Case study a case for case studies exploring the use
Upcoming SlideShare
Loading in...5

Case study a case for case studies exploring the use






Total Views
Views on SlideShare
Embed Views



0 Embeds 0

No embeds



Upload Details

Uploaded via as Adobe PDF

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
Post Comment
Edit your comment

Case study a case for case studies exploring the use Document Transcript

  • 1. Journal of Advanced Nursing, 2000, 31(4), 926±934 Methodological issues in nursing researchA case for case studies: exploring the useof case study design in community nursingresearchAnn Bergen BA MSc RGN DipN DNCert Cert Ed DNTHonorary Research Fellowand Alison While BSc MSc PhD RGN RHVProfessor of Community Nursing, Research in Health and Social Care Section,Florence Nightingale School of Nursing and Midwifery,Kings College London, London, EnglandAccepted for publication 17 September 1999 BERGEN A. & WHILE A. (2000) Journal of Advanced Nursing 31(4), 926±934 A case for case studies: exploring the use of case study design in community nursing research The case study has become an accepted vehicle for conducting research in a variety of disciplines. However, the meaning behind the term is not always made explicit by researchers and this has given rise to a number of assumptions which are open to challenge, and to questions about the robustness of the method. This paper explores some of the issues arising from one particular de®nition of case study research, used in a study by Yin which examined the practice of case management in community nursing. Four main areas are discussed. First, de®ning `case is seen to pose questions about the relationship of the phenomenon to its context, the degree of researcher control over case de®nition, the limits to what may constitute a `case and what is meant by the term `unit of analysis. Second, the relevance of external validity to case study research is supported through the use of a number of tactics, in particular Yins concept of replication logic, which involves generalizing to theory, rather than to empirical data. Third, the use of method triangulation (multiple methods of data collection) is advanced as a means of enhancing construct validity in research where data converge around a particular theory. Finally, the relation- ship of the case study to theory construction, through the prior development of `propositions is discussed. Each of these issues is applied to the design and conduct of a research study based closely on Yins multiple case study framework. Thirteen `cases were selected of case management practice and data were collected through interviews and examination of literature and docu- mentation, to explore the suitability of community nurses for the role. It is concluded that, given the appropriate subject matter, context and research aims, the case study method may be seen as a credible option in nursing research.Correspondence: Ann Bergen, Florence Nightingale School of Nursing andMidwifery, Kings College London, James Clerk Maxwell Building, WaterlooRoad, London SE1 8WA, England.926 Ó 2000 Blackwell Science Ltd
  • 2. Methodological issues in nursing research Case study design Keywords: case, case management, case study, unit of analysis, research methods, nursing, health care account of the Chicago School of Sociology, has illustratedINTRODUCTION this in putting forward their view:The case study is a familiar yet elusive approach to In essence, the term `case study referred to the collection ofresearch. It is familiar because it has been promoted by detailed, relatively unstructured information from a range ofresearchers and writers from a number of disciplines, for sources about a particular individual, group or institution,example education (Ball 1983, Burgess 1985, Hammersley usually including the accounts of subjects themselves.1986, Stake 1995), experimental psychology (Barlow & (Hammersley 1989 p. 93)Hersen 1984, Robson 1993, Yin 1994) and nursing (Hutch-inson 1990, Ross & Tissier 1994, Woods 1998). It is It may be suggested that `relatively unstructured infor-elusive, at least in nursing, because the case study method mation sits uneasily with the traditional de®nition ofis usually only given minimal attention in general research as a systematic process, involving discipline andresearch textbooks (for example Polit & Hungler 1999, structure (for example Macleod Clark & Hockey 1989,Burns & Grove 1997) and because much research Powers & Knapp 1990), aimed at extending a given body ofappearing in academic nursing journals which claims to knowledge. This is not to say that Hammersley himselfuse the design fails either to de®ne the authors interpret- supported the view of the Chicago School Ð indeed hisation, or to offer a rationale (Woods 1997). earlier collection of case studies in classroom research Stake (1994, 1995), who has written extensively on the (Hammersley 1986), which included a variety of approa-case study method, has commented that: `labels contribute ches, does not indicate any particular view. Nevertheless,little understanding of what researchers do (Stake 1994 his edited collection of school studies (Hammersley 1983)p. 236), implying the need to describe and justify a chosen appears to support the placing of case study researchresearch method, rather than assume an accepted ®rmly within the qualitative (and, more speci®cally, themeaning. This is particularly true when the method in ethnographic) paradigm (Ball 1983).question Ð unlike, for example, the questionnaire or This is a not uncommon position amongst researchers. Inexperiment Ð has different interpretations. Thus: nursing, the case study has been categorized by Parse et al. (1985) as an example of the descriptive method within the¼ custom is not so strong that researchers (other than graduate qualitative framework. She saw the purpose of the methodstudents) will get into trouble by calling anything they please a as the in-depth investigation of a particular unit or institu-case study. (Stake 1994 p. 237) tion, a view subsequently accepted by other writers (for The aim of this paper is to clarify what is meant by the example, Hutchinson 1990, Powers & Knapp 1990).label `case study research, through an exploration of its In contrast stands Yins (p. 13) de®nition, which saw thede®ning characteristics and some of the issues arising case study as an empirical enquiry that:from its application. The discussion will be illustrated · investigates a contemporary phenomenon within itswith particular reference to the work of one of the chief real-life context; whenproponents of the method, R. K. Yin, and to a particular · the boundaries between phenomenon and context areresearch study making use of case study principles, not clearly evident.which sought to identify the current and potential rele-vance and value of case management to community His de®nition was further elaborated in a second set ofnursing. conditions (Ying 1994 p. 13). The case study enquiry: · copes with the technically distinctive situation inDEFINING CASE STUDY RESEARCH which there will be many more variables of interest than data points, and as one resultA variety of schools of thought have in¯uenced the · relies on multiple sources of evidence, with datadevelopment of the case study method throughout the converging in a triangulating fashion; and as another19th and 20th centuries, from disciplines as disparate as resultanthropology and sociology, on the one hand · bene®ts from the prior development of theoretical(Hammersley 1989, Hamel et al. 1993), to pure science propositions to guide data collection and analysis.and the single case experiment on the other (Barlow &Hersen 1984). This poses problems when seeking a This at once differentiates case study research, not onlyde®nition of the method and Hammersley (1989), in his from experimental and quasi-experimental designs, whichÓ 2000 Blackwell Science Ltd, Journal of Advanced Nursing, 31(4), 926±934 927
  • 3. A. Bergen and A. Whileaim to divorce the phenomenon under study from its Case and contextcontext, but also from historical research, which does notstudy contemporary events, and surveys, which attempt to In stating, in his de®nition, that the boundaries betweenlimit the number of variables under investigation. Yin phenomena and context are not clearly demarcated, Yinsubsequently challenged many of the commonly held (1994 p. 13) seemed clear that the case study wouldassumptions about case study research; it was not to be necessarily include data relating to that context becauseseen as synonymous with ethnography, nor did it always the researcher `deliberately wanted to cover contextualfocus on one single person, group or institution. Finally, conditions. This appears to have marked a developmentthough in-depth in nature, data were certainly not to be in Yins thinking, since the ®rst and revised editions of hisseen as lacking structure. basic text (Yin 1984, 1989) failed to mention this inclu- However, though persuasive and comprehensively siveness, while his later book of applications of case studyexplored, Yins concept of case study research does bear method (Yin 1993) implied the point only through thescrutiny, particularly in the context of any particular addition to his de®nition for the ®rst time of the clauseapplication. Four major areas of debate will therefore be about variables and data points (see above). At the sameconsidered in relation to the strategy adopted in a piece of time, his emphasis on the need clearly to de®ne the caseempirical research. These issues focus on: (i) case and meant that case-related and context-related data were notcontext; (ii) external validity; (iii) triangulation; and to be treated in the same way.(iv) the relationship to theory, and are roughly equivalent This poses challenges for the researcher, not leastto the four major sections of Yins de®nition. The discus- because his examples in the `applications book (Yinsion will be preceded by a brief outline of the studys 1993) fail to illustrate in detail how this principle shouldsubject matter. be operationalized. Nevertheless, Yins unfolding theory is perhaps more helpful than that of Stake, who wrote, on the one hand, about the case as `a bounded system (StakeTHE RESEARCH STUDY: CASE MANAGEMENT 1995 p. 2) and `the boundaries of the case (Stake 1994AND COMMUNITY NURSING p. 237), while referring, on the other hand, to the `in®n-The term `case management (or the governments more itely complex nature of case and context, where `therecently preferred usage `care management) has, over the phenomena are ¯uid and elusive (Stake 1995 p. 33).last 10 years, become familiar to community nurses in all Contextual issues were shown in the literature to bespecialisms. This is largely due to the prominence it was important to the phenomenon of case management, in thegiven in the White Paper Caring for People (Department form of national policy, professional theory and localof Health, DoH 1989), where it was proposed, along with organizational structures, so Yins speci®c counsel ofjoint assessment, as the `cornerstone of high quality care inclusion would certainly seem appropriate here.(para.1á11). The government was not prescriptive about Most researchers appear to support Yins emphasis onthe professional background most appropriate for the the importance of a clear de®nition of `the case. However,case manager role; however, the research undertaken to experts differ as to whether this should be imposed by thedate displays, in the main, a social services orientation. researcher or be evolutionary in nature. Hamel et al.Although there is evidence of nurses undertaking case (1993) recommended the active selection of the ideal casemanagement, the nursing perspective has not been a in order to grasp the object of study and that the researcherfeature of either individual projects nor reviews (for intervene to produce a de®nition rather than having itexample, Challis 1994, Phillips & Penhale 1996). imposed by the ®eld or key informants. Robson (1993),A search of the literature (Bergen 1992) highlighted a who, in many respects, comes closer to Yin in hisnumber of questions relating to the potential of commu- conceptualization of the case study than most othernity nurses within the case management remit. It was writers, has referred to both `prestructured and `emergentthese questions which the research study attempted to designs, with a necessary trade-off between looseness andaddress. selectivity, meaning that most research falls somewhere The research was based on a multiphase design, further between these two extremes. But Robsons `real worlddetails of which may be found elsewhere (Bergen 1993, research leaves room for a type of pragmatism which Yin1994, 1995, 1997). The case studies comprised the third would probably denounce as unsatisfactory.phase (following a telephone survey and a questionnaire Although not overtly allied to one particular camp here,survey) and aimed to provide an in-depth analysis of case Yin implied researcher control over the case de®nition inmanagement projects involving community nurses along a his assertion that it should follow logically from the naturenumber of parameters identi®ed in the literature and the of the research question and focus on the possession ofearlier phases. The issues confronting the researcher in characteristics of interest. Moreover, so that ®ndings canthis undertaking are addressed in terms of the four main be compared with previous research, `key de®nitionsareas of debate identi®ed above. should not be idiosyncratic. Rather, each case study¼.928 Ó 2000 Blackwell Science Ltd, Journal of Advanced Nursing, 31(4), 926±934
  • 4. Methodological issues in nursing research Case study designshould be similar to those previously studied by others(Yin 1994 p. 25). This approach would sit readily with thesubject of case management, as it appears in the literature. In seeking a de®nition, the question arises of whetheranything (within, perhaps, the above limitations) can be acase if so designated. Stake (1995) strongly countered thisposition with the example of a teacher who, he suggested,could appropriately constitute a case, but not her teaching,since this would lack speci®city. Interestingly, Yin (1994)disagreed with this assertion, claiming Stake himself to betoo broad in his approach. Yins wording, in a footnote tothe subject of de®nitions, implied that Stake failed todifferentiate the case study (a research method) and thecase (its object). De®ning the former in terms of an entity Figure 1 Basic types of design for case studies.would mean that any object could then be labelled a casestudy regardless of methodology used. This would clearly the `case within a case as he studied classroom, teacherbelittle the discipline. The case, on the other hand, and pupil in order to build up the picture of the schoolaccording to Yin (1994 p. 22) `can be some event or entity itself. Stake is one of only a few researchers, like Yin, tothat is less well de®ned than a single individual (the detail the case study approach in action, which is bothusual choice and that adopted by Stake) and `case studies useful for other researchers and adds credibility to theirhave been done about decisions, about programs, about theory.the implementation process and about organizational In the case management research illustrated here, thechange. issue was dealt with in two ways. First, it was decided to There is a substantial difference here, and this broader adopt Yins (and Stakes) notion of designating a unitview was felt to hold potential for the present study, smaller than the case for purposes of analysis, in order towhere case management practice, within a given location build up the case picture. Thus the case (community nurse(case study site), rather than the practitioner, would case management practice within the case study site) wasprovide the most useful focus because the research ques- subdivided into its component parts (the practices oftion revolved around professional theory and practice individual case managers). In order to be clear aboutfrom two disciplines (nursing and case management) and terminology, the former (the case) was termed the maintheir interrelationship. Individual practitioners could be unit and the latter (individual practice) a subunitseen as the vehicles for this practice. (Figure 2). This overcomes Yins rather confusing and Discussion about de®ning `the case prompts clari®ca- various references to the `unit, `embedded units and, intion of other terminology, in particular the term `unit of one instance, `subunits (Yin 1994 p. 41). However, unlikeanalysis. This latter term is one example of research Yin and Stake in their implications (though the point wasterminology which is commonly used and accepted, butrarely de®ned. Yin appeared, initially, to imply equiv-alence with case in his reference to `the de®nition of theunit of analysis (and therefore of the case)¼ (Yin 1994p. 22, authors emphasis) and this may be taken further asread in his summary of the same section, where hereferred to `the unit of analysis (or the case itself)¼ (Yin1994 p. 44). However, in a later discussion on different types of casestudy designs, Yin described a 2 ´ 2 matrix, where thehorizontal axis represented single and multiple casedesigns, while the vertical axis differentiated between`holistic and `embedded designs, since `within these twotypes [that is, single and multiple case designs] there canalso be unitary or multiple units of analysis (Figure 1)(Yin 1994 p. 38). In other words, there is an implicationthat the case may be disaggregated for the purpose ofanalysis. This is a point not addressed by the other mainwriters on case study research, although Stakes (1995)illustrative vignette of Harper School made reference to Figure 2 The case and the context.Ó 2000 Blackwell Science Ltd, Journal of Advanced Nursing, 31(4), 926±934 929
  • 5. A. Bergen and A. WhileFigure 3 Variables, data points and multiple sources of evidence.made explicit by neither) the unit and subunit were data on professional disciplinary theory and nationalqualitatively the same, that is, comprised of case manage- policy, respectively (though interviews could providement practice, rather than, for example, individuals. It was further information on how these were operationalized).thought this provided the sounder rationale for overall Thus the data collection methods equated with each of theconclusions about the case. multiple sources of evidence used, while the individual Second, the term `unit of analysis, like the term `case, sources of data within these methods equated to what Yinwas rede®ned to embrace two distinct meanings, and here termed `data points. Although Yin failed to de®ne thisthe terminology of Moser & Kalton (1971) was found term, this interpretation is consistent with his usage. Forhelpful. This differentiated sampling units (the sources of national policy documentation this meant the White Paperdata collection) and units of enquiry (the subjects or (DoH 1989), statute (the NHS & Community Care Act 1990)variables to be measured). Thus, units of enquiry and implementation guidance (for example DoH/SSIcomprised the case as the main unit and individual case 1991); for literature this meant papers on case manage-management practices as the subunits. Since data about ment, nursing and research, for local documentation thisthe relevant contexts were also to be included, these were meant policies, protocols and standards, and for interviewtermed contextual units (of enquiry). Determination of the data this meant interviews with the individual nurse-casesampling units depended on the chosen methods of data managers, their line managers, patients and carers.collection, and in this instance it seemed that interviews A schema of this terminology used in the research designwith case management project participants had the best is depicted in Figure 3 and 17 case study sites werepotential to elicit the type of data required to expand on identi®ed from the questionnaire data for in-depth ana-the questionnaire replies and ful®l the research aims. This lysis, based on this plan.was identi®ed as a principle data collection method incase study research by Stake (1995). In addition, the External validityscrutiny of documentation relating to local policies andprotocols provided contextual data on local organizational An understanding of the nature of the sampling unit isstructures and practices. crucial to an understanding of the nature of external The literature on case management and national policy validity within case study design. Critics have argued thatdocuments already reviewed supplied further contextual case study research is a poor basis for generalization930 Ó 2000 Blackwell Science Ltd, Journal of Advanced Nursing, 31(4), 926±934
  • 6. Methodological issues in nursing research Case study design(Stake 1995) but this criticism is based on traditional Yin (1994) advanced the notion of generalizability insampling theory, itself based on the representativeness of case study design through his idea of analytical general-sample selection and the consequent ability to make ization. The vast number of variables under study ininferences about a population. Since case selection is comparison to the number of data points available hasbased largely on other factors, as noted above, the question already been seen to make statistical generalizationarises as to whether generalizability is a redundant impossible, but generalization at the level of theory (ratherconcept in this context. than empirical data) becomes possible providing a the- Most proponents would dispute this view, although the oretical framework is developed ®rst. Yin outlined aterminology may need to be changed. Stake (1995 p. 7) number of strategies for achieving this and in multiplesuggested that the purpose of much research, not just case case study design the favoured approach is a two-stagestudy approaches, is seldom an entirely new under- process. The ®rst stage consists of `pattern matching,standing of phenomena, but rather a `re®nement of whereby `several pieces of information from the same caseunderstanding. This is particularly so with what he may be related to some theoretical proposition (Yin 1994termed the collective case study design, where several p. 25) and this strengthens the internal validity of thecases are selected in order that `by counterexample [each] research where causal relationships are being study invites modi®cation of the generalization¼ The second stage consists of replication logic, which isthough `a positive example is likely neither to establish a analogous to multiple experiments, where the results ofgeneralization nor to modify one, but may increase the the entire case are compared both with other cases andcon®dence that readers have in their (or the researchers) with the proposed theory. Literal replication occurs wheregeneralization (Stake 1995 p. 8). there are predicted similar results. Theoretical replication Stake (1995) identi®ed two further types of case study occurs where there are contrasting results for predictablewhich developed his conceptualization of external validity reasons, and if the cases do not produce results asas applied to this method. The `intrinsic study is a `given predicted, then the initial theory must be revised andin the sense that there is little researcher in¯uence over retested with another set of cases. An important pointcase selection, and the researcher is interested in it: `not with replication logic is that data from sub-units are notbecause by studying it we learn about other cases or about pooled across cases, but analysed within cases prior to asome general problem, but because we need to learn about cross-case analysis at the main unit level.that particular case (Stake 1995 p. 3). Analysis is based This approach seemed suitable for the study of caseupon the direct interpretation of this individual instance management. In many ways the research question and thealone. The `instrumental study, on the other hand, is case de®nition could be seen in terms of Stakes instru-chosen to answer a speci®c research question, and it is the mental study, where the issue (case management practice),underlying issue it exempli®es, rather than the case itself, rather than the particular case, provided the focus. Find-which is important. Analysis is based on a search for ings from one case would also become more meaningfulpatterns across a number of cases and for a re®nement of through the use of replication logic, use being made ofunderstanding through what Stake (1995 p. 74) termed both the literature already reviewed to develop a frame-`categorical aggregation of instances. Thus case studies are work, and the ®ndings from the survey stages to provide aof interest for both their uniqueness and their commonality number of suitable cases, thought to possess the charac-but, as with the positivist approach, a single case as a teristics under examination. However, both stages of Yinsnegative example can limit generalizability. analytical strategy need to be further detailed, and can be Hamel et al. (1993 p. 34) similarly developed a two-fold discussed under the remaining two sections of his de®ni-terminology in their view of the purpose of case study tion given above.research, which was to move `from local to global. Here,the case study was `only microscopic for want of a Triangulationsuf®cient number of cases (Hamel et al. 1993 p. 34),although the authors added that the number was not The use of triangulation in order to undertake the `pattern-paramount providing each case was suitable for the aim. matching analysis was said by Yin (1994) to strengthenHowever, to Hamel et al. (1993) the single case could also the construct validity of the research Ð the establishmentbe seen as representative in its own right provided that of the correct operational measures for the concepts beingthere was suf®ciently detailed description, since this studied. Although Yin followed the commonly acceptedwould lead to a clearer understanding and hence to typology of triangulation (data triangulation, investigatorexplanation. The use of terms such as `explanation of its triangulation, theory triangulation and methodologicalproperties (Hamel et al. 1993 p. 37) and `relationships triangulation), his de®nition of these terms differed fromthat link the parts (Hamel et al. 1993 p. 39) is strongly other exponents. Yin (1994) advocated what he termedsuggestive of Yins (1994) much more explicit view of data triangulation for the case study, though hissubunits within a case. illustration of this on page 93 featured different methodsÓ 2000 Blackwell Science Ltd, Journal of Advanced Nursing, 31(4), 926±934 931
  • 7. A. Bergen and A. While(interviews, surveys, observation, records, etc.), equival- Relationship to theoryent to what Polit & Hungler (1999), for example, termed`methodological triangulation. To Polit and Hungler, data To Yin, the relationship of the case study to theorytriangulation implied one method (for example, inter- development, made explicit in the last of his de®ningviewing) but using multiple data sources (for example, key clauses, underpinned the assumptions underlying hisinformants). Powers & Knapp (1990) implied that both other assertions about the method. The limitations of theusages are acceptable, but made the point that there is no scienti®c method, and a tendency by many researchers toclear agreement on the meaning of triangulation in view the case study as part of the naturalistic researchresearch and it is used in very different ways. Therefore, paradigm, have already been noted. Nevertheless, Yinsthose who use the device must specify its meaning within (1994) favoured analytical device of generalizing to sometheir work. preformulated theory through replication logic strongly The purpose, as well as the type, of triangulation in any suggests an adherence to the positivist approach, aresearch needs to be made clear, since it affects the logic of conclusion which is further enhanced by his reference tothe overall design. Both Powers & Knapp (1990) and a similarity between the case study method and multipleRedfern & Norman (1994) have pointed out a dual appli- experiments.cation here, namely, con®rmation and/or completeness. Yin (1994) accepted that case studies may be descriptiveCon®rmation occurs where data converge around a partic- in nature (indeed, the seminal Street Corner Society ofular theory or proposition and strengthen its claim to be Whyte (1955), often quoted in literature on the case study,upheld. Where the aim is completeness, a method is falls into this category) the emphasis in his writing wasselected for its unique contribution to the research ques- clearly on the explanatory case study. This, like thetion and assists in building up the picture through the experiment, deals with `how and `why questions, ashighlighting of new, and different, theory. Yin (1994) opposed to `who, `what, `how many or `how muchappeared to follow the convergent theorists in his illus- questions, which deal with frequencies and incidencestration of convergence and non-convergence. and are therefore more suitable to survey type approaches. Yin (1994) suggested that the `sources of evidence Questions alone, however, according to Yin, do not pointsuitable for case study research were almost unlimited, to what a researcher should study and therefore there is aand discussed speci®cally six sources Ð documentation, need to develop `propositions about the research, whicharchival records, direct observation, participant observa- re¯ect an important theoretical issue. Only where a topiction, interviews and physical artefacts. Though he is the subject of `exploration Ð and this condition maydetailed their relative strengths and weaknesses, he exist in experiments and surveys as well as case studies Ðconcluded that no single source had advantage over the does a study have a legitimate reason for not having anyrest. Other writers on case study methods have adopted propositions, though, even here, there should be a clearthe same position and Hutchinson (1990) added that it is purpose to guide data collection.the depth and breadth of evidence supporting the case The role of theory development, prior to the conduct ofstudy, rather than a de®nitive method, which is important. data collection, is one point of difference, to Yin (1994), For case management research it was felt helpful to between case studies and related methods, such astriangulate both data collection methods and data sources, ethnography and grounded theory, although, like these,with the object of convergence around the framework the case study may also be used for theory-building. This,constructed from the literature. As noted above, semi- with other comparative parameters, is clear in his ®gurestructured interviews were selected as the major feature, illustrating these methods (reproduced in Figure 4) which,conducted with different key informants (data sources) with the additional property of contextual inclusiveness,from each case and covering the same issues as the initial makes the case study unique among research strategies.questionnaire, but in greater depth. Additionally, docu- Ultimately, of course, the question of theory develop-mentation was scrutinized pertaining to case management ment must be addressed by all researchers if one is topractice within the case study site (mission statements, accept the received de®nition of research. Interestingly,assessment forms, protocols, etc.). Observation, another even the proponents of the ethnographic case studydata collection method common to the case study, was appear near to Yins position when pushed to defenddiscounted on the grounds that case management practice their logic on this point. Hamel et al. (1993) commentedwould be dif®cult to de®ne operationally, and suf®ciently that validating a theory is at the core of the methodologicalinclusively for the purposes of an observation schedule. con¯ict, and advanced their own notion of theoreticalFurther, the inferences made from observed behaviour validation as a way of dealing with the inadequacy of themay have had subjectivity bias in an area given to different single case study as explanatory theory. While stillinterpretations. Interviews, on the other hand, allowed the holding that `all theories are ultimately based on arespondent to speak within his/her own conceptualization particular case or object (Hamel et al. 1993 p. 29) theof the phenomenon, and to make this explicit. theory so produced may be validated by others to assess932 Ó 2000 Blackwell Science Ltd, Journal of Advanced Nursing, 31(4), 926±934
  • 8. Methodological issues in nursing research Case study designFigure 4 Differences in ass-umptions among four eval-uation methods (Yin 1993p. 64).their general applicability. As with Yin, Hamel et al. CONCLUSION(1993) saw knowledge generated in this way as not basedon empirical elements but on a set of proposals or abstract The entire process of case study conduct, in this research,theories. from design through to data collection and analysis, was One point thus almost universally accepted by case thus largely in keeping with Yins (1994) recommendedstudy theorists is that there is some relationship between procedures. However, this did not signify uncriticalthe study and theory, be it theory-testing or theory- acceptance of the interpretation; indeed there were somegenerating. Hutchinson (1990) has made a useful distinc- issues, in particular Yins use of terminology, which ittion between the case study and case history on this point, was felt needed to be re-examined and clari®ed in order tothe latter often taking an atheoretical stance. She pointed be `usable. However, overall, Yins conceptualization ofout, however, that case histories are often claimed, erro- case study research was found to be more comprehen-neously, as case studies by their authors, and readers sively stated than other accounts of the method found intherefore need to be on their guard when reading nursing the literature. Its application to a speci®c piece ofresearch making such claims. research has, it is felt, given the appropriate subject In researching case management, use was made of the matter, context and research aims Ð shown it to be aexistent theory reviewed in the literature. Although this method as credible and robust as other, perhaps moredid not constitute grand theory on the scale often familiar, approaches.associated with discipline-speci®c research, Yinaccepted that this would often be the case, and thatlesser research ®ndings could still be useful. Theory Referencesapplied at two levels. First, commonly used parameters Ball S.J. (1983) Case study research in education; some notes andfor describing case management projects were adopted problems. In The Ethnography of Schooling (Hammersley structuring the descriptive elements of the interview ed.), Studies in Education Ltd, Drif®eld, Northumberside.schedule. In addition, recurring issues from the litera- Barlow D.H. & Hersen M. (1984) Single Case Experimentalture were used to trigger discussion where respondent Designs. Strategies for Studying Behaviour Change 2nd edn.perceptions on role value and appropriateness were Pergamon Press, New York.required. Second, and more generally, the factors Bergen A. (1992) Case management in community care: concepts,appearing as contextual variables in the literature, and practices and implications for nursing. Journal of Advancedhence as units of enquiry in the data collection plan, Nursing 175, 1106±1113.were, by their nature, instrumental in the construction Bergen A. (1993) Towards community care Ð methodological and sampling issues in obtaining a preliminary overview of currentof the theoretical framework for the whole study nursing practice. Health and Social Care in the Community(already illustrated in Figure 3). The underlying `pro- 1, 307±318.positions, as Yin would term them, concerned the way Bergen A. (1994) Case management in the community: identifyingin which these variables combined to address the a role for nursing. Journal of Clinical Nursing 3, 251±257.research question, namely under what circumstances Bergen A. (1995) A Study to Identify the Current and Potentialcase managment may or may not be an appropriate role Relevance and Value of Case Management to Communityfor nurses. Nursing. Queens Nursing Institute, London.Ó 2000 Blackwell Science Ltd, Journal of Advanced Nursing, 31(4), 926±934 933
  • 9. A. Bergen and A. WhileBergen A. (1997) The role of community nurses as care managers. Phillips J. & Penhale B. (1996) Reviewing Case Management for British Journal of Community Health Nursing 2(10), 466±474. Older People. Jessica Kingsley/British Society of Gerontology,Burgess R. (1985) Case study and curriculum research; some London. issues for teacher researchers. In Issues in Educational Polit D.F. & Hungler B.P. (1999) Nursing Research. Principles and Research (Burgess R. ed.), The Falmer Press, London, Methods 6th edn. Lippincott, Philadelphia. pp. 177±196. Powers B.A. & Knapp T.R. (1990) A Dictionary of Nursing TheoryBurns N. & Grove S.K. (1997) The Practice of Nursing Research. and Research. Sage, Newbury Park, California. Conduct, Critique and Utilization. W.B. Saunders, Philadelphia. Redfern S.J. & Norman I.J. (1994) Validity through triangulation.Challis D. (1994) Case management Ð a review of UK develop- Nurse Researcher 2, 41±56. ments and issues. In Caring for People in the Community Ð The Robson C. (1993) Real World Research. A Resource for Social New Welfare (Titterton M. ed.), Jessica Kingsley, London, Scientists and Practitioner-Researchers. Blackwell Science, pp. 91±112. Oxford.DoH (1989) Caring for People: Community Care in the Next Ross F. & Tissier J. (1994) Monitoring and Action Research Decade and Beyond. Cm 849. Her Majestys Stationery Of®ce, Project of Assessment and Care Management in Two North London. Battersea GP Practices. St Georges Hospital Medical School,DoH/SSI (1991) Care Management and Assessment. Practitioners London. Guide. Her Majestys Stationery Of®ce, London. Stake R.E. (1994) Case studies. In Handbook of QualitativeHamel J., Dufour S. & Fortin D. (1993) Case Study Methods. Research (Denzin N.K. & Lincoln Y.S. eds), Sage, Thousand Qualitative Research Methods Series No. 32. Sage, Newbury Oaks. Park, California. Stake R.E. (1995) The Art of Case Study Research. Sage, ThousandHammersley M. (1986) Case Studies in Classroom Research. Open Oaks, California. University Press, Milton Keynes. Whyte W. (1955) Street Corner Society. University of ChicagoHammersley M. (1989) The Dilemma of Qualitative Research. Press, Chicago. Herbert Blumer and the Chicago Tradition. Routledge, London. Woods L.P. (1997) Designing and conducting case study researchHutchinson S.A. (1990) The case study approach. In Advancing in nursing. Nursing Times Research 2, 48±56. Nursing Science Through Research, Volume 2 (Moody L.E. ed.), Woods L.P. (1998) Implementing advanced practice: identifying Sage, Newbury Park, California. the factors that facilitate and inhibit the process. Journal ofMacleod Clark J. & Hockey L. (1989) Further Research for Nursing: Clinical Nursing 7, 265±274. A New Guide for the Enquiring Nurse. Education for Care Yin R.K. (1993) Applications of Case Study Research. Applied Series. Scutari Press, London. Social Research Methods Series, Volume 34. Sage, NewburyMoser G.A. & Kalton G. (1971) Survey Methods in Social Inves- Park. tigation 2nd edn. Heinemann, London. Yin R.K. (1984, 1989, 1994) Case Study Research: Design andParse R.R., Coyne A.B. & Smith M.J. (1985) Nursing Research Ð Methods. Applied Social Research Methods Series, Volume 5. Qualitative Methods. Brady Communications, Maryland. Sage, Thousand Oaks, California.934 Ó 2000 Blackwell Science Ltd, Journal of Advanced Nursing, 31(4), 926±934