Stake Multiple Case Study Analysis
Stake, R. E. (2006). Multiple case study analysis. New York: The Guilford Press.
Summary: This book is a fantastic book for anyone conducting multiple case study research. Stake gives an overview and rationale for analyzing cases and multiple cases. He also provides some practical information to help researchers embarking on this journey.
Preface: The Design of Multicase Projects
- Case study involves minimizing intrusion and observing the ordinariness of the case – that includes its activities and in its everyday context.
- Step By Step
- Studying Phenomena
Chapter 1: Single Cases
- The “case – quintain dilemma” is the tension between the case and the quintain over which one gets the attention.
- Cases need some similarity in order to have multicase research.
- It can include a sample of the whole or it can include all of the cases that exist.
- We use ase study research to understand function, but functions themselves are not cases. However, functions can be quintains.
1.1 Situation and Experience
- “It is an epistemological reason. Qualitative understanding of cases requires experiencing the activity of the case as it occurs in its contexts and in its particular situation. The situation is expected to shape the activity, as well as the experiencing and the interpretation of the activity. In choosing a case, we almost always choose to study its situation (p. 2).”
- “The case has an inside and an outside. Certain components lie within the system, within the boundaries of the case; certain features lie outside. A few of the outside features help define the contexts or environment of the case (p. 3).”
1.2 A Technical View of a Case
- “The case researcher needs to generate a picture of the case and then produce a portrayal of the case for others to see. In certain ways, the case is dynamic. It operates in real time. It acts purposively, encounters obstacles, and often has a strong sense of self. It interacts with other cases, playing different roles, vying and complying. It has stages of life – only one of which may be observed, but the sense of history and future are part of the picture (p. 3).”
1.3 The Quintain
- Cases in a multicase study are categorically bound; that is their common thread.
- “Each mini-case then will be constrained by its representation of or relationship to the program. But if the study is designed as a qualitative multicase study, then the individual cases should be studied to learn about their self-centering, complexity, and situational uniqueness. Thus each case is to be understood in depth, giving little immediate attention to the quintain (p. 6).”
- The researcher chooses the extent to which the focus is on the quintain or the cases. “Of course there is no one right way. Researchers can design a study to give either proportionate or disproportionate attention to the quintain and individual cases (p. 7).”
1.4 The Case-Quintain Dilemma
- The struggle! “Whether everything actually is a part of everything, or whether we have a human capacity for seeing everything as a part of everything, it all becomes more complex as it becomes better known, and it cries out for being still better known. It becomes increasingly worthy of being included in the study (p. 7).”
- Multiple case study research is usually instrumental by nature. Meaning that its focus is to go beyond the case.
- “When the purpose of case study is to go beyond the case, we call it ‘instrumental’ case study. When the main and enduring interest is in the case itself, we call it ‘intrinsic’ case study (Stake, 1998) (p. 8).”
1.5 The Research Questions
- Those deeper individualized research questions he calls issues. “To some extent, sometimes entirely, each case gets organized and studied separately around research questions of its own (p. 9).”
1.6 The Particular and the General
- If issues are important to the quintain, then they need to be a understood more generally.
1.7 The Contexts
- “Each case to be studied is a complex entity located in its own situation. It has its special contexts or backgrounds. Historical context is almost always of interest, but so are cultural and physical contexts. Others that are often of interest are the social, economic, political, ethical, and aesthetic contexts. The program or phenomenon operates in many different situations. One purpose of a multicase study is to illuminate some of these many contexts, especially the problematic ones (p. 12).”
1.8 Making the Individual Case Report
- “Even when the study is well done, the research questions will not be fully answered. Some assertions can be made that partially answer the question, but ways the questions need to be improved will become apparent. And new questions needing to be asked will become apparent (p. 14).”
- “The case study report is a summary of what has been done to try to get answers, what assertions can be made with some confidence, and what more needs to be studied (p. 14).”
Chapter 2: The Multicase Study
- Multiple case studies are so complex that having several minds can be beneficial, but they almost should be done by one person because that person maintains the vision of the quintain. It is difficult to keep that idea in many heads.
2.2 Selecting Cases
- The research director has voice in the case selection.
- Three main criteria for case selection:
- “Is the case relevant to the quintain?
- Do the cases provide diversity across contexts?
- Do the cases provide good opportunities to learn about complexity and contexts? (p. 23).”
- “When cases are selected carefully, the design of a study can incorporate a diversity of contexts (p. 23).”
- Purposeful selection is almost always used.
- “For qualitative fieldwork, we will usually draw a purposive sample of cases, a sample tailored to our study; this will build in a variety and create opportunities for intensive study (p. 24).”
- Additional criteria for case selection: “Balance and variety are important; relevance to the quintain and opportunity to learn are usually of greater importance (p. 26).” The last criteria – opportunity – is the key criteria.
2.3 Activity in its Situation
- “The multicase research team tries to tease out how the situation at each of several different sites influences program activity or the phenomena (p. 29).”
2.4 Data Gathering Across Cases
- “Usually the project director will carry out the cross-site analysis and write the synthesis, getting critiques from the team, people in the field, and (again) their critical friends (p. 31).”
2.5 Triangulation within Cases
- “Each important finding needs to have at least three (often more) confirmations and assurances that key meanings are not being overlooked. Each important interpretation needs assurance that it is supported by the data gathered and not easily misinterpreted by readers of the report (p. 33).”
- “Triangulation is mostly a process of repetitious data gathering and critical review of what is being said (p. 34).”
- “The author needs to repeat key assertions in several ways. He or she needs to give illustrations. He or she will leave some of the work for the readers to do, but should give them the makings of understanding (p. 35).”
Rules about triangulation:
- “If the description is trivial or beyond question, there is little need to triangulate.
- If the description is relevant and debatable, there is some need to triangulate.
- If the data are critical to a main assertion, there is much need to triangulate.
- If the data are evidence for a controversial finding, there is much need to triangulate.
- If a statement is clearly a speaker’s interpretation, there is little need to triangulate the quotation but not its content (p. 36).”
- Member checking is done after the data are processed and written in a report.
- “The qualitative researcher is interested in diversity of perception, even the multiple realities within which people live. Triangulation helps to identify these different realities (p. 38).”
Chapter 3: Cross-Case Analysis
- “The main reason for doing single-case study research often gets mangled in a cross-case analysis. The case researcher has tried to display the unique vitality of each case, noting its particular situation and how the context influences the experience of the program or phenomenon. Many readers look to the cross-case analysis to find what is common across the cases, not what is unique to each. I generally find that researchers doing cross-case analysis are emphasizing the common relationships across cases (p. 39).”
- “Findings from each – somehow combined as Assertions. Some of the important findings from the Cases will be context-bound. Given the load, some oversimplification is likely and perhaps inevitable (p. 41).”
3.2 Reading the Collected Reports
- All case reports of the quintain should be examined closely. Analysis is often best done with one person.
- Prominence is an indicator of relevance. A theme’s prominence reflects the relevance of the case to the quintain.
3.3 Cross-Case Procedure
- “For grand strategy, I think it is desirable for the analyst to set up a ‘case-quintain dialectic’ – a rhetorical, adversarial procedure, wherein attention to the local situations and attention to the program or phenomenon as a whole contend with each other for emphasis (p. 46).”
- He offers a site for downloading sample worksheets. In particular he suggests worksheets 2-6 for this part. www.ed.uiuc.edu/circe/worksheets/worksheet/
- 3.4 Expected Utility of Cases and Ordinariness of Situations
3.5 The Grounds for Assertion
Track 1: Emphasizing Case Findings
- Sorting Findings Strips
- Utility and Prominence of Cases
- Assertions by Bypass
- Tentative Assertions
Track II: Merging Case Findings
- Sorting and Merging Findings
- Sorting and Ranking Findings
- “A lengthy list diminishes the importance of individual Assertions (p. 62).”
- “Each Assertion should have a single focus, an orientation for understanding the Quintain, and evidence to support it (p. 62).”
Track III: Providing Factors for Analysis
- “A Theme or a Finding, as used here, is a central idea having importance related to its situation. It is at least somewhat context-bound, more local than universal. A Factor, as used here, is a widely found, sometimes influential variable of interest well beyond its situation (p. 64).”
- Identify factors for each case.
- Merging Factors
- Sorting and Ranking Factors
- Tentative Assertions
- Track 1: Emphasizing Case Findings
- 3.6 Cross-Case Assertions
3.7 Triangulation Across Cases
- “Triangulation also requires going further afield, checking with people who know some of the Quintain or related activity (p. 77).”
- “It means being redundant and skeptical in seeing, hearing, coding, analyzing, and writing. It benefits from discussion with both critical insiders and outsiders. The exchanges should be both routinized and spontaneous. Mature is the researcher who rejoices in finding a big mistake (p. 77).”
Chapter 4: The Report
- Executive summaries are optional. They are the short, abbreviated version of the report. They are located either at the beginning or at the end, or they can be published separately.
- 4.1 Planning the Multicase Report
4.2 Comparing Cases
- “Comparison is a search for similarity and difference in cases. There is comparison in all inquiry and all discourse, but comparison has only a minor role to play in individual case studies (p. 82).”
- “Comparative studies – whether of quantitative or qualitative design – seek similarities and differences among cases on a relatively few specified attributes (p. 82).”
- “The purpose of those studies is to make some grand comparison rather than to increase understanding of individual cases (p. 82).”
- “Multicase study is not a design for comparing cases. The cases studied are a selected group of instances chosen for better understanding of the quintain. Most case researchers report each case as case, knowing attributes for comparison (p. 83).”
- “Comparison is a competitor to probing study of a case. It is a grand research strategy, a powerful conceptualization, usually fixing attention on one or a few variables. In so doing, it obscures the situationality and complex interaction of case knowledge (p. 83).”
- “We conceptualize the case in various ways to facilitate learning about the quintain. The quintain is something that functions, that operates, that has life (p. 83).”
- It is important to recognize your bias. It adds credibility to your research if you recognize it rather than claim objectivity.
- “Most favoritism in research is to be found in what is not said – in underplaying a negative side of the picture. The reader wans a full account, as well as accountability, but the report writer cannot help being selective, for there is more to be said than can be said. What she or she selects to say will be based on the research question and the utility of alternative explanations; however, it will also be based partly on what the report writer cares about, partly on what is deemed important to know, and partly on what effects the findings might have. There is no value-free science in this world (p. 85).”
- “Seldom do we have a large conflict of interest, but we often have a confluence of interest. We hope to find the program working. We are disposed to see evidence of success (p.86).”
- “It is an ethical responsibility for us as case researchers to identify affiliations and ideological commitments that might influence our interpretations – not only for the contracting parties but for the readers of reports, and, of course, for ourselves. But there is no way for us as evaluators to identify all relevant predispositions, or even to know them (p. 87).”
- “In the complex representation of program operation and accomplishment, there is no single reality we can capture (p. 87).”
- “Were we to agree completely on what we see, we would presume that we are seeing correctly – and often we are not (p. 87).”
- “Researchers should be encouraged to ‘have a life’ and to ‘have a dream,’ so their interpretations are enriched by personal experience (p. 87).”
- “Comprehensive, idiosyncratic, irreproducible interpretations are a contribution to understanding and action (p. 87).”
- “The study of human activity loses too much if it reports primarily what is common among the several, and universal across the many (p. 88).”
- “It seems more likely to me that the act of generalizing is deeply set in the human repertoire and that it will continue to operate largely without protocol (pp. 89-90).”
Five levels of learning (Hubert & Stewart Dreyfus, 1986)
- “Only at the lowest leve, the novice level, is there concentration on epistemic generalizations, rule-bound and context-free. As readers advance up the steps, it is expected that their learning will increasingly involve situational elements, virtues, goals, plans, and the experience of the readers themselves, through the highest level, as described by Flyvbjerg (2001): ‘...experts’ behavior is intuitive, holistic, and synchronic, understood in the way that a given situation releases a picture of problem, goal, plan, decision and action in one instant and with no division into phases’ (p.21) (p. 90).”
- “Competent performers, experts, and virtuosos need case-based contextual understanding to add to their own direct and vicarious experience (p. 90).”
- “Because the reader knows the situations to which the assertions might aply, the responsibility of making generalizations should be more the reader’s than the writer’s (p. 90).”
Dreyfus, H. L., & Dreyfus, S. E. (1986). Mind over machine: The power of human intuition and expertise in the era of the computer. New York: Free Press.
Flyvbjerg, B. (2001). Making social science matter (S. Sampson, Trans.). Cambridge, UK: Cambridge University Press.