirth of qualitative research in the social sci- research process intended to guide qualitative and case-
ences. Some of these researchers have gone so far as to based research endeavors.
say that, given the low level of theoretical knowledge
about phenomena in which social science is interested, Case Studies
coupled with the known complexities and context-sen- Defined most generally, a case study is a description
sitivities of these same phenomena, qualitative research of a management situation. As such, it is the marketing
is the major or even the only valid knowledge-accrual
device open to scientists whose interests are focused on analogue of the physician's clinical examination (e.g.,
MacLeod 1979), and relies on a similar appeal to mul-
human behavior. Though we do not go so far, it may be
noted that many important marketing phenomena meet tiple data sources for reliable diagnosis (cf. Leenders and
Erskine 1978). Though case studies familiar from class-
the dual conditions of little theoretical knowledge and room use usually focus on some problem of high cur-
high complexity. Such phenomena should be suited to rency to firm management and have broad pedagogical
the application of qualitative research methods. How-
appeal, cases without any problem focus can be con-
ever, little trend toward qualitative research has yet been structed to learn about the operation of a healthy man-
observed in marketing.
Because of marketing's quantitative/deductive re- agement or marketing organization. Thus, though man-
search roots, many marketing subject areas not amenable agement "disease" often is the stimulus for case
to study by the methods oriented toward the top-left apex construction, a problem focus is not required.
of Figure 1 have received little research attention of any Second, case construction implicates multiple data
sources. Like other qualitative methods, cases often rely
sort. For instance, though much is written about nor-
mative pricing strategy formation, almost nothing is heavily on verbal reports (personal interviews) and un-
obtrusive observation as primarydata sources. However,
known descriptively about how (or whether!) managers case method is distinguished from other qualitative
engage these strategies under real-world pressures. In- methods in that it involves numerous other data sources,
deed, little is known about what constitutes effective some of which are quantitative. These other data sources
marketing management in practice (or whether practice serve as a means of "perceptualtriangulation"4and pro-
is consistent with what little is known from theory, sur- vide a fuller picture of the business unit under study.
vey verbalreports,or studentsimulations).What is known Prime among these sources are financial data (e.g.,
about such questions often evolves from practical ex-
budgets, operating statements), marketperformance data
perience, undocumented analogies with other disci- (e.g., share, sales by territory), and market and com-
plines, and common-sense reasoning. The apparent re- petitive data (e.g., product replacement rates, competi-
search bias toward types of investigation that preserve tive spending levels). Additional data sources consulted
data integrity at the expense of currency results in a include written archives (e.g., memoranda), business
methodological one-sidedness that may impair the de- plans, and direct observations of management interac-
velopment and testing of sound theories. tions.
In sum, there is a role and a need for a much broader
set of knowledge-accrual mechanisms than those con-
ventionally employed in marketing research. In partic-
ular, methods toward the lower-right apex of Figure 1 4Thenotionof "perceptualtriangulation" raises the point that, in
seem especially well-suited to aspects of marketingwhere all qualitativeresearch,knowledgedependsheavilyon the perceptions
there is a relatively thin theoretical base or complex ob- of theactorsandof the observer(orcase compiler).FollowingGeertz,
servational task. One such method found promising by however,knowledgecan only be consideredknowledgewithin the
confinesof someone'sperceptualframework,which is called "sig-
many researchers(e.g., Duncan 1979; McClintock, Bar- nification."The factthatwe knowso littleof how managersperceive
nard, and Maynard-Moody 1979) is case research. marketingrealitiesis a majorconcern.
5The idea that direct observation is required for case research means 6However, this is not to say that cases compiled for research pur-
that a common practice-using previously compiled cases either to poses cannot be used to good effect in the classroom. To the contrary,
illustrate theoretically important phenomena or to form the basis for one benefit of case research is that it can offer a substantial oppor-
theory building-cannot be called scientific and in no way constitutes tunity to integrate course development with research progress (see
case research as the term is used here. Bonoma 1984).
There are identifiable sets of research situations (and Fombrun,CharlesJ. (1982), "Strategiesfor NetworkResearch
stages of research design) where the qualitative, in-depth in Organizations,"Academyof ManagementReview,7 (2),
natureof these "high currency" approaches is desirable, 280-91.
even if accompanied by some risks to data integrity. In Geertz,CliffordA. (1973), TheInterpretationof Cultures.New
York:Basic Books.
particular,they are useful when a phenomenon is broad Gordon,R. (1976), "Ethnomethodology-RadicalCritique,"
and complex, where the existing body of knowledge is HumanRelations,29 (March),193-202.
insufficient to permit the posing of causal questions, and Guilford,J. P. (1954), PsychometricMethods,2nd ed. New
when a phenomenon cannot be studied outside the con- York:McGraw-HillBook Company.
text in which it naturally occurs. However, these are Kelly, GeorgeA. (1955), The Psychologyof Personal Con-
"features" of appropriate problems and not the major structs.New York:W. W. Norton(2 vols.).
"benefits" of nontraditional approaches. The real bene- Leenders,MichielR. andJamesA. Erskine(1978), Case Re-
fits are that clinical research methods offer the potential search: The Case WritingProcess, 2nd ed. London, On-
to deepen the perspective of persons performing and us- tario:Researchand PublicationsDivision, School of Busi-
ness Administration, The Universityof WesternOntario.
ing research, and that they expand the range of research
Leone,RobertP. andRandallL. Schultz(1980), "A Studyof
problems that can be considered. The implementation of
MarketingGeneralizations," Journal of Marketing, 44
Kelly's (1954) notion of "constructive altemativism" in (Winter),10-18.
research methods, the ability to construe knowledge in
London,P. (1964), TheModesand Moralsof Psychotherapy.
novel and useful ways, is central to maintaining the vi- New York:Holt, Rinehartand Winston.
tality of any area of inquiry. MacLeod, John (1979), Clinical Examination.Edinburgh:
ChurchillLivingstone.
REFERENCES Martin,Joanne(1982), "A GarbageCan Model of the Re-
searchProcess,"in JudgmentCalls in Research,JosephE.
Bonoma,ThomasV. (1984),ManagingMarketing:Text,Cases McGrath,JoanneMartin,and RichardA. Kulka,eds. Bev-
and Readings.New York:The Free Press. erly Hills, CA: Sage Publications,17-40.
(in press), The MarketingEdge: MakingStrategies McClintock,CharlesC., D. Barnard,andS. Maynard-Moody
Work.New York:The Free Press. (1979), "Applyingthe Logic of Sample Surveysto Quali-
, G. Zaltman,and W. J. Johnston(1977), Industrial tative Case Studies:The Case ClusterMethod,"Adminis-
BuyingBehavior. Cambridge,MA: MarketingScience In- trativeScience Quarterly,24 (December),612-29.
stitute. McGrath,Joseph E. (1982), "Dilemmatics:The Study of
Brunswik,E. (1952), "The ConceptualFrameworkof Psy- Choices and Dilemmas,"in JudgmentCalls in Research,
chology,"in InternationalEncyclopediaof UnifiedScience, JosephE. McGrath,JoanneMartin,andRichardA. Kulka,
O. Neurath,ed. Chicago:Universityof ChicagoPress. eds. BeverlyHills, CA: Sage Publications,69-102.
Calder,Bobby J. (1977), "FocusGroupsand the Natureof , JoanneMartin,andRichardA. Kulka(1982), "Some
QualitativeMarketingResearch,"Journalof MarketingRe- Quasi-Rulesfor MakingJudgmentCalls in Research,"in
search, 14 (August),353-64. JudgmentCalls in Research, Joseph E. McGrath,Joanne
Campbell,D. T. and D. W. Fiske (1959), "Convergentand Martin,andRichardA. Kulka,eds. BeverlyHills, CA: Sage
DiscriminantValidationby the Multitrait-MultimethodMa- Publications,103-18.
trix,"PsychologicalBulletin, 56, 81-105. Miles, MatthewB. (1979), "QualitativeData as an Attractive
and J. C. Stanley (1963), Experimentaland Quasi- Nuisance,"AdministrativeScience Quarterly,24 (Decem-
ExperimentalDesigns for Research. Chicago: Rand Mc- ber), 590-601.
Nally. Mitroff,Ian M. (1974), The SubjectiveSide of Science. New
Cook, Thomas, D. and DonaldT. Campbell(1979), Quasi- York:Elsevier.
Experimentation: Design and Analysisfor Field Settings. Nisbett,RichardE. and T. D. Wilson (1977), "TellingMore
Boston:HoughtonMifflin and Company. thanWe CanKnow:VerbalReportson MentalProcesses,"
Corey,E. R. (1978), ProcurementManagement:Strategy,Or- PsychologicalReview, 84 (March),231-59.
ganizationand Decision Making. Boston: CBI Publishing Piore,MichaelJ. (1979), "QualitativeResearchTechniquesin
Company. Economics,"Administrative ScienceQuarterly,24 (Decem-
and S. Star(1971), OrganizationalStrategy:A Mar- ber), 560-9.
ketingApproach. Boston: Division of Research, Harvard Runkel,P. J. and McGrath,JosephE. (1972), Researchon
BusinessSchool. HumanBehavior:A SystematicGuide to Method. Skokie,
Cronbach,L. J. and P. E. Meehl (1955), "ConstructValidity IL: RandMcNally.
in PsychologicalTests," PsychologicalBulletin, 52, 281- Salancik,GeraldR. (1979), "FieldStimulationsfor Organi-
302. zationalBehaviorResearch,"Administrative Science Quar-
Dabbs, James M., Jr. (1982), "MakingThings Visible," in terly, 24 (December),638-48.
Varietiesof QualitativeResearch,Johnvan Maanan,J. M. Sellitz, Claire,MarieJahoda,MortonDeutsch,and StuartW.
Dabbs,Jr., andRobertR. Faulkner,eds. BeverlyHills, CA: Cook (1959), ResearchMethodsin Social Relations. New
Sage Publications,31-64. York:Holt, Rinehartand Winston.
Duncan,RobertR. (1979),StrategicManagement: A New View Simon, JulianL. (1978), Basic ResearchMethodsin Social
of BusinessPolicy. Boston:Little, Brown. Science:The Art of EmpiricalInvestigation,2nd ed. New
Feinstein,Alvan R. (1977), "The Haze of Bayes, the Aerial York:RandomHouse.
Palacesof Decision Analysis, and the ComputerizedOuija van Maanen,John(1979a), "ReclaimingQualitativeMethods
Board,"ClinicalPharmacology and Therapeutics,
21 (April), for OrganizationalResearch:A Preface," Administrative
482-96. ScienceQuarterly,24 (December),520-6.
ADVERTISERS' INDEX
The Ehrhart-Babic
Group...................................................................... Fl
Namelab...................................................................... . 157
SurveySampling......................................................... Cover 2