11 Case studies

What is a case study? a unique instance. Hitchcock and Hughes (1995:

316) suggest that case studies are distinguished
A case study is a specific instance that is less by the methodologies that they employ
frequently designed to illustrate a more general than by the subjects/objects of their inquiry
principle (Nisbet and Watt 1984: 72), it is ‘the (though, as indicated below, there is frequently
study of an instance in action’ (Adelman et al. a resonance between case studies and interpretive
1980) (see http://www.routledge.com/textbooks/ methodologies). Hitchcock and Hughes (1995:
9780415368780 – Chapter 11, file 11.1. ppt). The 322) further suggest that the case study approach
single instance is of a bounded system, for example is particularly valuable when the researcher has
a child, a clique, a class, a school, a community. little control over events. They consider (p. 317)
It provides a unique example of real people in real that a case study has several hallmarks:
situations, enabling readers to understand ideas
O It is concerned with a rich and vivid description
more clearly than simply by presenting them with
of events relevant to the case.
abstract theories or principles. Indeed a case study
O It provides a chronological narrative of events
can enable readers to understand how ideas and
relevant to the case.
abstract principles can fit together (Nisbet and
O It blends a description of events with the
Watt 1984: 72–3). Case studies can penetrate
analysis of them.
situations in ways that are not always susceptible
O It focuses on individual actors or groups
to numerical analysis. As Robson (2002: 183)
of actors, and seeks to understand their
remarks, case studies opt for analytic rather than
perceptions of events.
statistical generalization, that is they develop a
O It highlights specific events that are relevant
theory which can help researchers to understand
to the case.
other similar cases, phenomena or situations.
O The researcher is integrally involved in the
Case studies can establish cause and effect,
case.
indeed one of their strengths is that they observe
O An attempt is made to portray the richness of
effects in real contexts, recognizing that context
the case in writing up the report.
is a powerful determinant of both causes and
effects. As Nisbet and Watt (1984: 78) remark, the Case studies are set in temporal, geographical,
whole is more than the sum of its parts. Sturman organizational, institutional and other contexts
(1999: 103) argues that a distinguishing feature that enable boundaries to be drawn around the
of case studies is that human systems have a case; they can be defined with reference to
wholeness or integrity to them rather than being a characteristics defined by individuals and groups
loose connection of traits, necessitating in-depth involved; and they can be defined by partici-
investigation. Further, contexts are unique and pants’ roles and functions in the case (Hitch-
dynamic, hence case studies investigate and report cock and Hughes 1995: 319) (see http://www.
the complex dynamic and unfolding interactions routledge.com/textbooks/9780415368780 –
of events, human relationships and other factors in Chapter 11, file 11.2. ppt).
254 CASE STUDIES

Case studies full of enlargements of enlargements’, alluding

to the tendency of some case studies to
O will have temporal characteristics which help overemphasize detail to the detriment of seeing
to define their nature the whole picture
O have geographical parameters allowing for their O pomposity: striving to derive or generate pro-
definition found theories from low-level data, or by wrap-
O will have boundaries which allow for definition ping up accounts in high-sounding verbiage
O may be defined by an individual in a particular O blandness: unquestioningly accepting only the
context, at a point in time respondents’ views, or including only those
O may be defined by the characteristics of the aspects of the case study on which people
group agree rather than areas on which they might
O may be defined by role or function disagree.
O may be shaped by organizational or institu-
tional arrangements. Case studies can make theoretical statements, but,
like other forms of research and human sciences,
Case studies strive to portray ‘what it is like’ to these must be supported by the evidence presented.
be in a particular situation, to catch the close up This requires the nature of generalization in case
reality and ‘thick description’ (Geertz 1973b) of study to be clarified. Generalization can take
participants’ lived experiences of, thoughts about various forms, for example:
and feelings for a situation. They involve looking
at a case or phenomenon in its real-life context, O from the single instance to the class of instances
usually employing many types of data (Robson that it represents (for example, a single-sex
2002: 178). They are descriptive and detailed, selective school might act as a case study to
with a narrow focus, combining subjective and catch significant features of other single-sex
objective data (Dyer 1995: 48–9). It is important selective schools)
in case studies for events and situations to be O from features of the single case to a multiplicity
allowed to speak for themselves, rather than to of classes with the same features
be largely interpreted, evaluated or judged by the O from the single features of part of the case to
researcher. In this respect the case study is akin to the whole of that case.
the television documentary.
Simons (1996) has argued that case study needs
This is not to say that case studies are
to address six paradoxes; it needs to:
unsystematic or merely illustrative; case study data
are gathered systematically and rigorously. Indeed O reject the subject–object dichotomy, regarding
Nisbet and Watt (1984: 91) specifically counsel all participants equally
case study researchers to avoid: O recognize the contribution that a genuine
creative encounter can make to new forms
O journalism: picking out more striking features of understanding education
of the case, thereby distorting the full account O regard different ways of seeing as new ways of
in order to emphasize these more sensational knowing
aspects O approximate the ways of the artist
O selective reporting: selecting only that evidence O free the mind of traditional analysis
which will support a particular conclusion, O embrace these paradoxes, with an overriding
thereby misrepresenting the whole case interest in people.
O an anecdotal style: degenerating into an end-
less series of low-level banal and tedious There are several types of case study. Yin
illustrations that take over from in-depth, (1984) identifies three such types in terms
rigorous analysis; one is reminded of Stake’s of their outcomes: exploratory (as a pilot to
(1978) wry comment that ‘our scrapbooks are other studies or research questions); descriptive
 WHAT IS A CASE STUDY? 255

(providing narrative accounts); explanatory critical case study and an extreme or unique case.

Chapter 11
(testing theories). Exploratory case studies that The former, he argues, is
act as a pilot can be used to generate hypotheses
when your theoretical understanding is such that
that are tested in larger scale surveys, experiments
there is a clear, unambiguous and non-trivial set
or other forms of research, e.g. observational.
of circumstances where predicted outcomes will be
However, Adelman et al. (1980) caution against
found. Finding a case which fits, and demonstrating
using case studies solely as preliminaries to other
what has been predicted, can give a powerful boost
studies, e.g. as pre-experimental or pre-survey;
to knowledge and understanding.
rather, they argue, case studies exist in their
(Robson 2002: 182)
own right as a significant and legitimate research
method (see http://www.routledge.com/textbooks/ One can add to the critical case study the issue that
9780415368780 – Chapter 11, file 11.3. ppt). the case in question might possess all, or most, of
Yin’s (1984) classification accords with Merriam the characteristics or features that one is investi-
(1988) who identifies three types: descriptive (nar- gating, more fully or distinctly than under ‘normal’
rative accounts); interpretative (developing con- circumstances, for example, a case study of student
ceptual categories inductively in order to exam- disruptive behaviour might go on in a very dis-
ine initial assumptions); evaluative (explaining ruptive class, with students who are very seriously
and judging). Merriam (1988) also categorizes disturbed or challenging, rather than going into a
four common domains or kinds of case study: class where the level of disruption is not so marked.
ethnographic, historical, psychological and socio- By contrast, Robson (2002: 182) argues that
logical. Sturman (1999: 107), echoing Stenhouse the extreme and the unique case can provide a
(1985), identifies four kinds of case study: an valuable ‘test bed’. Extremes include, he argues,
ethnographic case study – single in-depth study; the situation in which ‘if it can work here it
action research case study; evaluative case study; will work anywhere’, or choosing an ideal set of
and educational case study. Stake (1994) identifies circumstances in which to try out a new approach
three main types of case study: intrinsic case studies or project, maybe to gain a fuller insight into how
(studies that are undertaken in order to understand it operates before taking it to a wider audience
the particular case in question); instrumental case (e.g. the research and development model).
studies (examining a particular case in order to gain Case studies have several claimed strengths and
insight into an issue or a theory); collective case weaknesses. These are summarized in Box 11.1
studies (groups of individual studies that are under- (Adelman et al. 1980) and Box 11.2 (Nisbet
taken to gain a fuller picture). Because case studies and Watt 1984) (see http://www.routledge.
provide fine-grain detail they can also be used to com/textbooks/9780415368780 – Chapter 11, file
complement other, more coarsely grained – often 11.4. ppt).
large-scale – kinds of research. Case study material Shaughnessy et al. (2003: 290–9) suggest that
in this sense can provide powerful human-scale case studies often lack a high degree of control, and
data on macro-political decision-making, fusing treatments are rarely controlled systematically,
theory and practice, for example the work of Ball yet they are applied simultaneously, and with
(1990), Bowe et al. (1992) and Ball (1994a) on the little control over extraneous variables. This,
impact of government policy on specific schools. they argue, renders it difficult to make inferences
Robson (2002: 181–2) suggests that there are to draw cause-and-effect conclusions from case
an individual case study; a set of individual studies, and there is potential for bias in some
case studies; a social group study; studies of case studies as the therapist is both the participant
organizations and institutions; studies of events, and observer and, in that role, may overstate or
roles and relationships. All of these, he argues, understate the case. Case studies, they argue, may
find expression in the case study method. Robson be impressionistic, and self-reporting may be biased
(2002) adds to these the distinction between a (by the participant or the observer). Further, they
256 CASE STUDIES

Box 11.1
Possible advantages of case study

Case studies have a number of advantages that make them attractive to educational evaluators or researchers. Thus:

O Case study data, paradoxically, are ‘strong in reality’ but difficult to organize. In contrast, other research data are often
‘weak in reality’ but susceptible to ready organization. This strength in reality is because case studies are down-to-earth
and attention-holding, in harmony with the reader’s own experience, and thus provide a ‘natural’ basis for generalization.
O Case studies allow generalizations either about an instance or from an instance to a class. Their peculiar strength lies in
their attention to the subtlety and complexity of the case in its own right.
O Case studies recognize the complexity and ‘embeddedness’ of social truths. By carefully attending to social situations,
case studies can represent something of the discrepancies or conflicts between the viewpoints held by participants. The
best case studies are capable of offering some support to alternative interpretations.
O Case studies, considered as products, may form an archive of descriptive material sufficiently rich to admit subsequent
reinterpretation. Given the variety and complexity of educational purposes and environments, there is an obvious value
in having a data source for researchers and users whose purposes may be different from our own.
O Case studies are ‘a step to action’. They begin in a world of action and contribute to it. Their insights may be directly
interpreted and put to use; for staff or individual self-development, for within-institutional feedback; for formative
evaluation; and in educational policy-making.
O Case studies present research or evaluation data in a more publicly accessible form than other kinds of research report,
although this virtue is to some extent bought at the expense of their length. The language and the form of the
presentation is (we hope) less esoteric and less dependent on specialized interpretation than conventional research
reports. The case study is capable of serving multiple audiences. It reduces the dependence of the reader upon unstated
implicit assumptions and makes the research process itself accessible. Case studies, therefore, may contribute towards
the ‘democratization’ of decision-making (and knowledge itself). At its best, they allow readers to judge the implications
of a study for themselves.

Source: adapted from Adelman et al. 1980

Box 11.2
Strengths and weaknesses of case study

Strengths

O The results are more easily understood by a wide audience (including non-academics) as they are frequently written in
everyday, non-professional language.
O They are immediately intelligible; they speak for themselves.
O They catch unique features that may otherwise be lost in larger scale data (e.g. surveys); these unique features might
hold the key to understanding the situation.
O They are strong on reality.
O They provide insights into other, similar situations and cases, thereby assisting interpretation of other similar cases.
O They can be undertaken by a single researcher without needing a full research team.
O They can embrace and build in unanticipated events and uncontrolled variables.

Weaknesses

O The results may not be generalizable except where other readers/researchers see their application.
O They are not easily open to cross-checking, hence they may be selective, biased, personal and subjective.
O They are prone to problems of observer bias, despite attempts made to address reflexivity.

Source: Nisbet and Watt 1984

 WHAT IS A CASE STUDY? 257

argue that bias may be a problem if the case study What kind of case study is this (what is its

Chapter 11
O

relies on an individual’s memory. purpose)?

Dyer (1995: 50–2) remarks that, reading a O What is reliable evidence?
case study, one has to be aware that a process O What is objective evidence?
of selection has already taken place, and only O What is an appropriate selection to include
the author knows what has been selected in or from the wealth of generated data?
out, and on what criteria; indeed, the participants O What is a fair and accurate account?
themselves may not know what selection has taken O Under what circumstances is it fair to take an
place. Dyer (1995: 48–9) observes that case studies exceptional case (or a critical event – see the
combine knowledge and inference, and it is often discussion of observation in Chapter 18)?
difficult to separate these; the researcher has to be O What kind of sampling is most appropriate?
clear on which of these feature in the case study O To what extent is triangulation required and
data. how will this be addressed?
From the preceding analysis it is clear that case O What is the nature of the validation process in
studies frequently follow the interpretive tradition case studies?
of research – seeing the situation through the O How will the balance be struck between
eyes of participants – rather than the quantitative uniqueness and generalization?
paradigm, though this need not always be the O What is the most appropriate form of writing
case. Its sympathy to the interpretive paradigm up and reporting the case study?
has rendered case study an object of criticism. O What ethical issues are exposed in undertaking
Consider, for example, Smith (1991: 375), who a case study?
argues that not only is the case study method
the logically weakest method of knowing but also A key issue in case study research is the selection
studying individual cases, careers and communities of information. Although it is frequently useful
is a thing of the past, and that attention should to record typical, representative occurrences, the
be focused on patterns and laws in historical researcher need not always adhere to criteria of
research. representativeness. It may be that infrequent,
This is prejudice and ideology rather than unrepresentative but critical incidents or events
critique, but signifies the problem of respectability occur that are crucial to the understanding of
and legitimacy that case study has to conquer the case. For example, a subject might only
among certain academics. Like other research demonstrate a particular behaviour once, but it is
methods, case study has to demonstrate reliability so important as not to be ruled out simply because
and validity. This can be difficult, for given it occurred once; sometimes a single event might
the uniqueness of situations, they may be, occur which sheds a hugely important insight
by definition, inconsistent with other case into a person or situation (see the discussion of
studies or unable to demonstrate this positivist critical incidents in Chapter 18); it can be a key
view of reliability. Even though case studies to understanding a situation (Flanagan 1949).
do not have to demonstrate this form of For example, it may be that a psychological
reliability, nevertheless there are important case study might happen upon a single instance
questions to be faced in undertaking case of child abuse earlier in an adult’s life, but the
studies, for example (Adelman et al. 1980; effects of this were so profound as to constitute
Nisbet and Watt 1984; Hitchcock and Hughes a turning point in understanding that adult. A
1995) (see http://www.routledge.com/textbooks/ child might suddenly pass a single comment that
9780415368780 – Chapter 11, file 11.5. ppt): indicates complete frustration with or complete
fear of a teacher, yet it is too important to
O What exactly is a case? overlook. Case studies, in not having to seek
O How are cases identified and selected? frequencies of occurrences, can replace quantity
258 CASE STUDIES

with quality and intensity, separating the significant structured interviews that yielded simple frequency
few from the insignificant many instances of counts. Blease and Cohen’s (1990) study of
behaviour. Significance rather than frequency is coping with computers used highly structured
a hallmark of case studies, offering the researcher observation schedules, undertaken by non-
an insight into the real dynamics of situations and participant observers, with the express intention
people. of obtaining precise, quantitative data on the
classroom use of a computer programme. This was
part of a longitudinal study in primary classrooms,
Examples of kinds of case study and yielded typical profiles of individual
Unlike the experimenter who manipulates behaviour and group interaction in students’
variables to determine their causal significance usage of the computer programme. Antonsen’s
or the surveyor who asks standardized questions (1988)) study was of a single child undergoing
of large, representative samples of individuals, psychotherapy at a child psychiatric unit,
the case study researcher typically observes the and uses unstructured observation within the
characteristics of an individual unit – a child, a artificial setting of a psychiatric clinic and
clique, a class, a school or a community. The is a record of the therapist’s non-directive
purpose of such observation is to probe deeply and approach. Finally Houghton’s (1991) study uses
to analyse intensively the multifarious phenomena data from structured sets of test materials
that constitute the life cycle of the unit with together with focused interviews with those with
a view to establishing generalizations about the whom this international student had contact.
wider population to which that unit belongs. Together these case studies provide a valuable
Antipathy among researchers towards the insight into the range and types of case
statistical – experimental paradigm has created study (see http://www.routledge.com/textbooks/
something of a boom industry in case study 9780415368780 – Chapter 11, file 11.6. ppt).
research. Delinquents (Patrick 1973), dropouts There are two principal types of observa-
(Parker 1974), drug-users (Young 1971) and tion – participant observation and non-participant
schools (King 1979) attest to the wide use of observation. In the former, observers engage in the
the case study in contemporary social science and very activities they set out to observe. Often, their
educational research. Such wide use is marked by ‘cover’ is so complete that as far as the other par-
an equally diverse range of techniques employed ticipants are concerned, they are simply one of the
in the collection and analysis of both qualitative group. In the case of Patrick (1973) for example,
and quantitative data. Whatever the problem or born and bred in Glasgow, his researcher role re-
the approach, at the heart of every case study lies mained hidden from the members of the Glasgow
a method of observation. gang in whose activities he participated for a pe-
In Box 11.3 we set out a typology of riod of four months. Such complete anonymity
observation studies on the basis of which our is not always possible, however. Thus in Parker’s
six examples are selected (for further explication (1974) study of downtown Liverpool adolescents,
of these examples, see the accompanying it was generally known that the researcher was
web site: http://www.routledge.com/textbooks/ waiting to take up a post at the university. In the
9780415368780 – Chapter 11, file 11.1.doc). mean time, ‘knocking around’ during the day with
Acker’s (1990) study is an ethnographic the lads and frequenting their pub at night rapidly
account that is based on several hundred established that he was ‘OK’. The researcher was,
hours of participant observation material, while in his own terms, ‘a drinker, a hanger-arounder’
Boulton’s (1992) work, by contrast, is based who could be relied on to keep quiet in illegal
on highly structured, non-participant observation matters.
conducted over five years. The study by Wild Cover is not necessarily a prerequisite of
et al. (1992) used participant observation, loosely participant observation. In an intensive study
 EXAMPLES OF KINDS OF CASE STUDY 259

of a small group of working-class boys during we identify a continuum of settings ranging from

Chapter 11
their last two years at school and their first the ‘artificial’ environments of the counsellor’s
months in employment, Willis (1977) attended and the therapist’s clinics (cells 5 and 6) to
all the different subject classes at school – ‘not the ‘natural’ environments of school classrooms,
as a teacher, but as a member of the class’ – and staffrooms and playgrounds (cells 1 and 2). Because
worked alongside each boy in industry for a short our continuum is crude and arbitrary we are
period (see http://www.routledge.com/textbooks/ at liberty to locate studies of an information
9780415368780 – Chapter 11, file 11.7. ppt). technology audit and computer usage (cells 3
Non-participant observers, on the other hand, and 4) somewhere between the ‘artificial’ and
stand aloof from the group activities they are the ‘natural’ poles.
investigating and eschew group membership – no Although in theory each of the six examples
great difficulty for King (1979), an adult observer of case studies in Box 11.3 could have been
in infant classrooms. King (1979) recalls how undertaken either as a participant or as a non-
he firmly established his non-participant status participant observation study, a number of factors
with young children by recognizing that they intrude to make one or other of the observational
regarded any adult as another teacher or surrogate strategies the dominant mode of inquiry in a
teacher. Hence he would stand up to maintain particular type of setting. Bailey (1994: 247)
social distance, and deliberately decline to show explains that it is hard for a researcher who
immediate interest, and avoided eye contact. wishes to undertake covert research not to act
The best illustration of the non-participant as a participant in a natural setting, as, if the
observer role is perhaps the case of the researcher researcher does not appear to be participating,
sitting at the back of a classroom coding up then why is he/she there? Hence, in many natural
every three seconds the verbal exchanges between settings the researchers will be participants. This
teacher and pupils by means of a structured set of is in contrast to laboratory or artificial settings, in
observational categories. which non-participant observation (e.g. through
Often the type of observation undertaken by video recording) may take place.
the researcher is associated with the type of setting What we are saying is that the unstructured,
in which the research takes place. In Box 11.3 ethnographic account of teachers’ work (cell 1)

Box 11.3
A typology of observation studies

Degree of
structure
imposed
by observer Degree of structure in the observational setting
Natural Artificial
Unstructured 1 3 5
Acker (1990) ‘Teachers’ Wild et al. Antonsen (1988) ‘Treatment
culture in an English (1992) ‘Evaluating information of a boy of twelve’
primary school’ technology’

Boulton (1992) ‘Participation Blease and Cohen (1990) Houghton (1991) ‘Mr Chong:
in playground activities’ ‘Coping with computers’ a case study of a dependent
learner’
Structured 2 4 6

Source: adapted from Bailey 1978

260 CASE STUDIES

is the most typical method of observation in the Recording observations

natural surroundings of the school in which that
I filled thirty-two notebooks with about half a million
study was conducted. Similarly, the structured
words of notes made during nearly six hundred hours
inventories of study habits and personality
[of observation].
employed in the study of Mr Chong (cell 6) reflect
(King 1979)
a common approach in the artificial setting of a
counsellor’s office. The recording of observations is a frequent source
of concern to inexperienced case study researchers.
Why participant observation? How much ought to be recorded? In what
form should the recordings be made? What does
The natural scientist, Schutz (1962) points out, one do with the mass of recorded data? Lofland
explores a field that means nothing to the (1971) gives a number of useful suggestions about
molecules, atoms and electrons therein. By collecting field notes:
contrast, the subject matter of the world in
which the educational researcher is interested is O Record the notes as quickly as possible after
composed of people and is essentially meaningful observation, since the quantity of information
to them. That world is subjectively structured, forgotten is very slight over a short period
possessing particular meanings for its inhabitants. of time but accelerates quickly as more time
The task of the educational investigator is very passes.
often to explain the means by which an orderly O Discipline yourself to write notes quickly and
social world is established and maintained in reconcile yourself to the fact that although it
terms of its shared meanings. How do participant may seem ironic, recording of field notes can
observation techniques assist the researcher in be expected to take as long as is spent in actual
this task? Bailey (1994: 243–4) identifies some observation.
inherent advantages in the participant observation O Dictating rather than writing is acceptable if
approach: one can afford it, but writing has the advantage
O Observation studies are superior to experiments of stimulating thought.
and surveys when data are being collected on O Typing field notes is vastly preferable to
non-verbal behaviour. handwriting because it is faster and easier to
O In observation studies, investigators are able to read, especially when making multiple copies.
discern ongoing behaviour as it occurs and are O It is advisable to make at least two copies of
able to make appropriate notes about its salient field notes and preferable to type on a master
features. for reproduction. One original copy is retained
O Because case study observations take place for reference and other copies can be used as
over an extended period of time, researchers rough draught to be cut up, reorganized and
can develop more intimate and informal rewritten.
relationships with those they are observing, O The notes ought to be full enough adequately
generally in more natural environments than to summon up for one again, months later,
those in which experiments and surveys are a reasonably vivid picture of any described
conducted. event. This probably means that one ought to
O Case study observations are less reactive than be writing up, at the very minimum, at least
other types of data-gathering methods. For a couple of single-space typed pages for every
example, in laboratory-based experiments and hour of observation.
in surveys that depend upon verbal responses
to structured questions, bias can be introduced The sort of note-taking recommended
in the very data that researchers are attempting by Lofland (1971) and actually undertaken by King
to study. (1979) and Wolcott (1973) in their ethnographic
 PLANNING A CASE STUDY 261

accounts grows out of the nature of the unstruc- Box 11.4

Chapter 11
tured observation study. Note-taking, confessed The case study and problems of selection
Wolcott (1973), helped him fight the acute bore-
dom that he sometimes felt when observing the Among the issues confronting the researcher at the
interminable meetings that are the daily lot of the outset of the case study are the problems of selection.
The following questions indicate some of the obstacles
school principal. Occasionally, however, a series
in this respect:
of events would occur so quickly that Wolcott
(1973) had time only to make cursory notes which O How do you get from the initial idea to the working
design (from the idea to a specification, to usable
he supplemented later with fuller accounts. One
data)?
useful tip from this experienced ethnographer is O What do you lose in the process?
worth noting: never resume your observations un- O What unwanted concerns do you take on board as
til the notes from the preceding observation are a result?
complete. There is nothing to be gained merely by O How do you find a site which provides the best
location for the design?
your presence as an observer. Until your observa-
O How do you locate, identify and approach key
tions and impressions from one visit are a matter informants?
of record, there is little point in returning to the O How they see you creates a context within which
classroom or school and reducing the impact of one you see them. How can you handle such social
set of events by superimposing another and more complexities?
O How do you record evidence? When? How much?
recent set. Indeed, when to record one’s data is but
O How do you file and categorize it?
one of a number of practical problems identified O How much time do you give to thinking and
by Walker (1980), which are listed in Box 11.4. reflecting about what you are doing?
O At what points do you show your subject what you
are doing?
Planning a case study O At what points do you give them control over who
In planning a case study there are several is- sees what?
O Who sees the reports first?
sues that researchers may find useful to con-
sider (e.g. Adelman et al. 1980) (see http://www.
routledge.com/textbooks/9780415368780 – Source: adapted from Walker 1980
Chapter 11, file 11.8. ppt):
O The particular circumstances of the case, in-
cluding: the possible disruption to individ- in conducting interviews, it may be wiser to
ual participants that participation might en- interview senior people later rather than ear-
tail; negotiating access to people; negotiating lier so that the most effective use of discussion
ownership of the data; negotiating release of time can be made, the interviewer having been
the data. put into the picture fully before the inter-
O The conduct of the study, including: the use view. Finally, the conduct of research involves
of primary and secondary sources; the opportu- data analysis, theory generation where appro-
nities to check data; triangulation (including priate, and writing the report. Nisbet and Watt
peer examination of the findings, respon- (1984) suggest that it is important to sepa-
dent validation and reflexivity); data collec- rate conclusions from the evidence, with the
tion methods – in the interpretive paradigm, essential evidence included in the main text,
case studies tend to use certain data collection and to balance illustration with analysis and
methods, e.g. semi-structured and open in- generalization.
terviews, observation, narrative accounts and O The consequences of the research (for partici-
documents, diaries, maybe also tests, rather pants). This might include the anonymizing of
than other methods, e.g. surveys, experi- the research in order to protect participants,
ments. Nisbet and Watt (1984) suggest that, though such anonymization might suggest that
262 CASE STUDIES

a primary goal of case study is generalization Box 11.5

rather than the portrayal of a unique case, Continua of data collection, types and analysis in
i.e. it might go against a central feature case study research
of case study. Anonymizing reports might
render them anodyne, and Adelman et al. Data collection
(1980) suggest that the distortion that is Unstructured Structured
(field notes) (interviews – (survey, census
involved in such anonymization – to render open to closed) data)
cases unrecognizable might be too high a price
to pay for going public.
Data types
Narrative Numeric
(field notes) (coded (ratio scale data)
Nisbet and Watt (1984: 78) suggest three main qualitative data
and non-
stages in undertaking a case study. Because case parametric
studies catch the dynamics of unfolding situations statistics)
it is advisable to commence with a very wide
field of focus, an open phase, without selectivity Data analysis
or prejudgement. Thereafter progressive focusing Journalistic Statistical
enables a narrower field of focus to be established, (impressionistic) (content (inferential
identifying key foci for subsequent study and data analysis) statistics)

collection. At the third stage a draft interpretation

is prepared which needs to be checked with Source: adapted from Sturman 1997
respondents before appearing in the final form.
Nisbet and Watt (1984: 79) advise against the
generation of hypotheses too early in a case study; www.routledge.com/textbooks/9780415368780 –
rather, they suggest, it is important to gather data Chapter 11, file 11.9. ppt).
openly. Respondent validation can be particularly At one pole we have unstructured, typically
useful as respondents might suggest a better way of qualitative data, while at the other we have
expressing the issue or may wish to add or qualify structured, typically quantitative data. Researchers
points. using case study approaches will need to decide
There is a risk in respondent validation, how- which methods of data collection, which type of
ever, that they may disagree with an interpretation. data and techniques of analysis to employ.
Nisbet and Watt (1984: 81) indicate the need to
have negotiated rights to veto. They also recom- Writing up a case study
mend that researchers promise that respondents
The writing up of a case study abides by the
can see those sections of the report that refer to
twin notions of ‘fitness for purpose’ and ‘fitness
them (subject to controls for confidentiality, e.g.
for audience’. Robson (2002: 512–13) suggests six
of others in the case study), and take full account
forms of organizing the writing-up of a case study:
of suggestions and responses made by respondents
and, where possible, to modify the account. In the O In the suspense structure the author presents
case of disagreement between researchers and re- the main findings (e.g. an executive summary)
spondents, researchers should promise to publish in the opening part of the report and
respondents’ comments and criticisms alongside then devotes the remainder of the report
the researchers’ report. to providing evidence, analysis, explanations,
Sturman (1997) places on a set of continua justifications (e.g. for what is selected in or out,
the nature of data collection, types and analy- what conclusions are drawn, what alternative
sis techniques in case study research. These are explanations are rejected), and argument that
presented in summary form (Box 11.5) (see http:// leads to the overall picture or conclusion.
 CONCLUSION 263

In the narrative report a prose account is pro- reader to know which areas are important

Chapter 11
O

vided, interspersed with relevant figures, tables, or unimportant, or whether there are any
emergent issues, analysis and conclusion. omissions. It risks the caprice of the writer.
O In the comparative structure the same case is
Some case studies are of a single situation – a
examined through two or more lenses (e.g.
single child, a single social group, a single class, a
explanatory, descriptive, theoretical) in order
single school. Here any of the above six approaches
either to provide a rich, all-round account
may be appropriate. Some case studies require
of the case, or to enable the reader to have
an unfolding of events, some case studies operate
sufficient information from which to judge
under a ‘snapshot’ approach (e.g. of several schools,
which of the explanations, descriptions or
or classes, or groups at a particular point in
theories best fit(s) the data.
time). In the former it may be important to
O In the chronological structure a simple sequence
preserve the chronology, whereas in the latter
or chronology is used as the organizational
such a chronology may be irrelevant. Some
principle, thereby enabling not only cause
case studies are divided into two main parts
and effect to be addressed, but also possessing
(e.g. Willis 1977): the data reporting and then
the strength of an ongoing story. Adding to
the analysis/interpretation/explanation.
Robson’s (2002) comments, the chronology
can be sectionalized as appropriate (e.g.
key events or key time frames), and
Conclusion
intersperse commentaries on, interpretations
of and explanations for, and summaries of The different strategies we have illustrated in
emerging issues as events unfold (e.g. akin our six examples of case studies in a variety
to ‘memoing’ in ethnographic research). The of educational settings suggest that participant
chronology becomes an organizing principle, observation is best thought of as a generic term that
but different kinds of contents are included at describes a methodological approach rather than
each stage of the chronological sequence. one specific method.1 What our examples have
O In the theory-generating structure, the structure shown is that the representativeness of a particular
follows a set of theoretical constructs or a sample often relates to the observational strategy
case that is being made. Here, Robson (2002) open to the researcher. Generally speaking, the
suggests, each succeeding section of the case larger the sample, the more representative it is,
study contributes to, or constitutes, an element and the more likely that the observer’s role is of a
of a developing ‘theoretical formulation’, participant nature.
providing a link in the chain of argument, For examples of case studies, see the ac-
leading eventually to the overall theoretical companying web site (http://www.routledge.com/
formulation. textbooks/9780415368780 – Chapter 11, file
O In the unsequenced structures the sequence, e.g. 11.2.doc, http://www.routledge.com/textbooks/
chronological, issue-based, event-based, theory 9780415368780 – Chapter 11, file 11.3.doc and
based, is unimportant. Robson (2002) suggests http://www.routledge.com/textbooks/
that this approach renders it difficult for the 9780415368780 – Chapter 11, file 11.4 doc).