Volume 9, No. 1, Art. 34 – January 2008
Cultivating the Under-Mined: Cross-Case Analysis as Knowledge Mobilization
Samia Khan & Robert VanWynsberghe
Abstract: Despite a plethora of case studies in the social sciences, it is the authors' opinion that case studies remain relatively under-mined sources of expertise. Cross-case analysis is a research method that can mobilize knowledge from individual case studies. The authors propose that mobilization of case knowledge occurs when researchers accumulate case knowledge, compare and contrast cases, and in doing so, produce new knowledge. In this article, the authors present theories of how people can learn from sets of cases. Second, existing techniques for cross-case analysis are discussed. Third, considerations that enable researchers to engage in cross-case analysis are suggested. Finally, the authors introduce a novel online database: the Foresee (4C) database. The purpose of the database is to mobilize case knowledge by helping researchers perform cross-case analysis and by creating an online research community that facilitates dialogue and the mobilization of case knowledge. The design of the 4C database is informed by theories of how people learn from case studies and cross-case analysis techniques. We present evidence from case study research that use of the 4C database helps to mobilize previously dormant case study knowledge to foster greater expertise.
Key words: case study, cross-case analysis, computer-assisted analysis, knowledge mobilization, researcher, database
Table of Contents
1. Cross-Case Analysis: Introducing the Foresee Database
2. Literature Review
2.1 Learning from and with cases
3. Review of Several Cross-Case Analysis Approaches and Techniques
3.1 Variable-oriented approaches to cross-case analysis
3.2 Case-oriented approaches to cross-case analysis
4. Several Issues for the Case Study Researcher Engaged in Cross-Case Analysis
5. The Foresee Database Project
5.1 Design principles of the 4C database
5.2 Affordances of the 4C database
6. How 4C is Different from Computer-Assisted Qualitative Data Analysis Tools and Online Repositories
1. Cross-Case Analysis: Introducing the Foresee Database
Cross-case analysis is a research method that facilitates the comparison of commonalities and difference in the events, activities, and processes that are the units of analyses in case studies.1) Despite a plethora of case studies in the social science literature and archived on web sites, few are adequately mined again by researchers or are known to inform practitioners or policy at a broader level. The expertise embedded within the vast number of case studies in the fields of education and sociology remains relatively dormant. In this paper, we propose cross-case analysis as a mechanism for mining existing case studies so that knowledge from cases can be put into service for broader purposes. To mobilize case knowledge across subject domains and across communities, we introduce the creation of a novel database. The database represents a workspace to perform cross-case analysis and a workspace where expertise can flow in systematic and unexpected ways through the representation, transfer and mobilization of case studies. 
Engaging in cross-case analysis extends the investigator's expertise beyond the single case. It provokes the researcher's imagination, prompts new questions, reveals new dimensions, produces alternatives, generates models, and constructs ideals and utopias (STRETTON, 1969). Cross-case analysis enables case study researchers to delineate the combination of factors that may have contributed to the outcomes of the case, seek or construct an explanation as to why one case is different or the same as others, make sense of puzzling or unique findings, or further articulate the concepts, hypotheses, or theories discovered or constructed from the original case. Cross-case analysis enhances researchers' capacities to understand how relationships may exist among discrete cases, accumulate knowledge from the original case, refine and develop concepts (RAGIN, 1997), and build or test theory (ECKSTEIN, 2002). Furthermore, cross-case analysis allows the researcher to compare cases from one or more settings, communities, or groups. This provides opportunities to learn from different cases and gather critical evidence to modify policy. 
2. Literature Review
2.1 Learning from and with cases
Assuming that the researcher's learning process parallels the ways in which individuals develop expertise, the authors will, in this section, examine four learning theories that support the notion that cross-case analysis is a method for mobilizing case study knowledge: AUSUBEL, NOVAK, and HANESIAN's (1978) cognitive theory of meaningful learning, KOLODNER's (1993) case-based reasoning, FLYVBJERG's (2001) notions of developing expertise from cases, and DONMOYER's (1990) theory of vicarious learning from case knowledge. These learning theories support the notion that researchers develop expertise from cases, and they conceptualize the processes through which this expertise is cultivated. 
KOLODNER and AUSUBEL et al.'s theories primarily emphasize human learning as a cognitive and experiential undertaking and do so while pointing to cognitive processes that are similar to those required for engagement in cross-case analysis. FLYVBJERG and DONMOYER stress the importance of learning from one case to another, arguably emphasizing a form of case-based reasoning, that is, the process of reasoning about the similarities and differences across diverse cases, as key to the development of expertise. Cumulatively, these theories appear to hypothesize that cognition involves cases of experiences and that learning from cases is accomplished by cross-case analysis. The authors extend these hypotheses on learning and suggest that case study researchers can develop expertise through learning from and comparing cases. When the case study researcher makes this comparison public, case knowledge becomes mobilized. 
AUSUBEL et al.'s cognitive theory of meaningful learning. AUSUBEL et al.'s cognitive theory of learning (1978) emphasizes that people learn meaningfully by developing cross-connections between related concepts. This allows them to engage in inferential and analogical reasoning. These cross-connections can take the forms of either cognitive assimilation or accommodation of concepts. Assimilation of concepts increases knowledge while preserving the cognitive structure, whereas accommodation modifies existing knowledge to account for the new experience. AUSUBEL et al.'s conception of cross-connections can be applied to cross-case analysis: relating one case to another, building cross-connections between cases, preserving the essence of the original case knowledge while changing the character of the current case, can accumulate and produce new knowledge. 
Case-based reasoning. KOLODNER (1993) extends AUSUBEL et al.'s theory of cross-connections to memory. KOLODNER's case-based reasoning (CBR) explains learning as a cognitive process in which the individual interprets a new situation in terms of its relevance to a previous case. KOLODNER further theorizes that the lessons learned from the combination of previous and new cases are encoded and indexed in memory as abstract generalizations. This process of memory storage and retrieval implies that a person will be able to evaluate possible solutions through an indexing process that discriminates among cases. At memory retrieval time, when the person is engaged in a new situation, a memory probe searches through the index for cases that are similar to the new one. KOLODNER describes this probing as a creative process and suggests that the more astute the person is at conceptualizing a situation, the more likely he or she is to find relevant knowledge about previously learned, memorable cases (KOLODNER et al., 2003; SCHANK & BERMAN, 2002). This ability to enlighten oneself develops over time through case-based reasoning. It appears that analyses of a variety of cases are necessary to learn well. 
FLYVBJERG's notion of expertise. Drawing heavily upon DREYFUS and DREYFUS' (1988) work on skill acquisition in experts, FLYVBJERG (2001) extends the notion of case-based cognition to experts' ways of reasoning. Experts think quickly, intuitively, holistically, interpretive, and visually. As DREYFUS and DREYFUS explain, "bodily involvement, speed, and an intimate knowledge of concrete cases in the form of good examples are a prerequisite for true expertise" (1988, p.15). According to FLYVBJERG (2001), expertise or virtuosity is intimate knowledge of concrete cases. This intimate knowledge is gained through reflection upon thousands of cases directly, holistically, and intuitively. Case studies are the domain of expertise, which is neither guesswork nor a conscious analytical division of situations into parts and rules but rather, the recognition, interpretation and discrimination of cases and new situations. 
DONMOYER's theory of learning from cases. DONMOYER's (1990) conception of generalization reveals how an expert might simultaneously access numerous cases to make a comparison among these cases. DONMOYER suggests that new understanding takes root when an individual begins to generalize across cases that were derived or constructed from different contexts. According to DONMOYER, generalization across cases is not a formal act of generating working hypotheses that are to be tested in new cases. Instead, he views learning from cases as a meaning-making endeavor in which cross-case analysis is essential. DONMOYER suggests that learning from case knowledge can be better characterized as assimilating, accommodating, and integrating case knowledge from previously learned cases. His own example of becoming a better teacher over the years exemplifies this kind of learning. DONMOYER suggests that his development as a teacher was not an effort to consciously test hypotheses in the different schools he taught at but rather, an attempt to learn from individual cases of teaching that he and others experienced over the years. 
In sum, learning through cross-case analysis empowers the learner to access the experience of others and thus, to extend their personal experience. These new connections made across cases produce new knowledge and augment existing knowledge and experience. While learning theorists invoke different cognitive structures and processes to explain cross-case analysis, there are the following commonalities:
cases represent rich holistic examples of experiences;
cases are comparable in relation to patterns of similarities and differences;
memorable cases are accessed through memory;
comparisons among cases can construct and yield meaningful linkages, and
cognitive cross-case analyses are a useful way to produce analogies, make inferences, and develop conditional generalizations for the individual. 
Similarly, for researchers who develop expertise through cross-case analysis:
cases represent rich examples of cases they have learned or know about;
the cases are deemed comparable in relation to patterns of similarities and differences;
the cases are accessible;
meaningful connections between cases can be made explicit by the researcher, and
the researcher can produce and share new knowledge through cross-case analysis. 
3. Review of Several Cross-Case Analysis Approaches and Techniques
There are several well-known cross-case analysis approaches and techniques available to the case study researcher. RAGIN (1997) for example delineates between variable and case-oriented research as two approaches to cross-case comparisons. In variable-oriented research, variables take center stage; that is, the outcome observed in the cases varies across observations and causes appear to compete with one another. The cases are selected in advance with an eye toward randomness or the degree to which they represent the general population. The goal is to explain why the cases vary. Variable-oriented approaches to cross-case analysis are a challenge to conduct because fair comparisons are difficult to achieve and the multitude of factors that are associated with social phenomena are often too numerous to disentangle. In case-oriented research, commonalities across multiple instances of a phenomenon may contribute to conditional generalizations (MILES & HUBERMAN, 1994). The researcher can thus demonstrate that the outcomes in the cases selected are in fact enough alike to be treated as instances of the same thing. The central question of interest to the case-oriented researcher is in what ways the cases are alike. Therefore, special emphasis is given to the case itself instead of on variables across cases. Examples that illustrate the complexity of this approach are case studies that focus on the role of violence in schoolyard bullying and national warfare. Both case studies are about violence, but the scale and scope of the violence in the respective contexts are likely incommensurable and difficult to compare or contrast. Still one is immediately attracted to the prospect of crossover and mutual illumination. Thus, in a variable-oriented approach, factors known to be involved in violence, such as resources and perceptions of vulnerabilities, could be used to evaluate both cases independently before comparing factors between a case of schoolyard bullying and a case of war-mongering states to explain and predict violent behavior. On the other hand, in a case-oriented approach, one could conceivably compare two cases of "swarming" in schools with two cases of "swarming"-like behavior in war-mongering nation states to search for or construct similar processes that appear to lead to violent behaviors. 
In this section, several variable-oriented and case-oriented approaches that are applicable to cross-case analysis are discussed by drawing upon the more extensive reviews of these approaches by GEORGE and BENNETT (2005) and MILES and HUBERMAN (1994). For variable-oriented cross-case analyses, several well-known research techniques include: MILLS' Methods, Case Survey, and Before-After research design. For case-oriented cross-case analyses, several well-known techniques include:Most different design, Typologies, Multicase Methods, and Process-tracing. 
3.1 Variable-oriented approaches to cross-case analysis
MILLS' methods. MILLS' (1843) famous comparative system of logic involves a method of agreement and a method of difference as two potential analytic techniques for comparing cases. The method of agreement identifies a similarity in the independent variable associated with a common outcome in two or more cases. The method of difference identifies independent variables associated with different outcomes. MILLS' methods require eliminating candidate causes for the outcome. In the method of difference for example, the condition that is not present in both cases where the outcomes were different, could be considered a possible causal factor in the variance between outcomes. The factor(s) that survive this systematic process of elimination are inferentially connected to the outcomes. MILLS himself noted some serious obstacles to his comparative system of logic, especially when applied to studies in social science. Social phenomena are often rooted in a complex web of causes, which are difficult if not impossible to isolate as deterministic. That leaves the researcher open to the danger of false positives. GEORGE and BENNETT (2005), who conducted an extensive review of comparative techniques, suggest that MILLS' methods can work if the causal relationship involves only one factor that is either necessary or sufficient for a specified outcome, if all causally relevant variables are identified prior to the analysis, and if cases that represent the full range of possible causal paths are available for study. GEORGE and BENNETT contend that there are few theories in the social sphere that are strong enough to support general claims of necessity or sufficiency for single variables (2005, p.157). 
Case survey method. The case survey method (YIN, 1994, 2003) involves gathering evidence from a large set of cases (e.g., 250) so that statistical analyses can be performed on the variables pertinent to all the cases. Case surveys are challenging to carry out because researchers seldom study so many cases and they rarely find perfectly comparable cases. Furthermore, increasing the number of cases often means making assumptions of homogeneity that are simply unjustifiable. An example of a case survey method is a study of the cultural antecedents of procrastination wherein large numbers of individuals from all over the world would be analyzed as separate case studies within a case survey method. 
Before-after design. Another method for cross-case analysis is the before-after design. The before-after design offers some level of control by dividing one case into two sub-cases. Some event or critical juncture in a natural setting creates the conditions for a before and after investigation. One of the assumptions on which the before-after design is based, is that only one variable changes, dividing the longitudinal case neatly in two. Determining the change in a variable is difficult unless a careful analysis of all factors involved in the case is conducted over the same period of time. An example of this type of cross-case analysis is the study of online communication in a science course where patterns of communication are analyzed before and after a major course assignment. 
3.2 Case-oriented approaches to cross-case analysis
Most different design. Some social scientists have abandoned the quest for controlled comparison in favor of PRZEWORKSI and TEUNE's (1982) most different design. A most-different research design deliberately seeks to compare cases that differ as much as possible in order to find similar processes or outcomes in diverse sets of cases. This case-oriented approach emphasizes diversity in the selection of cases (GEORGE & BENNETT, 2005, p.165). The power of the most different design lies in its ability to extend the lessons learned in single cases to inform another case and to uncover similar processes in unexpected contexts. Cross-case comparisons of school principals and CEO's of large auto companies would be one example of a most different design. While schools and auto companies do not, on the surface, appear to be meaningfully comparable, it may be fruitful to compare the work habits of CEOs who produce cars and their organization techniques with those of school principals who view schools as organizations with students as products. 
Typologies. Cross-case comparison can support the creation of clusters or families of phenomena. Sets of cases are categorized into clusters of groups that share certain patterns or configurations. Sometimes the clusters can be ordered or sorted along several dimensions. For example, DENZIN (1989) suggests deconstructing prior conceptions of a particular phenomenon and then collecting multiple cases and bracketing them for essential elements and components across cases. The essential elements are then rebuilt into an ordered whole (e.g., construction of the alcoholic self) and put back into the social context. In another typologizing effort, the pathway to the outcome is inspected and compared among a set of cases. Like process tracing below, the same outcome is theorized according to different pathways. For example, science education reforms that better integrate technology would be considered a sub-class of the general category of educational reforms. Typologies share a specified combination of factors, but these are not necessarily causal, mutually exclusive or exhaustive. GEORGE and BENNETT (2005) argue that a typological regularity can be sought through previously unexamined causal paths or a building block approach. Typologizing supports the construction of theories by identifying the sub-classes of a major phenomenon. 
Multicase methods. This method has recently been introduced by STAKE (2006) and focuses on the quintain, which is a common focus (organization, campaign, problem) for a set of case studies. The quintain, for example, might be mega-events, like the Olympic Games or a school district that wishes to incorporate technology at all of its sites. The quintain is comprised of case studies that have both common and unique issues. The common issues address important and complex problems about which disagreement exists. The impacts of mega-events on host regions might be elicited from case studies done at different Olympic sites. Common research questions (e.g., what is the economic impact of enhanced international image of the host region?) tie together all of the case studies. A cross-case analysis of these cases facilitates a greater understanding of the quintain (again mega-events). According to STAKE, after cross-case analysis, researchers can make assertions about the quintain. These assertions are then applied to the individual case studies to determine the extent to which the case studies reflect the quintain. The degree of congruity or disparity speaks to the uniformity of the quintain and the power of cross-case analysis (STAKE, 2006). 
Process-tracing. In this method, the progression of events that may have led to an outcome in a single case is traced (GEORGE & BENNETT, 2005). Process-tracing forces the researcher to consider alternative paths through which the outcome could have occurred, and it offers the possibility of mapping out one or more potential causal paths that are consistent with the outcome. Cross-case analysis allows the researcher to develop a typological theory by charting the repertoire of causal paths that reveal given outcomes as well as the conditions under which they occur. In process-tracing, all the intervening steps within a case must be predicted by a hypothesis or else the hypothesis is amended. Process-tracing generally takes the form of a detailed narrative in which the unfolding of a story is theoretically oriented. 
In addition to variable and case-oriented approaches, some analytic techniques are worth mentioning, such as stacking, building truth tables, and constructing narrative models. MILES and HUBERMAN (1994) suggest that these techniques are a mixture of variable and case-oriented approaches. These mixed techniques are mentioned here because any of the approaches discussed above can also utilize these techniques. The authors refer to these three techniques as data display and analysis techniques because they help to visualize sets of cases, and they bring case relationships to the surface in ways that invites and facilitates comparison. In the stacking comparable cases technique, a series of cases are displayed in a meta-matrix by fields of interest (MILES & HUBERMAN, 1994). Each case is condensed in a form that permits a systematic visualization and comparison of all the cases at once. 
The "qualitative comparative analysis" or QCA technique, developed by RAGIN (1993), allows for the analysis of certain aspects of the case without obscuring it. QCA is based on Boolean analysis where relationships among the cases are built by the use of conjunctions (and, or, not). This approach to synthesizing cases involves a technique that arranges cases in a "truth table" by variable in order to study common causes or outcomes. Conjunctions are utilized to locate relationships within the truth tables. 
The third technique discussed here was developed by GOLDSTONE (1997). He suggests that narratives are the keys to cross-case analysis. Narratives can preserve the essence of the case during cross-case analysis. It could also be argued that constructing narrative models helps to facilitate comparison by encapsulating the case as a storyline. 
In summary, there are multiple research techniques to conduct cross-case analyses. Variable-oriented approaches to cross-case comparison tend to pay greater attention to the variables across cases rather than the case itself. Variables are compared across cases in order to delineate pathways that may have led to particular outcomes. These pathways are often represented as probabilistic relationships among variables. The complexity and context of individual cases is not at the center of variable-oriented approaches. Case-oriented approaches, on the other hand, such as creating typologies, are more particularistic. This approach can show how a story unfolded in different cases, how researchers can make sense of the original case, or suggest new typologies, classes or families of a social phenomenon. Visualization techniques, such as stacking cases, can be utilized by either approach to invite and show comparison. Advantages of cross-case analysis that emerge from these techniques are:
the case content is made available to the researcher in an easily accessible form;
cases are clustered and represented in a visual display to facilitate comparison by the researcher and by others;
cases are compared in a method that either centers on the case or on the variables, depending on the goal of the researcher, and
findings of the case and the cross-case comparison are shared with others. 
4. Several Issues for the Case Study Researcher Engaged in Cross-Case Analysis
While there are a number of scholars who suggest that cross-case analysis can enhance a researcher's contribution to theory and method (cf. ECKSTEIN, 2002; RUESCHEMEYER, 2003), there are others who are less optimistic about comparing cases. Counter- arguments stem from an epistemological conviction that case knowledge emerges from a dense descriptive study of the particularities of a case. Comparison, the counter-argument goes, obscures case knowledge including knowledge not germane to the comparison (PEATTIE, 2001). Indeed, there are long-standing tensions between deeply contextualized and particularistic case knowledge and multiple case study research (FOREMAN, 1948; ALLPORT, 1962; MOLENAAR, 2004). To begin to reduce the tensions among idiographic and nomothetic research traditions, case study researchers must recall their original goals for the cross-case analysis. As mentioned previously, goals for engaging in a cross-case analysis can include, for example: further illustration, concept and hypothesis development, prediction, and empathic portrayals. 
Researchers' goals notwithstanding, the cross-case analyst will also be confronted with questions about the generalizability of the conclusions emerging from the analysis and the ability of the researcher to justify any comparison beyond the set of cases studied. As suggested by KHAN (2007), positivist notions of generalizability have been largely abandoned or modified in social science and case study scholarship (SCHOFIELD, 1990; DONMOYER, 1990; GUBA & LINCOLN; 1981). Generalizations have been recognized as contextual, having half-lives (CRONBACH, 1975) that require updating (even in experimental research). It is far easier, and more epistemologically sound, simply to give up on the idea of generalization; if generalizations are accepted, they should be as indeterminate, relative, and time and context-bound (LINCOLN & GUBA, 2000, p. 32). 
Instead of positivist notions of generalizability, new concepts have emerged to extend and amplify the impact of a single case beyond the case itself (YIN, 2003; BECKER, 1990; SMALING, 2003). For example, GOETZ and LECOMPTE (1984) recognized that the findings from case studies cannot be generalized in a probabilistic sense, but that findings from case studies may still be relevant to other contexts. "Comparability" is a concept they proposed to address the issue of generalizability from a single case or cross-case analysis. Comparability is the degree to which the parts of a study are sufficiently well described and defined that other researchers can use the results of the study as a basis for comparison. "Translatability" is a similar concept but refers to a clear description of one's theoretical stance and research techniques. 
While it is not the purpose of this paper to elaborate on idiographic and nomothetic debates or delineate all classes of generalization for the cross-case analyst, we recommend that interested case study researchers explore idiographic generalization (ALLPORT, 1962), analogical generalization (SMALING, 2003), analytic generalization (YIN, 2003), and naturalistic generalization (STAKE, 2005) as alternative forms of generalization that can be invoked to rationalize cross-case analyses. In addition to developing a stance on generalizability, there will be at least three accompanying, practical concerns for case study researchers to attend to before embarking upon their cross-case analysis:
preserving the essence of the cases,
reducing or stripping the case of context, and
selecting appropriate cases to compare 
Preserving the uniqueness of cases. SILVERSTEIN (1988) states that cross-case analysis must reconcile the preservation of the uniqueness of the case while attempting to analyze the case across other cases. The concern is that the complexity of meaning (from each case) might get lost when the content is simplified in order to make comparison possible (TESCH, 1990). While comparing multiple case studies holds great potential to inform theory, RUESCHEMEYER cautions that the researcher must "increase the number of cross-case comparisons without losing the advantage of close familiarity with the complexity of cases" (2003, p.323). The authors' stance, and the stance of others (STAKE, 2006), is that it is possible to learn from both the uniqueness and commonality of a case. By providing ample contextualized details of the cases and findings of cross-case analysis, a researcher can conceivably preserve the uniqueness of a case and convey the value of their engagement with a cross-case analysis. 
Contextual stripping. In cross-case analysis, the contextualized origins of each case are in danger of being lost as cases are compared, especially if a variable-oriented approach is adopted. However, according to AYRES, KAVANAUGH, and KNAFL (2003), losing some contextual detail may be consistent with the goals of cross-case comparison, which is to identify themes across cases. TESCH (1990) described cross-case comparison as essentially a "decontextualization and recontextualization" of cases. The process is as follows: case study data are separated into units of meaning (decontextualized because they are separated from the individual cases) and then recontextualized as they are later integrated and clustered into themes. The themes, which are a reduced data set, can help to explore relationships. The origin of each unit of meaning is less important than its membership in a group of like units. AYRES et al. (2003) referred to this approach as "moving between across- and within-case comparisons" (2003, p.875). Such a cross-case synthesis, according to these authors, achieves its authenticity in the immersion within individual cases. 
In a similar approach to cross-case comparison, KNAFL (as cited in AYRES et al., 2003) reduced the contextual stripping in a cross-case analysis of family management styles during illness. KNAFL first identified general themes that shaped the experience of families dealing with illness (searched for commonalities across accounts). Secondly, she delineated variation within the themes (across individual family members), and thirdly, created a "thematic profile" for each family member and family unit (within case analysis). Finally, she offered a differentiation of family management styles (across families case analysis). Themes such as being a burden ended up playing a role in illness management style. Sub-themes emerged when the accounts of individual family members were compared with that of the family as a unit. Within-case comparisons were represented as narrative case summaries and cross-case comparisons were displayed as a grid using a database manager to identify clusters of families with similar configurations. In both AYRE’s and KNAFL's approaches to cross-case analysis, attempts to preserve the uniqueness and authenticity of the case were successful. 
Selection of cases. Generally, in variable-oriented approaches, the number of cases to compare should be high, whereas in the case-oriented approach, the number of cases to compare is generally low (but not less than two). In both instances, the researcher is advised to search for comparable cases until they are satisfied that the search is no longer yielding new insights or until theoretical saturation has been achieved. Variable-oriented researchers support comparison of cases that are fairly similar in order to achieve a level of control that can foster predictability and idiographic or nomothetic generalizations. Case-oriented research can support the comparison of cases that are ostensibly very different. Earlier, the example of a cross-case comparison study of school principals and CEO's was introduced. At face value, such a comparison might be challenging since the contexts and roles are so different. However, a principal focused on cultivating citizenship and academic achievement may have something in common with a CEO who runs a car manufacturing plant and is focused on production of vehicles and car performance. Both are attempting to motivate individuals to produce a set of outcomes in a certain time span. It is possible to imagine both case studies featuring interviews of a sampling of students and workers discussing their perceptions of accomplishment and alienation in regards to their duties and responsibilities. The selections of cases, and their corresponding units of analyses, are an important methodological consideration in case study comparisons and should be related to the overall goals of the case study researcher. 
5. The Foresee Database Project
With the above techniques and considerations regarding cross-case analysis in mind, an online database, known as the Foresee or 4C (Cross-Case Comparisons and Contrasts) was developed. The Foresee database utilizes Web 2.0 capacities to bring together case study researchers to perform cross-case analysis, and thus, to mobilize new knowledge. The long-term objectives of the database are, first, to promote cross-case analysis as a research method that facilitates the comparison of commonalities and difference in cases and second, to establish an online research community that facilitates dialogue and the mobilization of case study knowledge. In this section, the design principles and features of the 4C database are described. 
5.1 Design principles of the 4C database
The aforementioned assumptions regarding how researchers develop expertise informed the creation of four design principles to guide the development of the database. The four design principles are:
analyzing cases from different contexts can build common ground between case study researchers from multiple disciplines and diverse backgrounds;
cross-case analysis involves a set of cases that are indexed, accessible, and can be probed visually and conceptually by the researcher;
cross-case analysis can be facilitated by constructing meaningful linkages and relationships, and
in a cross-case analysis, researchers should attempt to preserve the richness and uniqueness of the case. 
These four principles were incorporated in the design of the database. The first and second principles were incorporated by applying the technique of stacking comparable cases. This means that one case is condensed and placed above and below another case or cases in a "meta-matrix" view, where cases are visualized in a table according to set fields. The matrix view offers a first pass visual comparison of cases. The meta-matrix also supports hyperlinking to uncondensed versions of the case to preserve the case in its original form. 
The third design principle incorporates RAGIN's qualitative comparative analysis. This method offers an attractive strategy for using conjunctions, and, or, but, which makes it possible to include some case studies and exclude others. For example, case studies that address both "Education" and "Chemistry" can be selected from the database. The third principle also dictates the use of "tags". Tags are personal, adaptable, and descriptive terms that can be applied to a body of information as metadata (CAMERON, 2004; HAMMOND et al., 2005; MATHES, 2004; SACCO, 2004). The ability to tag means a case study researcher could conceivably create a tag (e.g., "media") and apply it to his or her case study data on, for example, public anti-smoking advertisements. Another researcher could tag the same case with the tag "social marketing." In this way, one researcher could gain access to all the researchers who employed the same tag "media." Another researcher could determine that their case study data contains similar parameters and tag this information as "media." Thus, tagging can facilitate cross-case comparisons of media campaigns aimed at reducing smoking or media-based health promotion campaigns. 
The fourth design principle draws on PRZEWORKSI and TEUNE's (1982) most different research design, which argues for comparing diverse sets of cases because these could generate unforeseen discoveries. To promote this discovery, the authors opened the database to the possibility of researcher's building personal libraries of cases. In addition, researchers are also required to submit cases. The possibility of building a collective as well as personal library builds capacity by offering cases from many fields of endeavor. There are pragmatic and theoretical reasons for being able to do both, which will be discussed in the next section on the affordances of the 4C database. 
5.2 Affordances of the 4C database
Case study researchers can access the 4C database upon registration at http://www.foresee-database.com/. The database is currently housed on a university server. Figure 1 shows the splash page each case study researcher encounters once logging in and becoming members of the system.
Figure 1: 4C splash page 
Firstly, the 4C database records seven aspects of case studies that are submitted: title, focus of study, purpose, research tools, what was learned, related studies, and tags. These seven aspects, or categories, are based on the outcomes of a user study with case study researchers in 2004, and establish common ground among the 4C collective. The case categories are also congruent with most primary journal publishing requirements. Using our example, the "media" tag fits under the tag case category where it can be accessed and analyzed by other researchers much like a keyword. 
4C members can view the collection of submitted and archived case studies as a "list" or as a "meta- matrix" view. Clustering the cases in a central visual display affords what MILES and HUBERMAN (1994) call the "first deep dive" into cross-case analysis; that is, researchers can scroll through the meta-matrix, look across rows or down columns and perform a squint analysis. This gives 4C members the opportunity to scan potential cases for comparison. 
Secondly, 4C members can search the database and select candidate cases for comparison by using the search functions and conducting their search by title, author, content, authors' name, or researcher recommendation. Thirdly, once cases have been selected for comparison, 4C members have access to two methods for cross-case analysis that build relationships among cases. A set of comparison tools allows members to use Boolean terms and code multiple cases with tags. 
Finally, the 4C database helps to enable the publication of cross-case research by offering a multi-way dialogue forum among prospective researchers as well as the public annotations of case studies. A set of screenshots are included to illustrate these affordances. 
Collective Capacity. As Figure 2 depicts, 4C members contribute cases to a collective case study archive. The cases are indexed chronologically as well as by tags. Researchers can gain access to submitted cases and works in progress from a wide range of disciplines, and the public archiving of researchers' cases, tags and researchers' notes facilitates greater learning from cases. The contact information of the researchers who have submitted their case studies to the 4C database is available for other researchers, which enables researchers to further discuss cases and explore research connections.
Figure 2: Case archive in list view 
Personalization of database. Researchers can build their own personal library of cases suitable for work on contained research projects (see Figure 3). Researchers can also construct personal notes on each case submitted that are not viewable to the community. Researchers are able to create their own tags for cases that are different or the same as the submitter's tags.
Figure 3: A 4C member's personal library of cases 
Visual display. 4C database offers a visual display to view the studies as a "meta- matrix" where each study's text is structured and indexed into separate field or case categories. Figure 4 shows how a visual comparison is supported within a meta-matrix view of 4C case studies.
Figure 4: A meta-matrix view, see the PDF file for an improved version 
Advanced search and select tools. 4C's option to compare and contrast the same case categories between different studies with the use of Boolean search terms allows a researcher to find patterns across the database. Figure 5 outlines all the selection choices, and Figure 6 shows how Boolean search terms can be applied to compare cases.
Figure 5: Selection choices
Figure 6: Boolean search 
Conceptual and Conjunctive Relationships. 4C's use of tags helps researchers make comparisons. Tags provide not only a way of locating and comparing cases, but user-driven naming of relationships via tags also increases the flexibility of typical databases. The researcher can use terms to link various cases and search case studies by these terms (GRUDIN, 1994; GUERRERO & FULLER, 2001, PAHLEVI & KITAGAWA, 2003; SCHACHTER, N.D.; STAR, 1998; YEE, SWEARINGEN, LI & HEARST, 2003). The database enables the researcher to view, navigate, and subscribe to case content by researchers' tag(s). In addition, the database also makes it possible to see all the tags that have been applied by all the researchers to the case study. One researcher can compare their tags to another researcher's to learn about the kinds of terms that are applied to certain information. Thus, using the 4C database, researchers can create a personal view of all indexed (i.e., tagged) content, attach personalized tags attached to any indexed item and view, navigate, and subscribe to indexed content by researcher, tag(s) or any combination of these. 
As Figure 7 illustrates, the researcher is also able to see all tags that were applied by researchers of the case study. Tags for a case study are automatically available to the researcher. Furthermore, the researcher can read through tag lists and find other relevant case studies. Finally, Figure 7 shows related tags that are all the tags that include literacy "and" another tag term. Tagging has the potential to develop meaningful links across cases: the "cases" are indexed when they are stored in memory (or entered into the database). However, accessing this knowledge could also reflect the conditions under which the data are retrieved. The previous experience or case could be reframed in a way that was similar to the current one and retrieved as it was re-conceptualized or "tagged".
Figure 7: The "Literacy" tag "and" other tags 
Tutorials and support. The 4C database offers tutorials on how to use the database to conduct cross-case analysis. Although researchers can perform cross-case analysis in multiple ways from every page on the database, the database scaffolds the process of cross-case analysis for researchers by:
suggesting a trajectory involving: clustering cases, performing a squint analysis, selecting comparable cases, comparing cases, and publishing;
providing icons and a site map to ease navigation through the database;
offering a frequently asked questions page, which provides definitions of terms, and
including contacts to site administrators. 
6. How 4C is Different from Computer-Assisted Qualitative Data Analysis Tools and Online Repositories
The 4C database is different from computer-assisted qualitative data analysis software (CAQDAS) and database repositories such as libraries. Researchers often utilize CAQDAS software to code their data, construct categories and create themes. The result of analyzing data using CAQDAS is coded material that is often organized by an individual computer. Cross-case analysis, if conducted, is often done by hand after the data have been coded and contained in the computer. With the 4C database, however, the researcher can combine numerous case studies on any topic of interest (e.g. science education, urban sustainability). 4C members can utilize the database's distributed functionality to perform cross-case analysis from every page. Furthermore, the 4C database enables researchers to find other researchers with similar interests. 4C can establish dialog among researchers in a community and create an online environment that facilitates the discovery and sharing of case knowledge. In its support of collaboration among case study researchers, the 4C database is different from CAQDAS. 
The 4C database also favorably compares to scholarly online research library databases (e.g. ERIC, EDUDATA, CiteSeer, Medline, CINAHL, Web of Science, Canadian Education Fulltext, Pro-Quest Digital Dissertations, and The National Library of Canada) or e-libraries where people can post their research (e.g., SSRN http://ssrn.com/). Existing library databases lack user-driven search terms as well as effective ways of facilitating the comparison of case studies. Library searches are generally limited to metadata such as keyword, institution, author, or subject words, and do not adequately support the locating of meaningful case study research, or cross-case analyses. Moreover, traditional indexing methods are used to retrieve and analyze the research studies but these tools do not include research in progress, do not permit uploading and editing of data by the author, or do not involve researchers in building a community of users based on identifying and recommending research. On the other hand, the 4C database supports works in progress and allows further updates of case studies. The 4C database affords researchers with the opportunity to add their perspective or comments to these case studies and share these perspectives with other researchers. The user-driven naming of relationships via tags increases the flexibility and expansion of 4C databases by enabling the researcher to link cases with meaningful terms and to search case studies by these terms. Finally, unlike library repositories, 4C allows different individuals to present and recommend selected case studies of interest on a common problem and facilitates collaboration between these individuals. 
The authors know of no other currently available, single, online tool that supports collaboration amongst case study researchers or allows them to create communities of interest, contribute case study data, discover and analyze existing case studies, perform cross-case analyses, recommend case studies to one another, and foster dialogue about case studies. 
In this paper, the authors suggest that the fundamental power of cross-case analysis emerges from understanding how expertise can be built and shared. Turning to theories of how people learn, we detected a form of cognitive cross-case analysis as a plausible hypothetical process involved in building expertise. We proposed that case study researchers have mobilized their knowledge of the original case when their cross-case analysis is made public. To support the mobilization of case study knowledge, we introduced the Foresee (4C) Database. The design of the Foresee Database was based upon: 1) the above hypotheses on the development of expertise 2) known techniques in cross-case analysis, such as stacking and qualitative comparative analysis, and 3) emerging Web 2.0 capacities, such as tagging and multi-way interactivity, to construct meaningful relationships. 
"This [cross-case] comparison makes it possible for me to develop expertise regarding home-school literacy practices, it helps refine my concepts, and it helps me think about theory in terms of validity across similar events but in different contexts [i.e., Contexts from different studies in the database]. It let me see patterns between concepts and among data. It afforded me the opportunity to take a closer look at my study, and in particular, to look more closely at my data."
"The comparison of those studies [within the database] definitely brought new insight for me. The [comparison] showed me how all of them were carrying the same idea that there must be some kind of meaning to tobacco use prevention or control program in order for it to be successful."
"The potential value of the cross-case analysis that I looked at involved seeing the notion of processes and practices in a new light. This comparison [with other cases in the database] has allowed me to see that my own study is much more built upon literary practices than I had realized." 
More research and application awaits the authors. Having taken steps to locate a theoretical framework and develop a set of design principles, we invite others to join this dialogue on cross-case analysis and knowledge mobilization. 
Allport, Gordon (1962). The general and the unique in psychological science. Journal of Personality, 30(3), 405-423.
Ausubel, David; Novak, Joseph & Hanesian, Helen (1978). Educational psychology: A cognitive view. New York, NY: Holt, Rinehart, and Winston.
Ayres, Lioness; Kavanaugh, Karen & Knafl, Kathleen A. (2003). Within-case and across-case approaches to qualitative data analysis. Qualitative Health Research, 13(6), 871-883.
Becker, Howard S. (1990). Generalizing from case studies. In Elliot W. Eisner and Alan Peshkin (Eds.), Qualitative inquiry in education: The continuing debate (pp.233-242). New York, NY: Teachers College Press.
Cameron, Richard (2004). CiteULike About Pager. Retrieved May 27, 2007, from: http://citeulike.org/.
Cronbach, Lee J. (1975). Beyond the two disciplines of scientific psychology. American Psychologist, 30,116-127.
Denzin, Norman K. (1989). Interpretive interactionism. In Gareth Morgan (Ed.), Beyond methods: Strategies for social research (pp.129-146). Beverly Hills, CA: Sage Publications.
Donmoyer, Robert (1990). Generalizability and the single case study. In Elliot W. Eisner, & Alan Peshkin (Eds.), Qualitative inquiry in education: The continuing debate (pp.175-200). New York, NY: Teachers College Press.
Dreyfus, Hubert L. & Dreyfus, Stuart E. (1988). Mind over machine: The power of human intuition and expertise in the era of the computer. New York, NY: Free Press.
Eckstein, Harry (2002). Case study and theory in political science. In Roger Gomm, Martyn Hammersley, & Peter Foster (Eds.), Case study method: Key issues, key texts (pp.119-163). London: Sage Publications.
Flyvbjerg, Bent (2001). Making social science matter: Why social inquiry fails and how it can succeed again. Cambridge: Cambridge University Press.
Foreman, Paul G. (1948). The theory of case studies. Social Forces, 26(4), 408-419.
George, Alexander L. & Bennett, Andrew (2005). Case studies and theory development in the social sciences. Cambridge, MA: MIT Press.
Goetz, Judith P. & LeCompte, Margaret D. (1984). Ethnography and qualitative design in education research. Orlando, FL: Academic Press.
Goldstone, Jack A. (1997). Methodological issues in comparative macrosociology. Great Britain: JAI Press.
Grudin, Jonathan (1994). Groupware and social dynamics: Eight challenges for developers. Communications of the ACM, 37(1), 92-155.
Guba, Egon G. & Lincoln, Yvonna S. (1981). Effective evaluation. San Francisco: Jossey-Bass.
Guerrero, Luis A. & Fuller, David A. (2001). A pattern system for the development of collaborative applications. Information and Software Technology, 43(7), 457-467.
Hammond, Tony; Hannay, Timo; Lund,Ben & Scott, Joanna (2005). Social bookmarking tools (I): A general review. Dlib Magazine, 11(4). Retrieved May 27, 2007, from: http://www.dlib.org/dlib/april05/hammond/04hammond.html.
Khan, Samia (2007). The case in case-based design of educational software: A methodological interrogation. Educational Technology Research & Development, 1-25.
Kolodner, Janet L. (1993). Case-based reasoning. San Mateo, CA: Morgan.
Kolodner, Janet L.; Camp, Paul J.; Crismond, David; Fasse, Barbara; Gray, Jackie; Holbrook, Jennifer; Puntambekar, Sadhana & Ryan, Mike (2003). Problem-based learning meets case-based reasoning in the middle-school science classroom: Putting learning by design™ into practice. Journal of the Learning Sciences, 12(4), 495-547.
Lincoln, Yvonna S. & Guba, Egon G. (2000). The only generalization is: There is no generalization. In Roger Gomm, Martyn Hammersley, & Peter Foster (Eds.), Case study method (pp.27-44). London: Sage Publications.
Mathes, Adam (2004). Folksonomies cooperative classification and communication through shared metadata. Retrieved September 01, 2006, from: http://www.adammathes.com/academic/computer-mediated-communication/folksonomies.html.
Miles, Matthew B. & Huberman, A. Michael (1994). Qualitative data analysis. Thousand Oaks, CA: Sage Publications.
Mills, John Stuart (1843). A system of logic. London: John W. Parker.
Molenaar, Peter (2004). A manifesto on psychology as idiographic science: Bringing the person back into scientific psychology, this time forever. Measurement, 2(4), 201-218.
Pahlevi, Said Mirza & Kitagawa, Hiroyuki (2003). TAX-PQ dynamic taxonomy probing and query modification for topic-focused Web search.Proceedings of the Eighth International DASFAA Conference on Database Systems for Advanced Applications, 91-100.
Peattie, Lisa (2001). Theorizing planning. Some comments on Flyvbjerg's rationality and power. International Planning Studies, 6(3), 257-262.
Przeworski, Adam & Teune, Henry (1982). The logic of comparative social inquiry. Malabar, FL: Robert E. Krieger Publishing Co.
Ragin, Charles (1993). Introduction to qualitative comparative analysis. In Thomas Janoski & Alexander Hicks (Eds.), The comparative political economy of the welfare state (pp.299-319). New York: Cambridge University Press.
Ragin, Charles (1997). Turning the tables: How case-oriented research challenges variable- oriented research. Comparative Social Research, 16, 27-42.
Rueschemeyer, Dietrich (2003). Can one or a few cases yield theoretical gains? In James Mahoney & Dietrich Rueschemeyer (Eds.), Comparative historical analysis in the social sciences (pp.305-336). Cambridge, MA: Cambridge University Press.
Sacco, Giovanni Maria (2004). Uniform access to multimedia information bases through dynamic taxonomies. Proceedings of the Sixth IEEE International Symposium on Multimedia Software Engineering, 320-328.
Schachter, Joshua. Del.icio.us. Retrieved May 27, 2007, from: http://del.icio.us/.
Schank, Roger C. & Berman, Tamara (2002). The pervasive role of stories in knowledge & action. In Melanie Green, Jeffrey Srange, & Timothy Brock (Eds.), Narrative impact: Social and cognitive foundations (pp.287-314). Mahwah,NJ: Erlbaum & Associates.
Schofield, Janet W. (1990). Increasing the generalizability of qualitative research. In Elliot W. Eisner & Alan Peshkin (Eds.), Qualitative inquiry in education: The continuing debate (pp.201-232). New York: Teachers College Press.
Silverstein, A. (1988). An Aristotelian resolution of the idiographic versus nomothetic tension. American Psychologist, 43(6), 425-430.
Smaling, Adri (2003). Inductive, analogical, and communicative generalization. International Journal of Qualitative Methods, 2(1), 1-31.
Stake, Robert (2005). Qualitative case studies. In Norman K. Denzin & Yvonna S. Lincoln (Eds.), Qualitative research (3rd ed., pp.433-466). Thousand Oaks, CA: Sage Publications.
Stake, Robert (2006). Multiple case study analysis. New York, NY: Guilford Press.
Star, Susan Leigh (1998). Grounded classification: Grounded theory and faceted classification. Library Trends, 47(2), 218.
Stretton, Hugh (1969). The political sciences: General principles of selection in social science and history. London: Routledge & Kegan Paul.
Tesch, Renata (1990). Qualitative research: Analysis types and software tools. New York, NY: Palmer.
VanWynsberghe, Robert & Khan, Samia (2007). Redefining case study. International Journal of Qualitative Methods, 6(2), 1-10.
Yee, Ka-Ping; Swearingen, Kirsten; Li, Kevin & Hearst, Marti (2003). Faceted metadata for image search and browsing. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, 401-408.
Yin, Robert (1994). Case study research: Design and methods (2nd ed.). Thousand Oaks, CA: Sage Publications.
Yin, Robert (2003). Case study research: Design and methods (3rd ed.). Thousand Oaks, CA: Sage Publications.
Samia KHAN is a professor in the Curriculum Studies Department, Faculty of Education, at the University of British Columbia, Canada. Substantive research interests include case study research, knowledge mobilization, and pedagogical and technological innovations that are designed to enhance science education learning, especially among women.
Dr. Samia Khan
Department of Curriculum Studies
Faculty of Education
University of British Columbia
2125 Main Mall
Neville Scarfe Building
Vancouver, BC V6T 1Z4
Robert VANWYNSBERGHE is a professor in the School of Human Kinetics at the University of British Columbia, Canada. Substantive research interests include case study research and community mobilization for the purposes of achieving sustainability and health promotion goals.
Robert VanWynsberghe, PhD
Human Kinetics and Educational Studies
Rm. 156g Aud. Annex A
1924 West Mall
University of British Columbia
Vancouver, BC V6T 1Z2
Khan, Samia & VanWynsberghe, Robert (2008). Cultivating the Under-Mined: Cross-Case Analysis as Knowledge Mobilization [54 paragraphs]. Forum Qualitative Sozialforschung / Forum: Qualitative Social Research, 9(1), Art. 34, http://nbn-resolving.de/urn:nbn:de:0114-fqs0801348.
What is a case study?
A case study is a research approach that is used to generate an in-depth, multi-faceted understanding of a complex issue in its real-life context. It is an established research design that is used extensively in a wide variety of disciplines, particularly in the social sciences. A case study can be defined in a variety of ways (Table 5), the central tenet being the need to explore an event or phenomenon in depth and in its natural context. It is for this reason sometimes referred to as a "naturalistic" design; this is in contrast to an "experimental" design (such as a randomised controlled trial) in which the investigator seeks to exert control over and manipulate the variable(s) of interest.
Stake's work has been particularly influential in defining the case study approach to scientific enquiry. He has helpfully characterised three main types of case study: intrinsic, instrumental and collective. An intrinsic case study is typically undertaken to learn about a unique phenomenon. The researcher should define the uniqueness of the phenomenon, which distinguishes it from all others. In contrast, the instrumental case study uses a particular case (some of which may be better than others) to gain a broader appreciation of an issue or phenomenon. The collective case study involves studying multiple cases simultaneously or sequentially in an attempt to generate a still broader appreciation of a particular issue.
These are however not necessarily mutually exclusive categories. In the first of our examples (Table 1), we undertook an intrinsic case study to investigate the issue of recruitment of minority ethnic people into the specific context of asthma research studies, but it developed into a instrumental case study through seeking to understand the issue of recruitment of these marginalised populations more generally, generating a number of the findings that are potentially transferable to other disease contexts. In contrast, the other three examples (see Tables 2, 3 and 4) employed collective case study designs to study the introduction of workforce reconfiguration in primary care, the implementation of electronic health records into hospitals, and to understand the ways in which healthcare students learn about patient safety considerations[4-6]. Although our study focusing on the introduction of General Practitioners with Specialist Interests (Table 2) was explicitly collective in design (four contrasting primary care organisations were studied), is was also instrumental in that this particular professional group was studied as an exemplar of the more general phenomenon of workforce redesign.
What are case studies used for?
According to Yin, case studies can be used to explain, describe or explore events or phenomena in the everyday contexts in which they occur. These can, for example, help to understand and explain causal links and pathways resulting from a new policy initiative or service development (see Tables 2 and 3, for example). In contrast to experimental designs, which seek to test a specific hypothesis through deliberately manipulating the environment (like, for example, in a randomised controlled trial giving a new drug to randomly selected individuals and then comparing outcomes with controls), the case study approach lends itself well to capturing information on more explanatory 'how', 'what' and 'why' questions, such as 'how is the intervention being implemented and received on the ground?'. The case study approach can offer additional insights into what gaps exist in its delivery or why one implementation strategy might be chosen over another. This in turn can help develop or refine theory, as shown in our study of the teaching of patient safety in undergraduate curricula (Table 4)[6,10]. Key questions to consider when selecting the most appropriate study design are whether it is desirable or indeed possible to undertake a formal experimental investigation in which individuals and/or organisations are allocated to an intervention or control arm? Or whether the wish is to obtain a more naturalistic understanding of an issue? The former is ideally studied using a controlled experimental design, whereas the latter is more appropriately studied using a case study design.
Case studies may be approached in different ways depending on the epistemological standpoint of the researcher, that is, whether they take a critical (questioning one's own and others' assumptions), interpretivist (trying to understand individual and shared social meanings) or positivist approach (orientating towards the criteria of natural sciences, such as focusing on generalisability considerations) (Table 6). Whilst such a schema can be conceptually helpful, it may be appropriate to draw on more than one approach in any case study, particularly in the context of conducting health services research. Doolin has, for example, noted that in the context of undertaking interpretative case studies, researchers can usefully draw on a critical, reflective perspective which seeks to take into account the wider social and political environment that has shaped the case.
Example of epistemological approaches that may be used in case study research
How are case studies conducted?
Here, we focus on the main stages of research activity when planning and undertaking a case study; the crucial stages are: defining the case; selecting the case(s); collecting and analysing the data; interpreting data; and reporting the findings.
Defining the case
Carefully formulated research question(s), informed by the existing literature and a prior appreciation of the theoretical issues and setting(s), are all important in appropriately and succinctly defining the case[8,12]. Crucially, each case should have a pre-defined boundary which clarifies the nature and time period covered by the case study (i.e. its scope, beginning and end), the relevant social group, organisation or geographical area of interest to the investigator, the types of evidence to be collected, and the priorities for data collection and analysis (see Table 7). A theory driven approach to defining the case may help generate knowledge that is potentially transferable to a range of clinical contexts and behaviours; using theory is also likely to result in a more informed appreciation of, for example, how and why interventions have succeeded or failed.
Example of a checklist for rating a case study proposal
For example, in our evaluation of the introduction of electronic health records in English hospitals (Table 3), we defined our cases as the NHS Trusts that were receiving the new technology. Our focus was on how the technology was being implemented. However, if the primary research interest had been on the social and organisational dimensions of implementation, we might have defined our case differently as a grouping of healthcare professionals (e.g. doctors and/or nurses). The precise beginning and end of the case may however prove difficult to define. Pursuing this same example, when does the process of implementation and adoption of an electronic health record system really begin or end? Such judgements will inevitably be influenced by a range of factors, including the research question, theory of interest, the scope and richness of the gathered data and the resources available to the research team.
Selecting the case(s)
The decision on how to select the case(s) to study is a very important one that merits some reflection. In an intrinsic case study, the case is selected on its own merits. The case is selected not because it is representative of other cases, but because of its uniqueness, which is of genuine interest to the researchers. This was, for example, the case in our study of the recruitment of minority ethnic participants into asthma research (Table 1) as our earlier work had demonstrated the marginalisation of minority ethnic people with asthma, despite evidence of disproportionate asthma morbidity[14,15]. In another example of an intrinsic case study, Hellstrom et al. studied an elderly married couple living with dementia to explore how dementia had impacted on their understanding of home, their everyday life and their relationships.
For an instrumental case study, selecting a "typical" case can work well. In contrast to the intrinsic case study, the particular case which is chosen is of less importance than selecting a case that allows the researcher to investigate an issue or phenomenon. For example, in order to gain an understanding of doctors' responses to health policy initiatives, Som undertook an instrumental case study interviewing clinicians who had a range of responsibilities for clinical governance in one NHS acute hospital trust. Sampling a "deviant" or "atypical" case may however prove even more informative, potentially enabling the researcher to identify causal processes, generate hypotheses and develop theory.
In collective or multiple case studies, a number of cases are carefully selected. This offers the advantage of allowing comparisons to be made across several cases and/or replication. Choosing a "typical" case may enable the findings to be generalised to theory (i.e. analytical generalisation) or to test theory by replicating the findings in a second or even a third case (i.e. replication logic). Yin suggests two or three literal replications (i.e. predicting similar results) if the theory is straightforward and five or more if the theory is more subtle. However, critics might argue that selecting 'cases' in this way is insufficiently reflexive and ill-suited to the complexities of contemporary healthcare organisations.
The selected case study site(s) should allow the research team access to the group of individuals, the organisation, the processes or whatever else constitutes the chosen unit of analysis for the study. Access is therefore a central consideration; the researcher needs to come to know the case study site(s) well and to work cooperatively with them. Selected cases need to be not only interesting but also hospitable to the inquiry  if they are to be informative and answer the research question(s). Case study sites may also be pre-selected for the researcher, with decisions being influenced by key stakeholders. For example, our selection of case study sites in the evaluation of the implementation and adoption of electronic health record systems (see Table 3) was heavily influenced by NHS Connecting for Health, the government agency that was responsible for overseeing the National Programme for Information Technology (NPfIT). This prominent stakeholder had already selected the NHS sites (through a competitive bidding process) to be early adopters of the electronic health record systems and had negotiated contracts that detailed the deployment timelines.
It is also important to consider in advance the likely burden and risks associated with participation for those who (or the site(s) which) comprise the case study. Of particular importance is the obligation for the researcher to think through the ethical implications of the study (e.g. the risk of inadvertently breaching anonymity or confidentiality) and to ensure that potential participants/participating sites are provided with sufficient information to make an informed choice about joining the study. The outcome of providing this information might be that the emotive burden associated with participation, or the organisational disruption associated with supporting the fieldwork, is considered so high that the individuals or sites decide against participation.
In our example of evaluating implementations of electronic health record systems, given the restricted number of early adopter sites available to us, we sought purposively to select a diverse range of implementation cases among those that were available. We chose a mixture of teaching, non-teaching and Foundation Trust hospitals, and examples of each of the three electronic health record systems procured centrally by the NPfIT. At one recruited site, it quickly became apparent that access was problematic because of competing demands on that organisation. Recognising the importance of full access and co-operative working for generating rich data, the research team decided not to pursue work at that site and instead to focus on other recruited sites.
Collecting the data
In order to develop a thorough understanding of the case, the case study approach usually involves the collection of multiple sources of evidence, using a range of quantitative (e.g. questionnaires, audits and analysis of routinely collected healthcare data) and more commonly qualitative techniques (e.g. interviews, focus groups and observations). The use of multiple sources of data (data triangulation) has been advocated as a way of increasing the internal validity of a study (i.e. the extent to which the method is appropriate to answer the research question)[8,18-21]. An underlying assumption is that data collected in different ways should lead to similar conclusions, and approaching the same issue from different angles can help develop a holistic picture of the phenomenon (Table 2).
Brazier and colleagues used a mixed-methods case study approach to investigate the impact of a cancer care programme. Here, quantitative measures were collected with questionnaires before, and five months after, the start of the intervention which did not yield any statistically significant results. Qualitative interviews with patients however helped provide an insight into potentially beneficial process-related aspects of the programme, such as greater, perceived patient involvement in care. The authors reported how this case study approach provided a number of contextual factors likely to influence the effectiveness of the intervention and which were not likely to have been obtained from quantitative methods alone.
In collective or multiple case studies, data collection needs to be flexible enough to allow a detailed description of each individual case to be developed (e.g. the nature of different cancer care programmes), before considering the emerging similarities and differences in cross-case comparisons (e.g. to explore why one programme is more effective than another). It is important that data sources from different cases are, where possible, broadly comparable for this purpose even though they may vary in nature and depth.
Analysing, interpreting and reporting case studies
Making sense and offering a coherent interpretation of the typically disparate sources of data (whether qualitative alone or together with quantitative) is far from straightforward. Repeated reviewing and sorting of the voluminous and detail-rich data are integral to the process of analysis. In collective case studies, it is helpful to analyse data relating to the individual component cases first, before making comparisons across cases. Attention needs to be paid to variations within each case and, where relevant, the relationship between different causes, effects and outcomes. Data will need to be organised and coded to allow the key issues, both derived from the literature and emerging from the dataset, to be easily retrieved at a later stage. An initial coding frame can help capture these issues and can be applied systematically to the whole dataset with the aid of a qualitative data analysis software package.
The Framework approach is a practical approach, comprising of five stages (familiarisation; identifying a thematic framework; indexing; charting; mapping and interpretation), to managing and analysing large datasets particularly if time is limited, as was the case in our study of recruitment of South Asians into asthma research (Table 1)[3,24]. Theoretical frameworks may also play an important role in integrating different sources of data and examining emerging themes. For example, we drew on a socio-technical framework to help explain the connections between different elements - technology; people; and the organisational settings within which they worked - in our study of the introduction of electronic health record systems (Table 3). Our study of patient safety in undergraduate curricula drew on an evaluation-based approach to design and analysis, which emphasised the importance of the academic, organisational and practice contexts through which students learn (Table 4).
Case study findings can have implications both for theory development and theory testing. They may establish, strengthen or weaken historical explanations of a case and, in certain circumstances, allow theoretical (as opposed to statistical) generalisation beyond the particular cases studied. These theoretical lenses should not, however, constitute a strait-jacket and the cases should not be "forced to fit" the particular theoretical framework that is being employed.
When reporting findings, it is important to provide the reader with enough contextual information to understand the processes that were followed and how the conclusions were reached. In a collective case study, researchers may choose to present the findings from individual cases separately before amalgamating across cases. Care must be taken to ensure the anonymity of both case sites and individual participants (if agreed in advance) by allocating appropriate codes or withholding descriptors. In the example given in Table 3, we decided against providing detailed information on the NHS sites and individual participants in order to avoid the risk of inadvertent disclosure of identities[5,25].
What are the potential pitfalls and how can these be avoided?
The case study approach is, as with all research, not without its limitations. When investigating the formal and informal ways undergraduate students learn about patient safety (Table 4), for example, we rapidly accumulated a large quantity of data. The volume of data, together with the time restrictions in place, impacted on the depth of analysis that was possible within the available resources. This highlights a more general point of the importance of avoiding the temptation to collect as much data as possible; adequate time also needs to be set aside for data analysis and interpretation of what are often highly complex datasets.
Case study research has sometimes been criticised for lacking scientific rigour and providing little basis for generalisation (i.e. producing findings that may be transferable to other settings). There are several ways to address these concerns, including: the use of theoretical sampling (i.e. drawing on a particular conceptual framework); respondent validation (i.e. participants checking emerging findings and the researcher's interpretation, and providing an opinion as to whether they feel these are accurate); and transparency throughout the research process (see Table 8)[8,18-21,23,26]. Transparency can be achieved by describing in detail the steps involved in case selection, data collection, the reasons for the particular methods chosen, and the researcher's background and level of involvement (i.e. being explicit about how the researcher has influenced data collection and interpretation). Seeking potential, alternative explanations, and being explicit about how interpretations and conclusions were reached, help readers to judge the trustworthiness of the case study report. Stake provides a critique checklist for a case study report (Table 9).
Potential pitfalls and mitigating actions when undertaking case study research
Stake's checklist for assessing the quality of a case study report