A Case in Case Study Methodology.pdf

FIELD METHODS
Meyer / CASE STUDY METHODOLOGY
A Case in Case Study Methodology
CHRISTINE BENEDICHTE MEYER
Norwegian School of Economics and Business Administration
The purpose of this article is to provide a comprehensive view of the case study pro-
cess from the researcher’s perspective, emphasizing methodological consider-
ations. As opposed to other qualitative or quantitative research strategies, such as
grounded theory or surveys, there are virtually no specific requirements guiding
case research. This is both the strength and the weakness of this approach. It is a
strength because it allows tailoring the design and data collection procedures to the
research questions. On the other hand, this approach has resulted in many poor case
studies, leaving it open to criticism, especially from the quantitative field of research.
This article argues that there is a particular need in case studies to be explicit about
the methodological choices one makes. This implies discussing the wide range of
decisions concerned with design requirements, data collection procedures, data
analysis, and validity and reliability. The approach here is to illustrate these deci-
sions through a particular case study of two mergers in the financial industry in
Norway.
In the past few years, a number of books have been published that give use-
ful guidance in conducting qualitative studies (Gummesson 1988; Cassell
and Symon 1994; Miles and Huberman 1994; Creswell 1998; Flick 1998;
Rossman and Rallis 1998; Bryman and Burgess 1999; Marshall and
Rossman 1999; Denzin and Lincoln 2000). One approach often mentioned is
the case study (Yin 1989). Case studies are widely used in organizational
studies in the social science disciplines of sociology, industrial relations, and
anthropology (Hartley 1994). Such a study consists of detailed investigation
of one or more organizations, or groups within organizations, with a view to
providing an analysis of the context and processes involved in the phenome-
non under study.
As opposed to other qualitative or quantitative research strategies, such as
grounded theory (Glaser and Strauss 1967) or surveys (Nachmias and
Nachmias 1981), there are virtually no specific requirements guiding case
research. Yin (1989) and Eisenhardt (1989) give useful insights into the case
study as a research strategy, but leave most of the design decisions on the
table. This is both the strength and the weakness of this approach. It is a
Field Methods, Vol. 13, No. 4, November 2001 329–352
© 2001 Sage Publications
329

strength because it allows tailoring the design and data collection procedures
to the research questions. On the other hand, this approach has resulted in
many poor case studies, leaving it open to criticism, especially from the
quantitative field of research (Cook and Campbell 1979). The fact that the
case study is a rather loose design implies that there are a number of choices
that need to be addressed in a principled way.
Although case studies have become a common research strategy, the
scope of methodology sections in articles published in journals is far too lim-
ited to give the readers a detailed and comprehensive view of the decisions
taken in the particular studies, and, given the format of methodology sec-
tions, will remain so. The few books (Yin 1989, 1993; Hamel, Dufour, and
Fortin 1993; Stake 1995) and book chapters on case studies (Hartley 1994;
Silverman 2000) are, on the other hand, mainly normative and span a broad
range of different kinds of case studies. One exception is Pettigrew (1990,
1992), who places the case study in the context of a research tradition (the
Warwick process research).
Given the contextual nature of the case study and its strength in addressing
contemporary phenomena in real-life contexts, I believe that there is a need
for articles that provide a comprehensive overview of the case study process
from the researcher’s perspective, emphasizing methodological consider-
ations. This implies addressing the whole range of choices concerning spe-
cific design requirements, data collection procedures, data analysis, and
validity and reliability.
WHY A CASE STUDY?
Case studies are tailor-made for exploring new processes or behaviors or
ones that are little understood (Hartley 1994). Hence, the approach is particu-
larly useful for responding to how and why questions about a contemporary
set of events (Leonard-Barton 1990). Moreover, researchers have argued that
certain kinds of information can be difficult or even impossible to tackle by
means other than qualitative approaches such as the case study (Sykes 1990).
Gummesson (1988:76) argues that an important advantage of case study
research is the opportunity for a holistic view of the process: “The detailed
observations entailed in the case study method enable us to study many dif-
ferent aspects, examine them in relation to each other, view the process
within its total environment and also use the researchers’ capacity for
‘verstehen.’ ”
The contextual nature of the case study is illustrated in Yin’s (1993:59)
definition of a case study as an empirical inquiry that “investigates a contem-
330 FIELD METHODS

porary phenomenon within its real-life context and addresses a situation in
which the boundaries between phenomenon and context are not clearly
evident.”
The key difference between the case study and other qualitative designs
such as grounded theory and ethnography (Glaser and Strauss 1967; Strauss
and Corbin 1990; Gioia and Chittipeddi 1991) is that the case study is open to
the use of theory or conceptual categories that guide the research and analysis
of data. In contrast, grounded theory or ethnography presupposes that theo-
retical perspectives are grounded in and emerge from firsthand data.
Hartley (1994) argues that without a theoretical framework, the
researcher is in severe danger of providing description without meaning.
Gummesson (1988) says that a lack of preunderstanding will cause the
researcher to spend considerable time gathering basic information. This
preunderstanding may arise from general knowledge such as theories, mod-
els, and concepts or from specific knowledge of institutional conditions and
social patterns. According to Gummesson, the key is not to require research-
ers to have split but dual personalities: “Those who are able to balance on a
razor’s edge using their pre-understanding without being its slave” (p. 58).
DESCRIPTION OF THE ILLUSTRATIVE STUDY
The study that will be used for illustrative purposes is a comparative and
longitudinal case study of organizational integration in mergers and acquisi-
tions taking place in Norway. The study had two purposes: (1) to identify
contextual factors and features of integration that facilitated or impeded
organizational integration, and (2) to study how the three dimensions of orga-
nizational integration (integration of tasks, unification of power, and integra-
tion of cultures and identities) interrelated and evolved over time. Examples
of contextual factors were relative power, degree of friendliness, and eco-
nomic climate. Integration features included factors such as participation,
communication, and allocation of positions and functions.
Mergers and acquisitions are inherently complex. Researchers in the field
have suggested that managers continuously underestimate the task of inte-
grating the merging organizations in the postintegration process (Haspeslaph
and Jemison 1991). The process of organizational integration can lead to
sharp interorganizational conflict as the different top management styles,
organizational and work unit cultures, systems, and other aspects of organi-
zational life come into contact (Blake and Mounton 1985; Schweiger and
Walsh 1990; Cartwright and Cooper 1993). Furthermore, cultural change in
mergers and acquisitions is compounded by additional uncertainties, ambi-
Meyer / CASE STUDY METHODOLOGY 331

guities, and stress inherent in the combination process (Buono and Bowditch
1989).
I focused on two combinations: one merger and one acquisition. The first
case was a merger between two major Norwegian banks, Bergen Bank and
DnC (to be named DnB), that started in the late 1980s. The second case was a
study of a major acquisition in the insurance industry (i.e., Gjensidige’s
acquisition of Forenede), that started in the early 1990s. Both combinations
aimed to realize operational synergies though merging the two organizations
into one entity. This implied disruption of organizational boundaries and
threat to the existing power distribution and organizational cultures.
The study of integration processes in mergers and acquisitions illustrates
the need to find a design that opens for exploration of sensitive issues such as
power struggles between the two merging organizations. Furthermore, the
inherent complexity in the integration process, involving integration of tasks,
unification of power, and cultural integration stressed the need for in-depth
study of the phenomenon over time. To understand the cultural integration
process, the design also had to be linked to the past history of the two
organizations.
DESIGN DECISIONS
In the introduction, I stressed that a case is a rather loose design that
requires that a number of design choices be made. In this section, I go through
the most important choices I faced in the study of organizational integration
in mergers and acquisitions. These include: (1) selection of cases; (2) sam-
pling time; (3) choosing business areas, divisions, and sites; and (4) selection
of and choices regarding data collection procedures, interviews, documents,
and observation.
Selection of Cases
There are several choices involved in selecting cases. First, there is the
question of how many cases to include. Second, one must sample cases and
decide on a unit of analysis. I will explore these issues subsequently.
Single or Multiple Cases
Case studies can involve single or multiple cases. The problem of single
cases is limitations in generalizability and several information-processing
biases (Eisenhardt 1989).
332 FIELD METHODS

Onewaytorespondtothesebiasesisbyapplyingamulti-caseapproach(Leonard-
Barton 1990). Multiple cases augment external validity and help guard against
observer biases. Moreover, multi-case sampling adds confidence to findings.
By looking at a range of similar and contrasting cases, we can understand a sin-
gle-case finding, grounding it by specifying how and where and, if possible,
why it behaves as it does. (Miles and Huberman 1994)
Given these limitations of the single case study, it is desirable to include
more than one case study in the study. However, the desire for depth and a
pluralist perspective and tracking the cases over time implies that the number
of cases must be fairly few. I chose two cases, which clearly does not support
generalizability any more than does one case, but allows for comparison and
contrast between the cases as well as a deeper and richer look at each case.
Originally, I planned to include a third case in the study. Due to changes in
management during the initial integration process, my access to the case was
limited and I left this case entirely. However, a positive side effect was that it
allowed a deeper investigation of the two original cases and in hindsight
turned out to be a good decision.
Sampling Cases
The logic of sampling cases is fundamentally different from statistical
sampling. The logic in case studies involves theoretical sampling, in which
the goal is to choose cases that are likely to replicate or extend the emergent
theory or to fill theoretical categories and provide examples for polar types
(Eisenhardt 1989). Hence, whereas quantitative sampling concerns itself
with representativeness, qualitative sampling seeks information richness and
selects the cases purposefully rather than randomly (Crabtree and Miller
1992).
The choice of cases was guided by George (1979) and Pettigrew’s (1990)
recommendations. The aim was to find cases that matched the three dimen-
sions in the dependent variable and provided variation in the contextual fac-
tors, thus representing polar cases.
To match the choice of outcome variable, organizational integration, I
chose cases in which the purpose was to fully consolidate the merging par-
ties’ operations. A full consolidation would imply considerable disruption in
the organizational boundaries and would be expected to affect the
task-related, political, and cultural features of the organizations. As for the
contextual factors, the two cases varied in contextual factors such as relative
power, friendliness, and economic climate. The DnB merger was a friendly
combination between two equal partners in an unfriendly economic climate.

Gjensidige’s acquisition of Forenede was, in contrast, an unfriendly and
unbalanced acquisition in a friendly economic climate.
Unit of Analysis
Another way to respond to researchers’ and respondents’ biases is to have
more than one unit of analysis in each case (Yin 1993). This implies that, in
addition to developing contrasts between the cases, researchers can focus on
contrasts within the cases (Hartley 1994). In case studies, there is a choice of
a holistic or embedded design (Yin 1989). A holistic design examines the
global nature of the phenomenon, whereas an embedded design also pays
attention to subunit(s).
I used an embedded design to analyze the cases (i.e., within each case, I
also gave attention to subunits and subprocesses). In both cases, I compared
the combination processes in the various divisions and local networks. More-
over, I compared three distinct change processes in DnB: before the merger,
during the initial combination, and two years after the merger. The overall
and most important unit of analysis in the two cases was, however, the inte-
gration process.
Sampling Time
According to Pettigrew (1990), time sets a reference for what changes can
be seen and how those changes are explained. When conducting a case study,
there are several important issues to decide when sampling time. The first
regards how many times data should be collected, while the second concerns
when to enter the organizations. There is also a need to decide whether to col-
lect data on a continuous basis or in distinct periods.
Number of data collections. I studied the process by collecting real time
and retrospective data at two points in time, with one-and-a-half- and
two-year intervals in the two cases. Collecting data twice had some interest-
ing implications for the interpretations of the data. During the first data col-
lection in the DnB study, for example, I collected retrospective data about the
premerger and initial combination phase and real-time data about the second
step in the combination process.
Although I gained a picture of how the employees experienced the second
stage of the combination process, it was too early to assess the effects of this
process at that stage. I entered the organization two years later and found
interesting effects that I had not anticipated the first time. Moreover, it was
interesting to observe how people’s attitudes toward the merger processes
changed over time to be more positive and less emotional.
334 FIELD METHODS

When to enter the organizations. It would be desirable to have had the
opportunity to collect data in the precombination processes. However,
researchers are rarely given access in this period due to secrecy.
The emphasis in this study was to focus on the postcombination process.
As such, the precombination events were classified as contextual factors.
This implied that it was most important to collect real-time data after the par-
ties had been given government approval to merge or acquire. What would
have been desirable was to gain access earlier in the postcombination pro-
cess. This was not possible because access had to be negotiated. Due to the
change of CEO in the middle of the merger process and the need for renegoti-
ating access, this took longer than expected.
Regarding the second case, I was restricted by the time frame of the study.
In essence, I had to choose between entering the combination process as soon
as governmental approval was given, or entering the organization at a later
stage. In light of the previous studies in the field that have failed to go beyond
the initial two years, and given the need to collect data about the cultural inte-
gration process, I chose the latter strategy. And I decided to enter the organi-
zations at two distinct periods of time rather than on a continuous basis.
There were several reasons for this approach, some methodological and
some practical. First, data collection on a continuous basis would have
required use of extensive observation that I didn’t have access to, and getting
access to two data collections in DnB was difficult in itself. Second, I had a
stay abroad between the first and second data collection in Gjensidige. Col-
lecting data on a continuous basis would probably have allowed for better
mapping of the ongoing integration process, but the contrasts between the
two different stages in the integration process that I wanted to elaborate
would probably be more difficult to detect. In Table 1 I have listed the periods
of time in which I collected data in the two combinations.
Sampling Business Areas, Divisions, and Sites
Even when the cases for a study have been chosen, it is often necessary to
make further choices within each case to make the cases researchable. The
most important criteria that set the boundaries for the study are importance or
criticality, relevance, and representativeness. At the time of the data collec-
tion, my criteria for making these decisions were not as conscious as they
may appear here. Rather, being restricted by time and my own capacity as a
researcher, I had to limit the sites and act instinctively.
In both cases, I decided to concentrate on the core businesses (criticality
criterion) and left out the business units that were only mildly affected by the
integration process (relevance criterion). In the choice of regional offices, I

used the representativeness criterion as the number of offices widely
exceeded the number of sites possible to study. In making these choices, I
relied on key informants in the organizations.
SELECTION OF DATA
COLLECTION PROCEDURES
The choice of data collection procedures should be guided by the research
question and the choice of design. The case study approach typically com-
bines data collection methods such as archives, interviews, questionnaires,
and observations (Yin 1989). This triangulated methodology provides stron-
ger substantiation of constructs and hypotheses. However, the choice of data
collection methods is also subject to constraints in time, financial resources,
and access.
I chose a combination of interviews, archives, and observation, with main
emphasis on the first two. Conducting a survey was inappropriate due to the
lack of established concepts and indicators. The reason for limited observa-
tion, on the other hand, was due to problems in obtaining access early in the
study and time and resource constraints. In addition to choosing among sev-
eral different data collection methods, there are a number of choices to be
made for each individual method.
Interviews
When relying on interviews as the primary data collection method, the
issue of building trust between the researcher and the interviewees becomes
very important. I addressed this issue by several means. First, I established a
procedure of how to approach the interviewees. In most cases, I called them
first, then sent out a letter explaining the key features of the project and out-
336 FIELD METHODS
TABLE 1
Periods of Time for Studying the Two Combinations
Bergen Bank–DnC Gjensidige-Forenede
Announcement of intention October 1989 December 1991
to merge/acquire
Government approval February 1990 June 1992
First data collection Autumn 1991/Winter 1992 Winter 1993/Spring 1994
Second data collection Spring 1994 Autumn 1995

lining the broad issues to be addressed in the interview. In this letter, the sup-
port from the institution’s top management was also communicated. In most
cases, the top management’s support of the project was an important prereq-
uisite for the respondent’s input. Some interviewees did, however, fear that
their input would be open to the top management without disguising the
information source. Hence, it became important to communicate how I
intended to use and store the information.
To establish trust, I also actively used my preunderstanding of the context
in the first case and the phenomenon in the second case. As I built up an
understanding of the cases, I used this information to gain confidence. The
active use of my preunderstanding did, however, pose important challenges
in not revealing too much of the research hypotheses and in balancing
between asking open-ended questions and appearing knowledgeable.
There are two choices involved in conducting interviews. The first con-
cerns the sampling of interviewees. The second is that you must decide on
issues such as the structure of the interviews, use of tape recorder, and
involvement of other researchers.
Sampling Interviewees
Following the desire for detailed knowledge of each case and for grasping
different participant’s views the aim was, in line with Pettigrew (1990), to
apply a pluralist view by describing and analyzing competing versions of
reality as seen by actors in the combination processes.
I used four criteria for sampling informants. First, I drew informants from
populations representing multiple perspectives. The first data collection in
DnB was primarily focused on the top management level. Moreover, most
middle managers in the first data collection were employed at the head
offices, either in Bergen or Oslo. In the second data collection, I compensated
for this skew by including eight local middle managers in the sample. The
difference between the number of employees interviewed in DnB and
Gjensidige was primarily due to the fact that Gjensidige has three unions,
whereas DnB only has one. The distribution of interviewees is outlined in
Table 2.
The second criterion was to use multiple informants. According to Glick
et al. (1990), an important advantage of using multiple informants is that the
validity of information provided by one informant can be checked against
that provided by other informants. Moreover, the validity of the data used by
the researcher can be enhanced by resolving the discrepancies among differ-
ent informants’ reports. Hence, I selected multiple respondents from each
perspective.

Third, I focused on key informants who were expected to be knowledge-
able about the combination process. These people included top management
members, managers, and employees involved in the integration project. To
validate the information from these informants, I also used a fourth criterion
by selecting managers and employees who had been affected by the process
but who were not involved in the project groups.
Structured versus unstructured. In line with the explorative nature of the
study, the goal of the interviews was to see the research topic from the per-
spective of the interviewee, and to understand why he or she came to have
this particular perspective. To meet this goal, King (1994:15) recommends
that one have “a low degree of structure imposed on the interviewer, a pre-
ponderance of open questions, a focus on specific situations and action
sequences in the world of the interviewee rather than abstractions and general
opinions.” In line with these recommendations, the collection of primary data
in this study consists of unstructured interviews.
Using tape recorders and involving other researchers. The majority of
the interviews were tape-recorded, and I could thus concentrate fully on ask-
ing questions and responding to the interviewees’ answers. In the few inter-
338 FIELD METHODS
TABLE 2
Distribution of Interviewees
DnB Gjensidige
Phase 1 Phase 2 Phase 1 Phase 2 Total
Level of organization
Top management/board 13 7 10 5 35
Middle management 4 10 8 2 24
Union representative and employees 3 3 11 2 19
Organizational affiliation
Acquirer 10 7 15 4 36
Acquired company 9 11 12 5 37
Neither 1 2 2 5
Locales
Oslo 13 12 11 3 39
Bergen 6 2 8
Trondheim 10 3 13
Locales outside headquarter cities 1 6 8 8 18
Total 20 20 29 9 78

views that were not tape-recorded, most of which were conducted in the first
phase of the DnB-study, two researchers were present. This was useful as we
were both able to discuss the interviews later and had feedback on the role of
an interviewer.
In hindsight, however, I wish that these interviews had been tape-recorded
to maintain the level of accuracy and richness of data. Hence, in the next
phases of data collection, I tape-recorded all interviews, with two exceptions
(people who strongly opposed the use of this device). All interviews that
were tape-recorded were transcribed by me in full, which gave me closeness
and a good grasp of the data.
Documents
When organizations merge or make acquisitions, there are often a vast
number of documents to choose from to build up an understanding of what
has happened and to use in the analyses. Furthermore, when firms make
acquisitions or merge, they often hire external consultants, each of whom
produces more documents. Due to time constraints, it is seldom possible to
collect and analyze all these documents, and thus the researcher has to make a
selection.
The choice of documentation was guided by my previous experience with
merger and acquisition processes and the research question. Hence, obtain-
ing information on the postintegration process was more important than gain-
ing access to the due-diligence analysis. As I learned about the process, I
obtained more documents on specific issues. I did not, however, gain access
to all the documents I asked for, and, in some cases, documents had been lost
or shredded.
The documents were helpful in a number of ways. First, and most impor-
tant, they were used as inputs to the interview guide and saved me time,
because I did not have to ask for facts in the interviews. They were also useful
for tracing the history of the organizations and statements made by key peo-
ple in the organizations. Third, the documents were helpful in counteracting
the biases of the interviews. A list of the documents used in writing the cases
is shown in Table 3.
Observation
The major strength of direct observation is that it is unobtrusive and does
not require direct interaction with participants (Adler and Adler 1994).
Observation produces rigor when it is combined with other methods. When
the researcher has access to group processes, direct observation can illumi-

nate the discrepancies between what people said in the interviews and casual
conversations and what they actually do (Pettigrew 1990).
As with interviews, there are a number of choices involved in conducting
observations. Although I did some observations in the study, I used inter-
views as the key data collection source. Discussion in this article about obser-
vations will thus be somewhat limited. Nevertheless, I faced a number of
choices in conducting observations, including type of observation, when to
enter, how much observation to conduct, and which groups to observe.
The are four ways in which an observer may gather data: (1) the complete
participant who operates covertly, concealing any intention to observe the
setting; (2) the participant-as-observer, who forms relationships and partici-
pates in activities, but makes no secret of his or her intentions to observe
events; (3) the observer-as-participant, who maintains only superficial con-
tact with the people being studied; and (4) the complete observer, who merely
stands back and eavesdrops on the proceedings (Waddington 1994).
In this study, I used the second and third ways of observing. The use of the
participant-as-observer mode, on which much ethnographic research is
based, was rather limited in the study. There were two reasons for this. First, I
had limited time available for collecting data, and in my view interviews
made more effective use of this limited time than extensive participant obser-
vation. Second, people were rather reluctant to let me observe these political
and sensitive processes until they knew me better and felt I could be trusted.
Indeed, I was dependent on starting the data collection before having built
sufficient trust to observe key groups in the integration process. Neverthe-
less, Gjensidige allowed me to study two employee seminars to acquaint me
with the organization. Here I admitted my role as an observer but participated
fully in the activities. To achieve variation, I chose two seminars representing
polar groups of employees.
As observer-as-participant, I attended a top management meeting at the
end of the first data collection in Gjensidige and observed the respondents
during interviews and in more informal meetings, such as lunches. All these
observations gave me an opportunity to validate the data from the interviews.
Observing the top management group was by far the most interesting and
rewarding in terms of input.
Both DnB and Gjensidige started to open up for more extensive observa-
tion when I was about to finish the data collection. By then, I had built up the
trust needed to undertake this approach. Unfortunately, this came a little late
for me to take advantage of it.
340 FIELD METHODS

DATA ANALYSIS
Published studies generally describe research sites and data-collection
methods, but give little space to discuss the analysis (Eisenhardt 1989). Thus,
one cannot follow how a researcher arrives at the final conclusions from a
large volume of field notes (Miles and Huberman 1994).
In this study, I went through the stages by which the data were reduced and
analyzed. This involved establishing the chronology, coding, writing up the
data according to phases and themes, introducing organizational integration
into the analysis, comparing the cases, and applying the theory. I will discuss
these phases accordingly.
The first step in the analysis was to establish the chronology of the cases.
To do this, I used internal and external documents. I wrote the chronologies
up and included appendices in the final report.
The next step was to code the data into phases and themes reflecting the
contextual factors and features of integration. For the interviews, this implied
marking the text with a specific phase and a theme, and grouping the para-
TABLE 3
Documents Used in Analysis of the Combinations
Gjensidige DnB
Strategic plan for 2000 McKinsey reports from the premerger phase
Reports from the integration project groups Reports from the integration project groups
Report from the recruiting committee Guidelines for selecting and positioning in
1990
Internal letters Minutes from top management integration
groups
Letters to the Norwegian authorities Written submissions for the government
approval application
Declaration of intent Merger prospect
Internal job announcement magazines Articles from the press
Report from employee survey Annual reports
Internal newsletters Publications and documents from the banking
crisis projects
Articles from the press
Annual reports
Acquisition prospectus
Union magazines
Publication from the Association of
Norwegian Insurance Companies

graphs on the same theme and phase together. I followed the same procedure
in organizing the documents.
I then wrote up the cases using phases and themes to structure them.
Before starting to write up the cases, I scanned the information on each
theme, built up the facts and filled in with perceptions and reactions that were
illustrative and representative of the data.
The documents were primarily useful in establishing the facts, but they
also provided me with some perceptions and reactions that were validated in
the interviews. The documents used included internal letters and newsletters
as well as articles from the press. The interviews were less factual, as
intended, and gave me input to assess perceptions and reactions. The limited
observation was useful to validate the data from the interviews. The result of
this step was two descriptive cases.
To make each case more analytical, I introduced the three dimensions of
organizational integration—integration of tasks, unification of power, and
cultural integration—into the analysis. This helped to focus the case and to
develop a framework that could be used to compare the cases. The cases were
thus structured according to phases, organizational integration, and themes
reflecting the factors and features in the study.
I took all these steps to become more familiar with each case as a individ-
ual entity. According to Eisenhardt (1989:540), this is a process that “allows
the unique patterns of each case to emerge before the investigators push to
generalise patterns across cases. In addition it gives investigators a rich
familiaritywitheachcasewhich,inturn,acceleratescross-casecomparison.”
The comparison between the cases constituted the next step in the analy-
sis. Here, I used the categories from the case chapters, filled in the features
and factors, and compared and contrasted the findings. The idea behind
cross-case searching tactics is to force investigators to go beyond initial
impressions, especially through the use of structural and diverse lenses on the
data. These tactics improve the likelihood of accurate and reliable theory,
that is, theory with a close fit to the data (Eisenhardt 1989).
As a result, I had a number of overall themes, concepts, and relationships
that had emerged from the within-case analysis and cross-case comparisons.
The next step was to compare these emergent findings with theory from the
organizational field of mergers and acquisitions, as well as other relevant
perspectives.
This method of generalization is known as analytical generalization. In
this approach, a previously developed theory is used as a template with which
to compare the empirical results of the case study (Yin 1989). This compari-
son of emergent concepts, theory, or hypotheses with the extant literature
342 FIELD METHODS

involves asking what it is similar to, what it contradicts, and why. The key to
this process is to consider a broad range of theory (Eisenhardt 1989). On the
whole, linking emergent theory to existent literature enhances the internal
validity, generalizability, and theoretical level of theory-building from case
research.
According to Eisenhardt (1989), examining literature that conflicts with
the emergent literature is important for two reasons. First, the chance of
neglecting conflicting findings is reduced. Second, “conflicting results
forces researchers into a more creative, frame-breaking mode of thinking
than they might otherwise be able to achieve” (p. 544).
Similarly, Eisenhardt (1989) claims that literature discussing similar find-
ings is important because it ties together underlying similarities in phenom-
ena not normally associated with each other. The result is often a theory with
a stronger internal validity, wider generalizability, and a higher conceptual
level.
The analytical generalization in the study included exploring and devel-
oping the concepts and examining the relationships between the constructs.
In carrying out this analytical generalization, I acted on Eisenhardt’s (1989)
recommendation to use a broad range of theory. First, I compared and con-
trasted the findings with the organizational stream on mergers and acquisi-
tion literature. Then I discussed other relevant literatures, including strategic
change, power and politics, social justice, and social identity theory to
explore how these perspectives could contribute to the understanding of the
findings. Finally, I discussed the findings that could not be explained either
by the mergerand acquisitionliteratureor the fourtheoreticalperspectives.
In every scientific study, questions are raised about whether the study is
valid and reliable. The issues of validity and reliability in case studies are just
as important as for more deductive designs, but the application is fundamen-
tally different.
VALIDITY AND RELIABILITY
The problems of validity in qualitative studies are related to the fact that
most qualitative researchers work alone in the field, they focus on the find-
ings rather than describe how the results were reached, and they are limited in
processing information (Miles and Huberman 1994).
Researchers writing about qualitative methods have questioned whether
the same criteria can be used for qualitative and quantitative studies (Kirk
and Miller 1986; Sykes 1990; Maxwell 1992). The problem with the validity

criteria suggested in qualitative research is that there is little consistency
across the articles as each author suggests a new set of criteria.
One approach in examining validity and reliability is to apply the criteria
used in quantitative research. Hence, the criteria to be examined here are
objectivity/intersubjectivity, construct validity, internal validity, external
validity, and reliability.
Objectivity/Intersubjectivity
The basic issue of objectivity can be framed as one of relative neutrality
and reasonable freedom from unacknowledged research biases (Miles and
Huberman 1994). In a real-time longitudinal study, the researcher is in dan-
ger of losing objectivity and of becoming too involved with the organization,
the people, and the process. Hence, Leonard-Barton (1990) claims that one
may be perceived as, and may even become, an advocate rather than an
observer.
According to King (1994), however, qualitative research, in seeking to
describe and make sense of the world, does not require researchers to strive
for objectivity and distance themselves from research participants. Indeed, to
do so would make good qualitative research impossible, as the interviewer’s
sensitivity to subjective aspects of his or her relationship with the interviewee
is an essential part of the research process (King 1994:31).
This does not imply, however, that the issue of possible research bias can
be ignored. It is just as important as in a structured quantitative interview that
the findings are not simply the product of the researcher’s prejudices and
prior experience. One way to guard against this bias is for the researcher to
explicitly recognize his or her presuppositions and to make a conscious effort
to set these aside in the analysis (Gummesson 1988). Furthermore, rival con-
clusions should be considered (Miles and Huberman 1994).
My experience from the first phase of the DnB study was that it was diffi-
cult to focus the questions and the analysis of the data when the research
questions were too vague and broad. As such, developing a framework
before collecting the data for the study was useful in guiding the collection
and analysis of data. Nevertheless, it was important to be open-minded and
receptive to new and surprising data. In the DnB study, for example, the posi-
tive effect of the reorganization process on the integration of cultures came as
a complete surprise to me and thus needed further elaboration.
I also consciously searched for negative evidence and problems by inter-
viewing outliers (Miles and Huberman 1994) and asking problem-oriented
questions. In Gjensidige, the first interviews with the top management
revealed a much more positive perception of the cultural integration process
344 FIELD METHODS

than I had expected. To explore whether this was a result of overreliance on
elite informants, I continued posing problem-oriented questions to outliers
and people at lower levels in the organization. Moreover, I told them about
the DnB study to be explicit about my presuppositions.
Another important issue when assessing objectivity is whether other
researchers can trace the interpretations made in the case studies, or what is
called intersubjectivity. To deal with this issue, Miles and Huberman (1994)
suggest that: (1) the study’s general methods and procedures should be
described in detail, (2) one should be able to follow the process of analysis,
(3) conclusions should be explicitly linked with exhibits of displayed data,
and (4) the data from the study should be made available for reanalysis by
others.
In response to these requirements, I described the study’s data collection
procedures and processing in detail. Then, the primary data were displayed in
the written report in the form of quotations and extracts from documents to
support and illustrate the interpretations of the data. Because the study was
written up in English, I included the Norwegian text in a separate appendix.
Finally, all the primary data from the study were accessible for a small group
of distinguished researchers.
Construct Validity
Construct validity refers to whether there is substantial evidence that the
theoretical paradigm correctly corresponds to observation (Kirk and Miller
1986). In this form of validity, the issue is the legitimacy of the application of
a given concept or theory to established facts.
The strength of qualitative research lies in the flexible and responsive
interaction between the interviewer and the respondents (Sykes 1990). Thus,
meaning can be probed, topics covered easily from a number of angles, and
questions made clear for respondents. This is an advantage for exploring the
concepts (construct or theoretical validity) and the relationships between
them (internal validity). Similarly, Hakim (1987) says the great strength of
qualitative research is the validity of data obtained because individuals are
interviewed in sufficient detail for the results to be taken as true, correct, and
believable reports of their views and experiences.
Construct validity can be strengthened by applying a longitudinal
multicase approach, triangulation, and use of feedback loops. The advantage
of applying a longitudinal approach is that one gets the opportunity to test
sensitivity of construct measures to the passage of time. Leonard-Barton
(1990), for example, found that one of her main constructs, communicability,
varied across time and relative to different groups of users. Thus, the longitu-

dinal study aided in defining the construct more precisely. By using more
than one case study, one can validate stability of construct across situations
(Leonard-Barton 1990). Since my study only consists of two case studies, the
opportunity to test stability of constructs across cases is somewhat limited.
However, the use of more than one unit of analysis helps to overcome this
limitation.
Construct validity is strengthened by the use of multiple sources of evi-
dence to build construct measures, which define the construct and distinguish
it from other constructs. These multiple sources of evidence can include mul-
tiple viewpoints within and across the data sources. My study responds to
these requirements in its sampling of interviewees and uses of multiple data
sources.
Use of feedback loops implies returning to interviewees with interpreta-
tions and developing theory and actively seeking contradictions in data
(Crabtree and Miller 1992; King 1994). In DnB, the written report had to be
approved by the bank’s top management after the first data collection. Apart
from one minor correction, the bank had no objections to the established
facts. In their comments on my analysis, some of the top managers expressed
the view that the political process had been overemphasized, and that the
CEO’s role in initiating a strategic process was undervalued. Hence, an
important objective in the second data collection was to explore these com-
ments further. Moreover, the report was not as positive as the management
had hoped for, and negotiations had to be conducted to publish the report.
The result of these negotiations was that publication of the report was post-
poned one-and-a-half years.
The experiences from the first data collection in the DnB had some conse-
quences. I was more cautious and brought up the problems of confidentiality
and the need to publish at the outset of the Gjensidige study. Also, I had to
struggle to get access to the DnB case for the second data collection and some
of the information I asked for was not released. At Gjensidige, I sent a prelim-
inary draft of the case chapter to the corporation’s top management for com-
ments, in addition to having second interviews with a small number of peo-
ple. Beside testing out the factual description, these sessions gave me the
opportunity to test out the theoretical categories established as a result of the
within-case analysis.
Internal Validity
Internal validity concerns the validity of the postulated relationships
among the concepts. The main problem of internal validity as a criterion in
qualitative research is that it is often not open to scrutiny. According to Sykes
346 FIELD METHODS

(1990), the researcher can always provide a plausible account and, with care-
ful editing, may ensure its coherence. Recognition of this problem has led to
calls for better documentation of the processes of data collection, the data
itself, and the interpretative contribution of the researcher. The discussion of
how I met these requirements was outlined in the section on objectivity/sub-
jectivity above.
However, there are some advantages in using qualitative methods, too.
First, the flexible and responsive methods of data collection allow
cross-checking and amplification of information from individual units as it is
generated. Respondents’ opinions and understandings can be thoroughly
explored. The internal validity results from strategies that eliminate ambigu-
ity and contradiction, filling in detail and establishing strong connections in
data.
Second, the longitudinal study enables one to track cause and effect.
Moreover, it can make one aware of intervening variables (Leonard-Barton
1990). Eisenhardt (1989:542) states, “Just as hypothesis testing research an
apparent relationship may simply be a spurious correlation or may reflect the
impact of some third variable on each of the other two. Therefore, it is impor-
tant to discover the underlying reasons for why the relationship exists.”
Generalizability
According to Mitchell (1983), case studies are not based on statistical
inference. Quite the contrary, the inferring process turns exclusively on the
theoretically necessary links among the features in the case study. The valid-
ity of the extrapolation depends not on the typicality or representativeness of
the case but on the cogency of the theoretical reasoning. Hartley (1994:225)
claims, “The detailed knowledge of the organization and especially the
knowledge about the processes underlying the behaviour and its context can
help to specify the conditions under which behaviour can be expected to
occur. In other words, the generalisation is about theoretical propositions not
about populations.”
Generalizability is normally based on the assumption that this theory may
be useful in making sense of similar persons or situations (Maxwell 1992).
One way to increase the generalizability is to apply a multicase approach
(Leonard-Barton 1990). The advantage of this approach is that one can repli-
cate the findings from one case study to another. This replication logic is sim-
ilar to that used on multiple experiments (Yin 1993).
Given the choice of two case studies, the generalizability criterion is not
supported in this study. Through the discussion of my choices, I have tried to
show that I had to strike a balance between the need for depth and mapping

changes over time and the number of cases. In doing so, I deliberately chose
to provide a deeper and richer look at each case, allowing the reader to make
judgments about the applicability rather than making a case for
generalizability.
Reliability
Reliability focuses on whether the process of the study is consistent and
reasonably stable over time and across researchers and methods (Miles and
Huberman 1994). In the context of qualitative research, reliability is con-
cerned with two questions (Sykes 1990): Could the same study carried out by
two researchers produce the same findings? and Could a study be repeated
using the same researcher and respondents to yield the same findings?
The problem of reliability in qualitative research is that differences
between replicated studies using different researchers are to be expected.
However, while it may not be surprising that different researchers generate
different findings and reach different conclusions, controlling for reliability
may still be relevant. Kirk and Miller’s (1986:311) definition takes into
account the particular relationship between the researcher’s orientation, the
generation of data, and its interpretation:
For reliability to be calculated, it is incumbent on the scientific investigator to
document his or her procedure. This must be accomplished at such a level of
abstraction that the loci of decisions internal to the project are made apparent.
The curious public deserves to know how the qualitative researcher prepares
him orherself forthe endeavour, andhowthe dataiscollected andanalysed.
The study addresses these requirements by discussing my point of departure
regarding experience and framework, the sampling and data collection pro-
cedures, and data analysis.
DISCUSSION
Case studies often lack academic rigor and are, as such, regarded as infe-
rior to more rigorous methods where there are more specific guidelines for
collecting and analyzing data. These criticisms stress that there is a need to be
very explicit about the choices one makes and the need to justify them.
One reason why case studies are criticized may be that researchers dis-
agree about the definition and the purpose of carrying out case studies. Case
348 FIELD METHODS

studies have been regarded as a design (Cook and Campbell 1979), as a quali-
tative methodology (Cassell and Symon 1994), as a particular data collection
procedure (Andersen 1997), and as a research strategy (Yin 1989). Further-
more, the purpose for carrying out case studies is unclear. Some regard case
studies as supplements to more rigorous qualitative studies to be carried out
in the early stage of the research process; others claim that it can be used for
multiple purposes and as a research strategy in its own right (Gummesson
1988; Yin 1989). Given this unclear status, researchers need to be very clear
about their interpretation of the case study and the purpose of carrying out the
study.
This article has taken Yin’s (1989) definition of the case study as a
research strategy as a starting point and argued that the choice of the case
study should be guided by the research question(s). In the illustrative study, I
used a case study strategy because of a need to explore sensitive, ill-defined
concepts in depth, over time, taking into account the context and history of
the mergers and the existing knowledge about the phenomenon. However,
the choice of a case study strategy extended rather than limited the number of
decisions to be made. In Schramm’s (1971, cited in Yin 1989:22–23) words,
“The essence of a case study, the central tendency among all types of case
study, is that it tries to illuminate a decision or set of decisions, why they were
taken, how they were implemented, and with what result.”
Hence, the purpose of this article has been to illustrate the wide range of
decisions that need to be made in the context of a particular case study and to
discuss the methodological considerations linked to these decisions. I argue
that there is a particular need in case studies to be explicit about the method-
ological choices one makes and that these choices can be best illustrated
through a case study of the case study strategy.
As in all case studies, however, there are limitations to the generalizability
of using one particular case study for illustrative purposes. As such, the
strength of linking the methodological considerations to a specific context
and phenomenon also becomes a weakness. However, I would argue that the
questions raised in this article are applicable to many case studies, but that the
answers are very likely to vary. The design choices are shown in Table 4.
Hence, researchers choosing a longitudinal, comparative case study need
to address the same set of questions with regard to design, data collection pro-
cedures, and analysis, but they are likely to come up with other conclusions
given their different research questions.

REFERENCES
Adler, P. A., and P. Adler. 1994. Observational techniques. In Handbook of qualitative research,
edited by N. K. Denzin and Y. S. Lincoln, 377–92. London: Sage.
Andersen, S. S. 1997. Case-studier og generalisering: Forskningsstrategi og design (Case stud-
ies and generalization: Research strategy and design). Bergen, Norway: Fagbokforlaget.
Blake, R. R., and J. S. Mounton. 1985. How to achieve integration on the human side of the
merger. Organizational Dynamics 13 (3): 41–56.
Bryman, A., and R. G. Burgess. 1999. Qualitative research. London: Sage.
Buono, A. F., and J. L. Bowditch. 1989. The human side of mergers and acquisitions. San Fran-
cisco: Jossey-Bass.
Cartwright, S., and C. L. Cooper. 1993. The psychological impact of mergers and acquisitions
onthe individual:Astudyofbuilding societymanagers. HumanRelations 46(3):327–47.
Cassell, C., and G. Symon, eds. 1994. Qualitative methods in organizational research: A practi-
cal guide. London: Sage.
Cook, T. D., and D. T. Campbell. 1979. Quasi experimentation: Design & analysis issues for
field settings. Boston: Houghton Mifflin.
Crabtree, B. F., and W. L. Miller. 1992. Primary care research: A multimethod typology and
qualitative road map. In Doing qualitative research: Methods for primary care, edited by
B. F. Crabtree and W. L. Miller, 3–28. Vol. 3. Thousand Oaks, CA: Sage.
Creswell, J. W. 1998. Qualitative inquiry and research design: Choosing among five traditions.
Thousand Oaks, CA: Sage.
Denzin, N. K., and L. S. Lincoln. 2000. Handbook of qualitative research. London: Sage.
Eisenhardt, K. M. 1989. Building theories from case study research. Academy of Management
Review 14 (4): 532–50.
Flick, U. 1998. An introduction to qualitative research. London: Sage.
350 FIELD METHODS
TABLE 4
Choices and Steps in Case Study Design
1. Selection of cases Single or multiple
sampling
Unit of analysis
2. Sampling time Number of data collections
When to enter
3. Selection of data collection Interviews Sampling interviewees
procedures Structured versus unstructured
Use of tape recorder
Documents Sampling documents
Use of documents
Observation Choosing method
When to enter
How much
Which groups

George, A. L. 1979. Case studies and theory development: The method of structured, focused
comparison. In Diplomacy: New approaches in history, theory, and policy, edited by P. G.
Lauren, 43–68. New York: Free Press.
Gioia, D. A., and K. Chittipeddi. 1991. Sensemaking and sensegiving in strategic change initia-
tion. Strategic Management Journal 12:433–48.
Glaser, B. G., and A. L. Strauss. 1967. The discovery of grounded theory: Strategies for qualita-
tive research. Chicago: Aldine.
Glick, W. H, G. P. Huber, C. C. Miller, D. H. Doty, and K. M. Sutcliffe. 1990. Studying changes
in organizational design and effectiveness: Retrospective event histories and periodic
assessments. Organization Science 1 (3): 293–312.
Gummesson, E. 1988. Qualitative methods in management research. Lund, Norway:
Studentlitteratur, Chartwell-Bratt.
Hakim, C. 1987. Research design. Strategies and choices in the design of social research.
Boston: Unwin Hyman.
Hamel, J., S. Dufour, and D. Fortin. 1993. Case study methods. London: Sage.
Hartley, J. F. 1994. Case studies in organizational research. In Qualitative methods in organiza-
tional research: A practical guide, edited by C. Cassell and G. Symon, 209–29. London:
Sage.
Haspeslaph, P., and D. B. Jemison. 1991. The challenge of renewal through acquisitions.
Planning Review 19 (2): 27–32.
King, N. 1994. The qualitative research interview. In Qualitative methods in organizational
research: A practical guide, edited by C. Cassell and G. Symon, 14–36. London: Sage.
Kirk, J., and M. L. Miller. 1986. Reliability and validity in qualitative research. Qualitative
Research Methods Series 1. London: Sage.
Leonard-Barton, D. 1990. A dual methodology for case studies: Synergistic use of a longitudinal
single site with replicated multiple sites. Organization Science 1 (3): 248–66.
Marshall, C., and G. B. Rossman. 1999. Designing qualitative research. London: Sage.
Maxwell, J. A. 1992. Understanding and validity in qualitative research. Harvard Educational
Review 62 (3): 279–99.
Miles, M. B., and A. M. Huberman. 1994. Qualitative data analysis. 2d ed. London: Sage.
Mitchell, J. C. 1983. Case and situation analysis. Sociology Review 51 (2): 187–211.
Nachmias, C., and D. Nachmias. 1981. Research methods in the social sciences. London:
Edward Arnhold.
Pettigrew, A. M. 1990. Longitudinal field research on change: Theory and practice. Organiza-
tion Science 1 (3): 267–92.
. 1992. The character and significance of strategic process research. Strategic Manage-
ment Journal 13:5–16.
Rossman, G. B., and S. F. Rallis. 1998. Learning in the field: An introduction to qualitative
research. Thousand Oaks, CA: Sage.
Schramm, W. 1971. Notes on case studies for instructional media projects. Working paper for
Academy of Educational Development, Washington DC.
Schweiger, D. M., and J. P. Walsh. 1990. Mergers and acquisitions: An interdisciplinary view.
In Research in personnel and human resource management, edited by G. R. Ferris and K. M.
Rowland, 41–107. Greenwich, CT: JAI.
Silverman, D. 2000. Doing qualitative research: A practical handbook. London: Sage.
Stake, R. E. 1995. The art of case study research. London: Sage.
Strauss, A. L., and J. Corbin. 1990. Basics of qualitative research: Grounded theory procedures
and techniques. Newbury Park, CA: Sage.

Sykes, W. 1990. Validity and reliability in qualitative market research: A review of the litera-
ture. Journal of the Market Research Society 32 (3): 289–328.
Waddington, D. 1994. Participant observation. In Qualitative methods in organizational
research, edited by C. Cassell and G. Symon, 107–22. London: Sage.
Yin, R. K. 1989. Case study research: Design and methods. Applied Social Research Series,
Vol. 5. London: Sage.
. 1993. Applications of case study research. Applied Social Research Series, Vol. 34.
London: Sage.
CHRISTINE BENEDICHTE MEYER is an associate professor in the Department of
Strategy and Management in the Norwegian School of Economics and Business Admin-
istration, Bergen-Sandviken, Norway. Her research interests are mergers and acquisi-
tions, strategic change, and qualitative research. Recent publications include: “Alloca-
tion Processes in Mergers and Acquisitions: An Organisational Justice Perspective”
(British Journal of Management 2001) and “Motives for Acquisitions in the Norwegian
Financial Industry” (CEMS Business Review 1997).
352 FIELD METHODS

A Case in Case Study Methodology.pdf

More Related Content

Similar to A Case in Case Study Methodology.pdf (20)

More from Karla Long (20)

Recently uploaded (20)

A Case in Case Study Methodology.pdf