Dissecting the Case Study Research: Yin and Eisenhardt Approaches

Shreya Mishra

Birla Institute of Management Technology, e-mail: shreya.mishra@bimtech.ac.in

Mishra, Shreya (2021). Dissecting the Case Study Research: Yin and Eisenhardt Approaches. In Dey, A. K. (Ed.), Case Method for Digital Natives: Teaching and Research (1st ed., pp. 243-264). Bloomsbury, India. ISBN: 978-93-54355-21-9

Abstract

In the past six decades, case study research has gained the attention of novice and accomplished scholars from various domains. As case study research evolved, so did its school of thoughts. This led to the development of approaches by four most prominent proponents of the case study research: Robert K. Yin, Kathleen Eisenhardt, Sharan Merriam and Robert E. Stake. Taking into account this development, this two-part article aims to provide their detailed approaches towards conducting a successful case study research. In doing so, it also provides insights into their epistemic lineages while answering questions like when to use case study research, its design, the process of conducting case study research and ensuring its quality for the purpose of acceptability. The focus is to help new and established researchers, who are planning to use case study research, in deciding which approach would fit their research objective based on their own epistemic stance. This first part of the article shall discuss the evolution of the case study research followed by providing a description of two approaches, that is, by Yin and Eisenhardt. The second part of the article shall discuss Stake and Merriam’s approaches to case study research.

 Keywords: Case study research, positivists, research design

Introduction

One of the previous articles in this book discussed the nuances of a teaching case, the objective of which remains to inculcate decision-making and critical thinking among the students as they evolve to become business practitioners. This article, on the other hand, will delve into the other form of case study known as case study research, which is of particular interest to academic researchers. Case study research has its critiques as well as proponents. In recent years, it has become a common form of research design to understand management-related issues. As this approach has gathered attention, multiple forms have evolved and all are frequently used in academic research. However, due to the lack of knowledge about the difference between research-oriented cases and non-research cases, researchers attempting to use the approach end up making blunders that result in rejection. Thus, this article aims to provide an initial grounding on using case study research by presenting four commonly used approaches of case study research presented by Robert Yin, Kathleen Eisenhardt, Robert Stake and Sharan Merriam. In this article we shall focus on the approaches of the former two.

Evolution of Case Study

It is commonly known that while the case study was evolving as a pedagogical tool for business teaching in Harvard, about the same time in the 1920’s, the researchers at Chicago School of Sociology started using case study research. Gradually researchers from natural sciences, humanities and social sciences also started adopting case study methodology to explore and study their respective fields. Evidence of the use of case study can been found in the recorded history (Flyvbjerg, 2011; Sclafani, 2017). It was prominently used in France (Sclafani, 2017).The biography of Charles Darwin (Stewart, 2014) also states the use of cases. However, it gained traction as a methodology in its own right only in the early twentieth century, primarily as a qualitative research approach (Philliber et al., 1980). As noted by Harrison et al. (2017), case method has been extensively used by sociologists and anthropologists in ethnographic research on people and culture.

 Although case study research was predominantly popular among the qualitative researchers, it was used by quantitative and mixed method research designs as well (Mills et al., 2010), with an understandably different lens and objective.

 Despite the presence of qualitative research and particularly case study research in social sciences, the early generation of case study researcher faced the challenge of acceptability of their work. This was mostly because surveys, experiments and statistical analysis were considered the benchmark of a good research. Case studies were usually a mere part of a larger quantitative study or a simple descriptive study (Merriam, 2009). Flyvbjerg (2011) notes that the lack of acceptability of case study research was clearly depicted in the Dictionary of Sociology, which defines case study as:

 The detailed examination of a single example of a class of phenomena, a case study cannot provide reliable information about the broader class, but it may be useful in the preliminary stages of an investigation since it provides hypotheses, which may be tested systematically with a larger number of cases. (p. 301)

 It was driven by the mind set of ‘Let’s get it down to something we can count!’ (Kaplan, 1964, p. 171) that prevailed for decades. However, as noted by Kaplan, it ‘does not always formulate the best research strategy’ (1964, p. 171).

 With the advent of Grounded Theory in the 1960s, things began to change for qualitative researchers. Despite its rise in acceptability, case study research gained limited attention (Mills et al., 2010), and only off late researchers have started to realise the potential of case study research. As put by Mills et al. (2010), ‘Over time the case study approach garnered interest across various disciplines as researchers sought to illuminate phenomena through detailed study of their occurrence in a particular context’ (p. xxxi).

 As qualitative research started gaining attention, the second generation of case study researchers rose and so did the different school of thoughts. Although there are multiple case researchers who understood and provided their own approaches to the methodology, four of them particularly have made their mark in the field. These are Robert K. Yin, Kathleen Eisenhardt, Robert E. Stake and Sharan Merriam. To understand the meaning of case study research, it is imperative to have a fundamental understanding of the case research approach of these prominent scholars.

 Four Schools of Case Study Research

As a researcher, is it important to acknowledge and learn that each approach is driven by a research paradigm that is the epistemological and ontological view of the researcher. In other words, the researcher’s personal worldview. The same holds true for the four pioneers of case study research Yin, Eisenhardt, Stake and Merriam. Hence, in order to describe each school of thought, the scholars’ paradigms would be particularly kept in mind, which is broadly divided into Positivist and Interpretivist/Constructivist stance. Moreover, it is important to confess that the details provided for each approach are taken from single source of each of the scholars’ work. The decision of the main source was the availability of their work that provided precise details of their approach to case study research. For Yin’s approach, the author had access to the sixth edition of his book, Case Study Research and Applications: Design and Method, published in 2018, and hence used it as a base document for explaining his approach to case study research. Details of Eisenhardt’s approach was taken from her pioneer article ‘Building Theory from Case Study Research’, published in 1989, For Robert Stake’s version of case study research, this article used his 1995 book The Art of Case Study Research. Finally, the author of this article failed to get Sharan Merriam’s book on qualitative case study research that as published in 1988 or 1998. Hence, her section includes the details of her approach to case study research from Chapter 3 of her 2009 book Qualitative Research—A Guide to Design and Implementation.

 Positivist Approach

Both Eisenhardt and Yin are known to have a positivist perspective in developing their versions of the methodology. Piekkari and Welch (2018) call Eisenhardt and Yin’s approach to case study as ‘qualitative positivism’. They derive the meaning of the term from Prasad and Prasad (2002):

[Q]ualitative positivism adopts qualitative methods and methodologies, but accompanies them with positivistic assumptions about the nature of social reality (ontology) and the production of knowledge about this reality (epistemology). To be more specific about these (often unstated) commitments, ‘[r]eality is assumed to be concrete, separate from the researcher, and cognisable through the use of so-called objective methods of data collection’ (Prasad & Prasad, 2002, p. 6). (p. 246)

Thus, for those researchers who have a positivist epistemic inclination, following their approach would be appropriate. Each of their approaches to case study research is explicated hereafter.

 Robert Yin’s Approach

 As soon as the Grounded Theory approach was introduced in the 1960s, Yin used this progress as a chance to further the approach of case study research by providing a much-needed structured process of working with the methodology (Harrison et al., 2017. He aimed to make case study research accepted as an equally rigorous process as any other quantitative research (Piekkari & Welch, 2018). His approach makes his epistemic stance of being a realist loud and clear. Moreover, in the sixth edition of his book, Yin (2018) puts emphasis on his realist stance towards case study research. As stated by him, ‘Much of case study research as it is described in this book appears to be oriented toward a realist perspective, which assumes the existence of a single reality that is independent of any observer’. He thus defines case study research in two-fold definition:

 An empirical method that

·        investigates a contemporary phenomenon (the ‘case’) in depth and within its real-world context, especially when

·        the boundaries between phenomenon and context may not be clearly evident. (Yin, 2018, p. 14)

 The first part of the definition draws a line between case study and other empirical research, while the second part is an acknowledgement of the fact that in case study, the distinction between phenomenon and context is often a blur and hence emphasis on other features become inevitable. These features include presence of multiple variables, theoretical propositions, data collection and analysis process, and multiple sources of evidence, finally leading to triangulation.

 More importantly, he notes that the case study approach uses a ‘trilogy’ of three elements which are often used interchangeably but are distinct from each other. These are (a) case study research, which is the mode of inquiry used to reach the objective of the research, in other words the methodology; (b) the case studies as the method of inquiry, that is the concrete steps being taken within the methodology; and (c) case(s) as the unit of inquiry, or the subject that will actually be studied. (Yin, 2018, p. xx)

 Further, according to him, case study should be used when three criteria are met: (a) to ‘answer the “how” and “why” questions’; (b) when the researcher has ‘little or no control over behavioural events’; and (c) the ‘focus of the study is contemporary (as opposed to entirely historical) phenomenon—“case”’ (Yin, 2018, p. 4). Additionally, a case study research would be appropriate when the researcher needs to describe a social phenomenon in-depth.

 Consider that one has interest in studying mergers and acquisitions, and his question is ‘what factors make mergers successful’. Now the researcher looks into the literature and identifies previously identified factors. Using those established factors, he/she makes a questionnaire, with a plan to do a survey research (mode of inquiry) and distributes the survey questionnaire (method of inquiry) to the top management (the unit of inquiry) of successful mergers or acquisitions. Would the results of the survey provide him the answer to the questions? Not really, because the question was ‘what factors make mergers successful’. This question could be modified as ‘why some mergers are successful’ as not all mergers are successful. Hence, previously identified factors may not suffice to answer the question. Moreover, the success factors of certain successful mergers may differ from those of others and a one-size-fits-all approach may not suffice. Thus, to answer a ‘why’ question like this, a researcher may like to choose case study research (mode of inquiry), use a single or multiple case study (method of inquiry) and identify one or more successful mergers to identify the factors (unit of inquiry), where the factors are not in the researcher’s control. Further, because the factors may vary from case to case, the researcher may need to do an in-depth dive to better understand the reasons of the success.

 The Research Design

 Although case study research has been widely accepted as a qualitative research, Yin (2018) while acknowledging it, differs to this notion. According to him ‘case study research may be separate from qualitative research. Case studies may need to follow their own customised research procedures—as in identifying and defining the case to be studied, along with numerous other procedures’ (Yin, 2018, p. xxv).

 Yin’s idea of a case study isn’t limited to a single case study. He purports more than one research design that researchers can decide on—embedded single case or embedded multi-case and holistic single case or holistic multiple case, these will be discussed in the next section.

 Research design in a logical sequence of the research, which starts with the questions one plans to study, identifies the relevant data, identifies the appropriate way of collecting the data and finally analyses the data to be able to reach to the conclusion that could answer the research question one developed in the first place. Put simply, it is the ‘blueprint’ of one’s research (Philliber et al., 1980). Yin cautions researchers about developing the design with an example,

 The design’s main purpose is to avoid the situation in which the evidence does not address the research questions. In this sense, the design deals with a logical, not a logistical, problem. For example, suppose you want to study a single organization. Your research questions have to do with the organization’s competitive or collaborative relationships with other organizations. You can properly address such questions only if you collect information from the other organizations, not just the one you started with. (Yin, 2018, p. 28)

 Components of Research Design

 Yin lists down five main components of a research design, which are discussed hereafter—(a) the research question, (b) theoretical propositions, (c) the ‘case(s)’, (d) linking case to research questions and (e) strength of the findings. These are explicated below.

 Research design, under most circumstances, should start with the research question(s), and hence researchers should spend a considerable amount of time on it (the ‘how’ and ‘why’ for case study). However, Yin suggests that the time when researchers aren’t able to narrow down the research questions, they should begin with a fieldwork to capture the lacunas in the phenomenon they wish to study. This will help them with the research questions.

Once the research questions are identified, he encourages to develop some theoretical propositions. Theoretical propositions are expected outcomes, something like hypothesis but not formulated in the same way. These propositions simply state what the researcher expects to find out. The finding may turn out to be in contrast to the proposition, but it gives an initial direction to the study. He further acknowledges that in exploratory case study, theoretical propositions may not be relevant, however, they should have some purpose—What does the researcher aim to achieve? Thus, the purpose or the objective of the study should be clear from the beginning.

 Another important element of the research design is the ‘case’ itself. The case here means the actual case that researcher would study—from where they shall gather information.

The researcher should be careful in selecting the ‘case’ that needs to be defined as well as bound (Yin, 2018). If one’s study is about ‘how’ and ‘why’ some people opt for multiple career transitions, then the ‘case’ will be the individual or individuals who have gone through multiple career transitions. Or if the study is an event, for instance, a merger of two organisations from entirely different industry, then the ‘case’ will be the merging of those two organisations into one. Once it is defined which ‘case’ or ‘cases’ are taken into consideration, one should then set a boundary of the ‘case’. For instance, when identifying people for career transition, one needs to decide who will be those people, what will be their age groups, the duration of transition, etc. Hence, it needs to be bound by time, space, place, etc.

 Once the question, propositions/purpose of the study and the ‘case(s)’ is identified, one needs to be clear about how to link your data with your questions. Yin (2018) has provided different ways of linking the data to propositions or purpose, which include pattern matching, explanation building, time-series analysis, logic models and cross-case synthesis (Chapter 5 of Yin’s book [2018] provides details of these, although pattern matching, explanation building and logic models are briefly discussed later in the article). However, he warns that unless appropriate data isn’t collected the analysis techniques may not suffice. Hence the researcher should spend enough time to gather appropriate and significant data—not too much and not too less. He notes that there could be combinations of these ways for better linking, however the sufficient data needs to be collected.

 The final criterion of the Yin’s research design is the criterion for interpreting the strength of the case study findings. He notes that in other research designs, the criterion is statistical benchmarks to substantiate the strength of the findings. However, most case studies do not rely on such statistical benchmarks. In such cases, the researcher aims to discard plausible rival explanations that emerge during findings. Rival explanations are nothing but the alternative reasons that seem to be equally plausible when compared with the findings. Yin encourages the researchers to identify as many plausible rival explanations at the design stage itself, so that they can be discarded on the go. This is possible by including ‘cases’ that represent the rival explanations. If at all such rivals are found at the findings stage, ruling them out could be difficult and they will remain an open question for further research.

 Types of Case Study Research Design

 Yin lists four types of cases study research—single holistic, single embedded, multiple holistic and multiple embedded. Before explaining what is holistic and embedded, an explication of single and multiple case study research will make things less complicated.

 Single Case Study Research Design

When the researcher’s method of inquiry uses single ‘case’ to study the phenomenon, it is said to be using single case study research design. According to Yin, single cases are relevant under certain circumstances. One is studying a critical case wherein the case chosen is critical for the theoretical proposition. For instance, if you wish to study how an organisation became bullying free through changing organisational culture, it is important to identify such organisations that claim to be bullying free and have established credible standing among the fraternity and employees for its efforts. In this case, the theoretical proposition is already established (change in organisational culture) and there are rare organisations that are bullying free. Thus, it is a critical case, and hence a single case study design shall suffice to establish this finding.

 Another, reason for using single case study is the case being ‘unusual’. Given the COVID-19 crisis, there could be many cases that can be identified as unusual. For instance, when lockdown was imposed worldwide, many organisations were severely impacted. However, if one organisation was able to increase its revenue while the competitors face losses, it becomes an unusual case and hence can be studied using a single case study design. 

In contrast to the unusual case is the a ‘common’ case, which are happening on a daily basis. Yin suggests that the objective of a common case is to ‘capture the circumstances and conditions of an everyday situation’ (2018, p. 51). Using the example of Duneier’s (1999), he notes ‘a street scene and its sidewalk vendors can become the setting for learning about the potential social benefits created by informal entrepreneurial activity’ (2018, p. 51).

 The fourth situation is that when a case is of a revelatory nature. A case is revelatory when it has not been accessible earlier and the academic fraternity had little knowledge about the same. Hence, in such a situation, if the researcher gets access to a single case that can reveal a phenomenon, using a single case study design is justified. 

 The fifth and final rationale is when a research is longitudinal in nature. In that case, the researcher can justify their use of single case study research design, as the longitudinal nature of the research case can provide various insights over a period of time, which are otherwise not captured. Yin acknowledges that the rationale of using a single case can be more than five.

 Multiple Case Study Research Design

 Simply put, when a researcher uses more than one case to study the phenomenon, the study is said to be multiple case-study research design. Multiple cases can also be used to do a comparative study; in this design, the cases may not be homogenous. Another thing that Yin encourages is to use replications instead of sampling logic. So, for instance if one wishes to do a comparative study, it is advised to have more than one similar cases and more than one contrasting cases. Each should however, lead to corroborate with the initial theoretical propositions. There is no specific number of multiple cases. Anywhere between 2 and 10 can be an ideal number depending on what the researcher wishes to accomplish (2018, p. 56).

 Holistic Versus Embedded Case Study

 As mentioned earlier, both single and multiple cases can be holistic or embedded. So, when the method of inquiry is single case study and unit of inquiry is also single, it is a single holistic case study, while when the unit of inquiry is more than one, it is considered to be a single embedded case study. For instance, you are studying the success of a merger and the unit of inquiry is single branch, it the is former type, and if the unit of inquiry is multiple branch, it is the later type. In case of multiple case study research, the same logic applies. 

 Theory, Rival Theory and Generalisation

 Yin advocates that a study should be derived from a plausible theory; however, he also cautions that theory can digress the researchers from seeing the actual explanation of the theory.

In the simplest sense, a theory is an expected explanation of why something is happening the way it is happening. It may not necessarily be an established theory in the social or management sciences or for that matter any other sciences. Yin provides the following example of building on a theory in advance.

 The case study will show why implementation only succeeded when the organization was able to re-structure itself, and not just overlay the new MIS on the old organizational structure. (Markus, 1983)

The statement presents the nutshell of a theory of MIS implementation—that is, that implementing an MIS goes beyond adding a new technology to an existing organization but requires some organizational restructuring to work.

The same MIS case study then added the following theoretical statement:

The case study will also show why the simple replacement of key persons was not sufficient for successful implementation. (Markus, 1983)

This second statement presents the nutshell of a rival theory—that is, that successful MIS implementation mainly calls for overcoming individuals’ resistance to change (and not any organizational restructuring), leading to the rival theory that the replacement of such people will permit implementation to succeed. (2018, p. 34)

 As it can be seen in the given example, the case study not only was built on an initial theory but also, in priori, claims to rule out a plausible rival theory. The idea about having a rival theory to rule out in advance reduces the possibility of being blinded by one’s own belief about the theoretical explanation. Rival theories or explanations keeps ones abreast about the alternate possibilities and the need to rule them out to justify your theoretical propositions. Having said that, there could well be the possibility that a rival theory ends up being dominant theory that explains the phenomenon. Hence, the researcher needs to have an open mind about the same as well as be aware of various theories that can explain the research questions.

 Having a theoretical proposition also allows generalisation of your case study research. Yin calls this ‘analytical generalisation’ (2018). He enforces the reader to remember that ‘analytical generalisation’ is different from ‘statistical generalisation’. The latter is based on samples that represent a population. In Yin’s approach to case study research, cases are not samples. He states: ‘Rather than thinking about your case(s) as a sample, you should think of your case study as the opportunity to shed empirical light on some theoretical concepts or principles’ (2018, p. 38).

 Statistical generalisations are reached through inferences drawn from a statistically driven study, while analytical generalisation is reached when the findings are either in line with previous studies, provide additional understanding, contrast with previous studies or give entirely new directions. Moreover, in case of analytical generalisation, the findings provide a conceptual level understanding to the question ‘why’ and ‘how’, which lead to further conceptualisation of the phenomenon being studied.

 Quality Parameter of Yin’s Approach

Yin, as mentioned earlier, is a positivist; hence, it is no surprise that his criteria to establish the quality of a research case study are to establish construct validity—whether the measures or tests used actually measure what its claims to measure; external validity—whether the findings can be generalised; internal validity—whether or not the study can successfully establish cause and effect relationship as claimed (valid only for explanatory or causal studies); and reliability—whether the findings will remain consistent in a different situation if the same procedure is followed. He acknowledges that establishing these quality criteria is not simple in a case study research. Moreover, case study research has been criticised for this very reason. To counter these criticisms, Yin has suggested certain tactics that a researcher can use to establish the said validity and reliability of the case study research.

 To establish construct validity, he suggests that researchers should (a) capture data from multiple sources—hence, if one is studying the success of mergers, they should collect data from multiple stakeholders, like owners, top management, employees of both firms, customers and consumers, key informant, experts, internal documents, etc—and (b) on arriving the initial findings the draft should again be reviewed by key informants.

For establishing internal validity, Yin encourages the researchers to do pattern matching—identifying patterns during analysis that corroborates with the predicted pattern made during the time of theoretical propositions. So, for instance your proposition is that it is the leadership that has the most significant influence on the success of the merger, and as you are analysing the data you realise that the same can be found in the statements of the various sources that were interviewed, you are able to confirm your initial proposition. Another way to establish internal validity is to do ‘explanation building’—similar to pattern matching but more difficult to explain and conduct. It is an iterative process and is gradually built during the analysis process as the researcher provides causal explanations of how and why something is happening within the case. Explanation building is usually in a narrative form. So, for instance, in pattern matching you found that multiple sources are suggesting that the leadership is the cause of success, but then as a researcher one should find evidences within the data that suggests that some action of the leader led to certain element of success during the merger. Address rival explanations—as one is building explanations to confirm predicted plausible explanations, they are to simultaneously identify evidence to reject plausible rival explanations. For instance, one of the rival explanation is that transactional knowledge management may result in merger success, however, the researcher found evidence such systems would not be established without the vision of the leadership. Thus, the researcher can explain why the rival explanation may be ruled out. Finally, Yin suggests that the researcher can also use logic models to establish internal validity. Using logic needs a longitudinal study where the researcher can establish that the cause and effect is repeatedly occurring, thus creating an evidence that the findings are linked to the theoretical propositions. One can also use multiple case studies to establish logic models. For instance, by identifying more than one successful merger and unsuccessful merger, and establishing that in each case the cause is the leadership, it creates a logical explanation of why leadership is the main reason for success of mergers. Just a reminder, internal validity is required when one is doing an explanatory study and not for an exploratory or descriptive study.

 Coming to external validity, Yin provide two strategies. In case of using single case study, one should build it on a theory, thus proving along the way that the findings are corroborating to the theory. In case of multiple case studies, the researcher should use the logic of repetition, that each case study should lead to the same findings. For this the researcher needs to choose their ‘cases’ (unit of inquiry) carefully, which could establish repetitions.   

Finally, coming to the reliability, he encourages researchers to develop a protocol, where he provides five specific sections. The first section of the protocol details about the overview of the case study, that includes is objectives, funding or support, the issues that the study shall address and relevant readings about the field of inquiry. The second section pertains to the data collection procedures, sources of data, credentials of the sources, collection tools, anonymity maintaining information, etc. The next section is regarding the protocol questions, the questions that will guide the data collection and an audit trail of evidence while addressing each questions. Finally, the fourth and last section should include an outline of the case study report, which shall include format, documentation and presentation, and bibliography.

Apart from the protocol, a researcher should also maintain a case study database, which include, memos, fieldnotes, interviews, observational notes, etc., and a chain of evidence, where a researcher, if required, can trace back the steps from the findings. Hence, they should be able to reach to the question which is linked to the findings. Thus, Yin’s approach is very methodical and provides a detailed and structured process to case study research.

 Kathleen Eisenhardt’s Approach

 Eisenhardt’s seminal paper was developed on the premise that case study research can very well be used to developed theory, which was a popularly criticised notion at the time. In her article, Eisenhardt draws on the different sources from literature including that of Glaser and Strauss (1967) and Yin (1981, 1984), to provide a step-by-step process to use case study in building theory. Like Yin, Eisenhardt is also a known realist who propagated case study research as positivist approach. As mentioned, her argument that case study research can lead to theory development was based on an eight-step process that will be described hereafter.

 Before getting into the process, Eisenhardt explains what is case study research:

The case study is a research strategy which focuses on understanding the dynamics present within single settings…can involve either single or multiple cases, and numerous levels of analysis…can employ an embedded design, that is, multiple levels of analysis within a single study…[can] typically combine data collection methods such as archives, interviews, questionnaires, and observations. The evidence may be qualitative (e.g., words), quantitative (e.g., numbers), or both…. Finally, case studies can be used to accomplish various aims: to provide description, test theory, or generate theory. (1989, p, 535)

 Although she acknowledges the multidimentionality of using a case study research, she remains focused on theory generation in her process, which is discussed in the next section. She also indicates her approach of case study research should be preferred when ‘little is known about the phenomenon’, ‘current perspectives seem inadequate, due to lack of empirical evidence’, ‘current knowledge conflicts common sense’ and ‘serendipitous findings emerge during a study, which calls for more research.’  

 The Eight Steps to Theory Building

 As mentioned earlier, Eisenhardt provides eight steps that shall help the researchers to develop (mid-range) theory from case studies.

Defining the Research Question

The eight-step process starts with defining the research question. Eisenhardt calls this step ‘getting started’. It is logical to be focused about the research question for the study. As she puts it ‘The rationale for defining the research question is the same as it is in hypothesis-testing research. Without a research focus, it is easy to become overwhelmed by the volume of data’ (p. 536). Hence, the research question, like Yin’s theoretical proposition, guides the overall study and helps the researcher stay on track. She also encourages the researchers to identify relevant constructs in advance. These constructs are to be found from the literature of the phenomenon being studied. As the researcher conducts the study, these constructs can be part of the interviews or questionnaires, and confirmation of the same can later help establish empirical grounding for the theory. Although the listing of research questions and constructs in the starting of the study is vital, Eisenhardt cautions the researcher to not be fixated with them as the research progresses as otherwise the researcher may be blinded to emergent constructs that may prove to be critical in the theory development.

 Before going to the next step, Eisenhardt emphasises on an important component of getting started. She suggests that the researcher should do their best to avoid having a theory or theoretical proposition in advance. Here her stance of case study research deviates from Yin’s approach and in fact is in line with Glaser and Strauss’s Classical Grounded Theory. She encourages to start the research with a clean late and to allow the theory and relationships between variables emerge from the data rather than proposing a theory or relationship in advance.

 Identifying the ‘Cases’

 This next step is crucial in Eisenhardt’s approach to theory building with case study research. To explain what it means, she compares the selection of case studies to the selection of the population in a hypothesis driven case as it allows the researcher to define the boundaries for generalisation. Like in hypothesis research, where the researcher claims that the findings are generalised only to the population from which the sample is drawn, the same logic applies to case study research. However, a main point of difference is that the cases selected are not by random sampling and are rather by theoretical sampling. Like Grounded Theory, the aim is to better explain the emergent theory, and to explain the difference between random and theoretical sampling she notes:

The goal of theoretical sampling is to choose cases which are likely to replicate or extend the emergent theory. In contrast, traditional, within experiment hypothesis-testing studies rely on statistical sampling, in which researchers randomly select the sample from the population. In this type of study, the goal of the sampling process is to obtain accurate statistical evidence on the distributions of variables within the population. (p. 537) 

 Thus, while selecting the cases, the researchers are required to be cautious and not end up choosing cases just for the sake it. The chosen cases should be able to explain the emergent theory well. They can either provide contrasting views or provide replicating views. Researchers who have used theoretical sampling have done that keeping in mind its contribution in building theory.

 Data Collection Tools

 Eisenhardt calls this step Crafting Instruments and Protocols. In line with Yin, Eisenhardt also doesn’t confine case study research into qualitative study. She notes that there is no boundary in case study research, and thus use of qualitative and quantitative data collection tools are permitted. Hence, according to her, a researcher may choose any one data collection method or both. Often the aim is to triangulate the findings to provide better evidence for the theory building. To rationalise the combination of the two she states:

 The combination of data types can be highly synergistic. Quantitative evidence can indicate relationships which may not be salient to the researcher. It also can keep researchers from being carried away by vivid, but false, impressions in qualitative data, and it can bolster findings when it corroborates those findings from qualitative evidence. The qualitative data are useful for understanding the rationale or theory underlying relationships revealed in the quantitative data or may suggest directly theory which can then be strengthened by quantitative support, (Jick, 1979 in Eisenhardt, 1989, p. 538) 

 Within data collection, she also notes that having multiple investigators would be advantageous for the research as it will allow creativity, emergence of multiple perceptions and capturing of more details from the data. She also suggests that when multiple investigators are present, they can be given separate roles. This, in many cases, provides the opportunity of having both attached as well as detached perspective, and hence provide divergent views, substantiating the emergent theory.

Overlapping Analysis and Collection

Again in line with classical grounded theory approach, Eisenhardt advocates the simultaneous data collection and analysis. This is different from Yin’s approach who suggests the researchers should collect data first and then start the analysis. An important part of this step is the consistent capturing of the field notes by the researchers. She especially cautions to not lose sight of important observations as researchers often leave out one off incidences. She further suggests that these notes should also reflect questions about what the researcher is learning, how the case is different or similar to the other case, etc. Overall ,the overlapping of the analysis and data collection is an important aspect of Eisenhardt’s approach, as stated:

 Overlapping data analysis with data collection not only gives the researcher a head start in analysis but, more importantly, allows researchers to take advantage of flexible data collection. Indeed, a key feature of theory-building case research is the freedom to make adjustments during the data collection process. These adjustments can be the addition of cases to probe particular themes which emerge. (p. 539)

 Apart from allowing flexibility, this strategy gives the opportunity to include additional questions to the interviews or questionnaire. Thus, when new themes emerge that were not anticipated, one can justify it by capturing further data. Further, one may ask whether this approach is legitimate or not. May be this question isn’t relevant now, but in 1989 it was a question raised often by the critics of case study research.

 Within-case Analysis

 Eisenhardt notes that analysis is the backbone of a case study research. However, due to the enormous amount of collected data, analysis becomes a daunting process. Analysing within-case data, as she puts it, is one of the ways to simplify it. This step is primarily about narrating each case in as much detail as possible. And it is nothing but a long description of each case. This step allows researchers to capture an in-depth account of every single case that later on is used to identify similarity and contrasts across cases, resulting in an emergence of the pattern. She notes:

 Within-case analysis typically involves detailed case study write-ups for each site. These write-ups are often simply pure descriptions, but they are central to the generation of insight (Gersick, 1988; Pettigrew, 1988) because they help researchers to cope early in the analysis process with the often enormous volume of data. (540)

 Identifying Cross-case Patterns

 Apart from within-case analysis, Eisenhardt’s approach also includes cross-case analysis. She reminds us that researchers often jump to conclusion too soon due to inhibit biases. Hence, she demands researchers to delve into careful cross-case analysis by approaching data in multiple ways. To do so, she provides certain tactics that researchers can use. The first tactic is to identify different dimensions and put the cases within those dimensions. These dimensions can be either identified with the use of literature or based on research questions. The researcher can also choose dimensions as per their own understanding. Then they should do an intergroup and intragroup analysis. To explain this, she provides following example.  

 In a study of strategic decision making, Bourgeois and Eisenhardt (1988) sifted cases into various categories including founder run vs. professional management, high vs. low performance, first vs. second generation product, and large vs. small size. Some categories such as size and product generation revealed no clear patterns, but others such as performance led to important patterns of within-group similarity and across-group differences. (p. 540)

 She provides another tactic for identifying cross-case patterns, wherein she suggests to simply select pairs of cases and then compare the two with each other and compare each pair with other pairs. She argues that this tactic allows identification of subtle differences between pairs, which otherwise remains hidden from the eyes of the researcher. These subtle differences and similarities result in unanticipated patterns. She explains this position with the following illustration:

 Eisenhardt and Bourgeois (1988) found that CEO power differences dominated initial impressions across firms. However, this paired comparison process led the researchers to see that the speed of the decision process was equally important. (p. 541)

Another way to do this is to analysis data separately based on their sources, that is, survey, observation, documents, interviews, etc. When multiple investigators are present, each can analyse one source of data, finally comparing the findings from each source, thus establishing stronger theory.

 Thus, cross-case analysis increases the reliability of the emerging theory as it is corroborated by multiple structured analysis and increases the likelihood of finding novel patterns that could have gone unnoticed.

 Developing Hypothesis

Analysis is followed by development of hypothesis. The iterative analysis process allows the emergence of themes that closely fit the data, leading to construct identification along with their relationships. This is when the researchers can start refining the constructs and build evidence within the data that measures the constructs. The constant comparison that happens between the tentative construct and the evidences from the data finally results into well-defined constructs for theory building. Although it seems similar to that of hypothesis testing, there are striking differences. In hypothesis testing, the constructs are identified in advance—in all probability from the previous literature or pilot study—while here the constructs are identified from the data itself through a rigorous within and through cross-comparison of data. Another difference is that in hypothesis testing, the constructs are present across data, while in case study research (Eisenhardt’s approach), the constructs and their relationships may not be present in all the cases; hence, the researchers have to create an evidential trace of the data that justifies the constructs and their relationships. As Eisenhardt cautions, in most cases, this data is qualitative in nature and hence it needs careful display of evidence. In case of constructs that are not confirmed in all the cases, they may be removed from the theory-building process. However, such disconfirming evidence allows room for future research and further expansion of theory. This is something like Yin’s rival explanations, but not in entirety. The rival explanation is required to be ruled out to be able to prove the theoretical proposition. In case of Eisenhardt, it is an opportunity to extend the emergent theory. In this, she acknowledges, the advantage of qualitative data, through following example:

Eisenhardt and Bourgeois (1988) found a case which did not fit with the proposition that political coalitions have stable memberships. Further examination of this disconfirming case indicated that the executive team in this case had been newly formed at the time of the study. This observation plus replication in another case led to a refinement in the emergent theory to indicate that increasing stabilization of coalitions occurs over time. At this point, the qualitative data are particularly useful for understanding why or why not emergent relationships hold. When a relationship is supported, the qualitative data often provide a good understanding of the dynamics underlying the relationship, that is, the ‘why’ of what is happening.

 This process of replication of emergent pattern through multiple evidences results in establishing internal validity of the construct.

 Comparing with Extant Literature

 The penultimate step in Eisenhardt’s process of case study research is to compare the findings with previous literature. She encourages researchers to look for consistencies and contradictions with extant literature, while also mentioning that the literature should be from a broad range. She further emphasises that discussing the findings in the light of contradictory literature is essential for two specific reasons, one of which is to not let readers dwindling between two different findings that may result in questioning the generalisability of the study. The second reason is to push researchers to use critical thinking and providing an opportunity to extend the theory. For instance, when Eisenhardt and Bourgeois (1988) found contradiction between their finding, that centralization of power leads to politics and the literature states the same for decentralization, they looked for evidence for this contradiction. This resulted in the stronger emergent theory that integration of centralization and decentralization leads to personal efficacy and collaboration.

 While comparing and identifying the reasons for conflicting findings of emergent theory and previous literature is important, the need to present corroborations is equally important as it indicates internal validity of your findings and also extends the generalisability of the theory in different context.

 Eisenhardt states that this step of comparison between the emergent theory and previous literature not only increases internal validity, generalisability and conceptual level of the theory but is also particularly important in case study research as it is built on limited number of cases.

 Reaching Saturation

 Since Eisenhardt’s approach to case study research is iterative and quite similar to the Grounded Theory approach, she suggests researchers to stop adding cases and stop iterating between theory and data when they reach saturation.

She uses Glaser and Strauss (1987) explanation of saturation—it is ‘the point at which incremental learning is minimal because the researchers are observing phenomena seen before’ (1989, p. 545).

She states that generally saturation is reached anywhere between 4 and 10 cases, when initially adding cases. Further, during the iteration between theory and data, when one stops finding new evidences to extend the theory, the saturation is achieved. 

 Quality Parameter of Eisenhardt’s Approach

 Eisenhardt’s approach to case study research is primarily focused on theory building. Hence the quality parameters are also focused on the goodness of the theory emerged from the study. For this, Eisenhardt chooses to apply the Pfeffer’s (1982) suggestion, which notes that a good theory should be parsimonious, testable and logically coherent.

 Apart from this, Eisenhardt seems to follow Yin’s suggestions to ensure the quality of the study. This is evident as she states:

 [T]he assessment of theory-building research also depends upon empirical issues: strength of method and the evidence grounding the theory. Have the investigators followed a careful analytical procedure? Does the evidence support the theory? Have the investigators ruled out rival explanations? Just as in other empirical research, investigators should provide information on the sample, data collection procedures, and analysis. Also, they should display enough evidence for each construct to allow readers to make their own assessment of the fit with theory. While there are no concise measures such as correlation coefficients or F values, nonetheless thorough reporting of information should give confidence that the theory is valid. (p. 548)

By ensuring the aforementioned, the study shall stand the test of validity and reliability. It shall also sharpen the generalisability of the study, allowing it to get acceptability among the academic fraternity.

 Apart from Eisenhardt’s advocacy of theory building in case study research, it is important to mention that Gioia and Chittipeddi (1991) and Ann Langey (1999) have been done significant work in propagating case study research as a way of building theory. However, there approaches differ in many ways. Although, their work is not discussed here, one can find a synthesized comparison of their approach in the essay composed from the 2016 Showcase symposium held at the Annual Meet of Academy of Management (see Gehman et al., 2016).   

 Strength and Weaknesses

 The eight steps of Eisenhardt of theory building through cases, like Yin, provides a structured approach to applying case study research. She advocates this method as it allows development of novel theory, but provided ways of avoiding researchers bias. She advanced the use of case study as a legitimate way of theory building at a time when case studies were not well accepted. The theory emerged from a case study not only may be novel but also provides measurable constructs for further testing and hypothesis building. Finally, she argues that another strength of this approach is that the theory is considered to be empirically valid, and the theory closely fits the data that is captured intimately.

Although the theory developed through this is closer to reality, it can also be very complex due to overwhelming data due to extent of the data availability, Eisenhardt cautions the researchers from trying to capture everything in the theory and making it difficult to test. Further, this may lead to vague understanding of relationships between constructs. Moreover, due to few cases, the theory may lead to narrow generalisability. Thus, while this approach has its advantages, the researchers should be aware about the traps and avoid falling in them.

Both Yin and Eisenhardt, encourage the use of multiple sources for data collection, which include quantitative as well as qualitative ways. But clearly Eisenhardt gives significant credit to qualitative ways, as they provide further insights while conceptualising the emergent theory. As both elaborate their version of case study research, we can clearly see an epistemic inclination towards positivism. Terms like construct, variable, generalisability, validity and reliability clearly suggest that both provide ways of approaching a methodology that appeal to a wide audience of positive researchers. Awareness of research paradigm is a crucial first step that a researcher should take. It is advised to explore and acknowledge it to avoid future disappointments. Case study research has a long history of criticism; however, as the proponents and approaches to case study research have increased, so has its acceptability. With Yin’s and Eisenhardt’s approach to the mode of inquiry, it has become easier for positivist and realist/objectivist researchers to use the methodology, which also fulfils one’s own epistemic stance.

As mentioned earlier, in the second part of this study, we shall discuss the remaining two school of thoughts of case study that which come from the works or Robert Stake and Sharan Merriam. 

Author

Dr. Shreya Mishra is currently working as an Assistant Professor at BIMTECH and Assistant Editor of the South Asian Journal of Management Cases, SAGE. She has completed Doctoral Studies in 2019 in the area of Workplace Bullying. Published in Scopus and ABDC Journals, her research interests include studies on Workplace Bullying, Inequality, Power, Identity, Intersectionality, and Qualitative Research. Currently, her research focus is on coping and navigating through vulnerable life events.