Demystifying Theoretical Sampling in Grounded Theory Research

Jenna Breckenridge BSc(Hons),Ph.D.Candidate and Derek Jones,
PhD, BA (Hons), Dip COT

Abstract

Theoretical sampling is a central tenet of classic grounded theory
and is essential to the development and refinement of a theory
that is ‘grounded’ in data. While many authors appear to share
concurrent definitions of theoretical sampling, the ways in which
the process is actually executed remain largely elusive and
inconsistent. As such, employing and describing the theoretical
sampling process can present a particular challenge to novice
researchers embarking upon their first grounded theory study.
This article has been written in response to the challenges faced
by the first author whilst writing a grounded theory proposal. It
is intended to clarify theoretical sampling for new grounded
theory researchers, offering some insight into the practicalities of
selecting and employing a theoretical sampling strategy. It
demonstrates that the credibility of a theory cannot be
dissociated from the process by which it has been generated and
seeks to encourage and challenge researchers to approach
theoretical sampling in a way that is apposite to the core
principles of the classic grounded theory methodology.

Introduction

With the introduction of grounded theory, Glaser and
Strauss (1967) challenged the prevailing hypothetico-deductive
method of theory verification, questioning the gulf that existed
between abstract theory and empirical research. They advocated
that a theory developed in direct response to immediate problems
under investigation would ultimately be more relevant to the
studied area than any pre-existing theory. Thus proffered as a
potential means of bridging the theory-practice divide, it is
perhaps of little surprise that the grounded theory method has
been embraced widely by the health professions. Grounded theory
offers healthcare researchers a systematic and interpretive
means of generating a theory from data that has the potential to
explain, interpret and guide practice. However, a review of
healthcare literature would suggest that while many authors
profess to using grounded theory, they may only appear to have
‘borrowed’ a particular aspect of the method, most commonly the
constant comparative approach to data analysis (Draucker et al
2007). Furthermore, ‘grounded theory’ studies have been
criticised for possessing a somewhat “mystical” (Melia 1997 p.33)
quality whereby:

a sleight of hand produces a list of ‘themes’, and we are
invited to take it on trust that theory somehow emerges
from the data without being offered a step by step
explanation of how theoretical insights have been built up
(Barbour 2001 p.1116).

Ultimately, this inconsistent application of grounded theory
and the ambiguous way in which grounded theory studies are
often presented within healthcare literature can pose several
challenges to novice researchers. Without being able to refer to
useful exemplars of grounded theory studies it is difficult to
understand and prepare for the practicalities of carrying out one’s
own grounded theory research. Similarly, when using grounded
theory studies as evidence in practice or as part of a literature
review it is difficult to ascertain the credibility of the research if
the product cannot be linked explicitly with the process. This
article has been written in response to the challenges faced by the
first author whilst writing a classic grounded theory proposal,
particularly in relation to theoretical sampling. As an active and
ongoing process that controls and directs data collection and
analysis, theoretical sampling is pivotal in ‘building up
theoretical insights’. However, while many authors appear to
share concurrent definitions of theoretical sampling, the ways in
which the process is actually executed remain largely elusive and
inconsistent. The purpose of this article is thus to clarify
theoretical sampling, explore the practicalities of this strategy,
and offer insight into the appropriate selection, execution and
write-up of theoretical sampling in order to ensure credible and
trustworthy research.

Theoretical Sampling

Classic grounded theory is a general methodology that seeks
to develop, through a process of induction, a theory that is
‘grounded’ in the data from which it has been derived (Glaser
2002a). Sampling is thus theoretically oriented; it is directed
towards the generation and development of conceptual theory as
opposed to creating a descriptive account. It is continually
directed by the emerging theory, following up leads as they arise
in the data and progressively focusing data collection to refine
and integrate the theory (Glaser & Strauss 1967). Interestingly,
despite the evolution of grounded theory since its inception, the
original definition of theoretical sampling has remained largely
undisputed:

the process of data collection for generating theory
whereby the analyst jointly collects, codes and analyses
his data and decides what data to collect next and where
to find them, in order to develop his theory as it emerges
(Glaser and Strauss 1967 p.45).

Indeed, Strauss and Corbin (1998) have described theoretical
sampling as a means to “maximise opportunities to discover
variations among concepts and to densify categories in terms of
their properties and dimensions” (p.201). Furthermore, despite
assuming a different epistemological stance, Charmaz (2006) has
also similarly described theoretical sampling as a means of
focusing data collection and increasing the analytic abstraction of
theory by illuminating variation and identifying gaps that require
elaboration. However, upon closer consideration, it would seem
that while authors may at first glance appear to share a common
definition of theoretical sampling, their apparent congruence with
classic grounded theory is somewhat superficial.

The theoretical sampling process in classic grounded theory
begins with initial data collection and analysis (Glaser 1978).
Open coding of raw data generates initial codes, which in turn
stimulate further data collection. In the initial stages of analysis,
codes are elicited rapidly and it is through a joint process of
theoretical sampling and memo-writing that codes can be
corrected, trimmed, and continually fitted to the data (Glaser
1978). Memo-writing enables the researcher to conceptualise the
boundaries and properties of each category and illuminate gaps
in the emerging theory, thus highlighting where to sample next
and for what theoretical purpose (Glaser 1978). Constant
comparison of codes yields a provisional set of conceptual
categories, from which point new categories emerge and new
incidents are fitted and re-fitted into existing categories. The
researcher samples both for theoretical similarity and difference
in order to expound the properties of each category, attempting to
saturate all categories until the emergence of a core category
(Glaser & Strauss 1967). Theoretical sampling is thereafter
focused on data that is sufficiently and significantly relevant to
the core category and its related properties. Data analysis and
memo-writing become increasingly conceptual as the core
category and its properties, through constant comparison of
incident-category and category-category, become dense and
theoretically integrated (Glaser 1978). When the core category is
saturated – considered sufficiently dense and data collection no
longer generates new leads – theoretical sampling will cease
(Glaser & Strauss 1967).

Strauss and Corbin (1998) have broken down the theoretical
sampling process into stages of open sampling, relational and
variational sampling, and discriminate sampling, which
correspond directly with their stages of open, axial and selective
coding. According to Glaser (1992), this fracturing of the sampling
process offers the researcher little methodological help as all the
stages “occur anyway” (p.102). Indeed, the above outline of
theoretical sampling appears implicitly to parallel the open,
selective and theoretical coding stages in classic grounded theory.
Interestingly, Strauss and Corbin (1998) have received notably
more popularity within healthcare as a very direct result of the
‘help’ offered by a clear set of procedural steps. Similarly, Coyne
(1997) has noted that step by step guidance on theoretical
sampling may be useful for novice researchers. However, the
inherent risk within such a prescriptive approach to theory
generation is that creativity is stifled, and without creativity
there can only be limited conceptualisation (Glaser 2002b). While
as a novice researcher it is unnerving to trust in the emergent
nature of classic grounded theory, learning to be patient with the
data and remaining open to multiple possibilities will ultimately
generate a more relevant, and thus more useful, theory (Holton
2007). By adhering to strict procedures for collecting and
analysing data, the researcher is at risk of manipulating the data
rather than patiently allowing the theory to emerge inductively.
Boychuk-Duchscher & Morgan (2004) have captured this concern
aptly: “by focusing the researcher’s energies on the perfect
approach to finding the data, the true nature of the data may be
lost” (p.611). The theory should be grounded in the data, not in
the procedure.

Regardless of the debate about the usefulness of sampling
‘rules’, the more concerning distinction between classic grounded
theory and the method proposed by Strauss and Corbin (1998) is
the extent to which data are processed deductively. While the
open coding/sampling stages are notably similar within both
approaches – both involve sampling and coding for all
possibilities in the data – axial coding represents a significant
divergence from classic grounded theory. Axial coding involves
the application of a coding paradigm, otherwise known as the ‘6C’
coding family, to identify conditions, context, action/interactional
strategies, intervening conditions and consequences (Strauss &
Corbin 1998). Rather than allowing theoretical concepts to
emerge inductively, emerging concepts are tested against and
fitted deductively into this paradigm: “Strauss’ sampling is
controlled by the evolving relevance of concepts, and relevance
comes from testing out what is looked for, not what is emerging”
(Glaser 1992 p.103). Indeed, there exist several possible coding
families to explicate inter-relationships between categories
(Glaser 1978), none of which can be identified as relevant in
advance of the emerging theory. By pre-selecting the type of
theory they wish to generate, Strauss and Corbin (1998) have
effectively subverted the inductive nature of classic grounded
theory. An inductive approach requires that the theory emerges
after data collection begins, meaning that the researcher cannot
predict in advance the relevance of any one particular type of
data. As such, the constructivist revision of grounded theory
(Charmaz 2006) can be criticised for predetermining the lens
through which data are processed before data collection has even
begun. Glaser (1992) and Glaser and Holton (2004) have thus
contended that, rather than being grounded theory, these authors
have ‘remodelled’ the methodology as part of generic qualitative
data analysis.

The Use and Abuse of Theoretical Sampling

As a general methodology, classic grounded theory can use
either qualitative or quantitative data (Glaser 1978). Since its
inception, however, grounded theory has been embraced fervently
by qualitative researchers, ultimately leading to the dilution of
classic principles and erosion of the original methodology (Glaser
& Holton 2004). This dilution has been further exacerbated in
healthcare research, in which grounded theory ‘versions’ are
frequently confused or researchers have extracted particular
methods outwith the context of the original methodology.
Theoretical sampling in particular has become embroiled within
the multiple interpretations of sampling in qualitative research,
often being misconstrued as inter-changeable with purposeful
sampling (Sandelowski 1995). In Theoretical Sensitivity (Glaser
1978) sought to address this same concern, and thirty years later
this remains to be a notable problem. Ultimately, faced with
many ambiguous examples of the theoretical sampling process, it
is imperative to clarify and ‘demystify’ the distinction between
purposeful and theoretical sampling in order to prepare novice
researchers to produce trustworthy and credible grounded theory
research.

Hood (2007) has suggested that “all theoretical sampling is
purposeful, but not all purposeful sampling is theoretical” (p.158).
Purposeful sampling is defined as the selection of participants
with shared knowledge or experience of the particular
phenomena identified by the researcher as a potential area for
exploration (Sandelowski 1995). Typically, to ensure selection of
the most information rich participants, the researcher will
establish a set of inclusion or exclusion criteria based upon
research questions generated deductively from prior knowledge of
the area and a preliminary review of related literature. The
concern is with who or what to sample for the purpose of
answering questions about a predetermined topic. In contrast, the
selection of participants in theoretical sampling, and the reason
underpinning that selection, will change in accordance with the
theoretical needs of the study at any given time (Morse 2008).
Researchers using “theoretical sampling cannot know in advance
precisely what to sample for and where it will lead” (Glaser 1978
p.37). While a purposeful sample is selected at the outset of the
study for a predetermined purpose, theoretical sampling
progressively and systematically tailors data collection to serve
the emergent theory. Theoretical sampling is thus always
purpose-driven; the sample is selected for the purpose of
explicating and refining the emerging theory.

The Practical Realities of Theoretical Sampling

It has been clearly established that theoretical sampling is
guided by the emerging theory, and is concerned with where to
sample next and for what theoretical purpose. Yet for novice
researchers newly embarking upon a grounded theory study, the
most pressing practical concern is perhaps where to start. While
Glaser (1978) has advocated beginning the study with a sense of
‘abstract wonderment’, this poses a significant challenge for
researchers in the healthcare arena where detailed protocols are
required as a means of securing financial and ethical backing.
Furthermore, if the purpose of theoretical sampling is to seek
data that will contribute to developing categories of the emerging
theory, the researcher must surely first have the beginnings of a
theory – some tentative ideas – upon which to build. Evidently
there is an unavoidable need to begin somewhere. Dey (2007) has
cautioned researchers not to confuse an “open mind with an
empty head” (p.176). Initial ideas can benefit theoretical
development by providing a point of departure and by raising
important preliminary questions (Walker & Myrick 2006). Coyne
(1997) has explained that “the researcher must have some idea of
where to sample, not necessarily what to sample for, or where it
will lead” (p.625). In this sense theoretical sampling may involve
the purposeful selection of an initial starting point before moving
into theoretical sampling when data analysis begins to yield
theoretical concepts.

Beyond these initial decisions of where to start it is
impossible to anticipate the direction in which sampling will
proceed in advance of the emergence of a preliminary theoretical
framework (Glaser & Holton 2004). It is pertinent to remember
that the starting point is only that, and the researcher should
avoid formulating a preconceived conclusion that these initially
sampled characteristics will contribute to theoretical variation
(Glaser 1978). For example, to sample only according to
demographic characteristics is to deduce that they will be
relevant to the emerging theory (Glaser 1978; Morse 1991). It is
important to recognise that deductive logic does have a legitimate
place in classic grounded theory; themes emerge inductively from
the data but in following up these themes through further inquiry
the researcher is essentially engaged in a process of ‘deducing’
who or what to sample in order to do so (Dey 2007). Glaser (1978)
has referred to this deductive logic as ‘conceptual elaboration’
whereby theoretical possibilities and probabilities are deduced
from the emerging theory. However, because points of departure
such as demographic characteristics have not emerged from the
theory, they must be considered merely another variable awaiting
a verdict as to its relevance. Indeed, descriptive data may be
elevated into abstract theory only by way of comparing
theoretical categories and properties, not mere demographic
opposites (Hood 2007). Pre-existing knowledge can guide the
researcher in identifying a starting point for data collection, but
this knowledge should be awarded no relevance until validated or
dismissed by the formulation of the emerging theory. In the same
way as ideas must earn a way into the theory, the converse is also
true; it is possible that initial ideas will earn a way out.

Theoretical Saturation

For the novice grounded theorist, the initial concern about
where to start is often accompanied by a similar concern
regarding the decision to stop data collection. Given the inductive
nature of theory generation, it is understood that theoretical
sampling, including the point at which sampling will cease, is
controlled throughout the study by the emerging theory.
Sampling is discontinued once a point of saturation has been
reached, whereby categories and their properties are considered
sufficiently dense and data collection no longer generates new
leads (Glaser & Strauss 1967). Glaser (1992) has described this as
the point at which the researcher has reached the full extent of
the data, and thus “sampling is over when the study is over”
(p.107). While this definition carries a degree of simplicity,
theoretical saturation can be a difficult concept to understand,
particularly for first-time grounded theorists who are yet to
actually experience reaching the saturation point within a study.
Furthermore, much akin to ‘theoretical sampling’, the term
‘saturation’ has become somewhat ambiguous, ill-defined and
frequently misconstrued within the blurry boundaries of
qualitative research. It is imperative to understand, however,
that ‘saturation’ within generic qualitative data analysis and
‘saturation’ within classic grounded theory are inherently
different. While the qualitative researcher seeks descriptive
saturation, the grounded theorist is concerned with saturation at
a conceptual level.

Theoretical saturation is not mere descriptive redundancy.
That Glaser and Strauss (1967) have stipulated that categories be
sufficiently dense denotes an understanding that theoretical
saturation need not signal a point of complete coverage whereby
the researcher ‘knows everything’. Instead, theoretical sampling
does not aim for full descriptive coverage, but systematically
focuses and narrows data collection in the service of theoretical
development. While a predetermined, purposefully selected
sample might cause the researcher to worry if one has captured
enough relevant information, the theoretical sampling approach
assures relevance by progressively and systematically tailoring
data collection to serve the emergent theory (Glaser & Strauss
1967). In so doing, the grounded theorist is able to transcend the
descriptive level typical of qualitative research. By saturating
categories that seem to have the most explanatory power and
integrating these into and around a core variable, the grounded
theorist is able to present the theoretical essence of a substantive
area. Rather than presenting findings, debatably ‘accurate’ facts
or descriptions, grounded theory seeks only to present plausible
hypotheses that are grounded in the data (Glaser & Holton 2004).
While the saturation point indicates theoretical stability whereby
the core category accounts for as much variation in the data as
possible, it is crucial to understand that these concepts and
hypotheses are openly modifiable within the substantive area.
Saturation in classic grounded theory is thus neither concerned
with verifying hypotheses or exhausting the description of a
particular situation at a particular point in time. Instead, the
researcher should be concerned with generating a theory that can
cope with changing situations (a particularly important
consideration within the ever-changing healthcare arena) and
less with in-the-moment accuracy that has little temporal
transferability.

Writing up Theoretical Sampling

Ultimately, it is difficult to clarify or ‘demystify’ theoretical
sampling if researchers continue to misconstrue grounded theory
as a qualitative method and not a general methodology. This
article seeks to encourage novice researchers to be mindful that,
as a general methodology, grounded theory should not necessarily
be subject to generic ‘qualitative’ guidelines. For example, one
element of ‘trustworthy’ qualitative research is that researchers
provide a detailed description of participants (Curtin & Fossey
2007). For healthcare in particular, this is considered central to
evidence based practice; a sample that is described sufficiently
will enable the reader to transfer the research findings to a
particular context, allowing comparison between the evidence
presented in the research article and their own sphere of
experience (Curtin & Fossey 2007). From a grounded theory
perspective, however, there lies an inherent risk in the excessive
description of potentially irrelevant detail. This is of particular
concern in relation to the above discussion, whereby researchers
should not automatically assume the relevance of participants
socio-demographic characteristics to the emerging theory. While
demographic or social characteristics may provide a starting
point for data collection, by presenting a thick, isolated
description of participants at the start of a grounded theory
research article the researcher is at risk of either belying an
inappropriate approach to sampling, or obscuring the analytic
flow and progression of theoretical insights thus compromising
the credibility of an otherwise trustworthy study.

Morse (2008) has criticised the way in which theoretical
samples are presented as static without detailing and justifying
the selection and sequencing of the sampling process. Typically,
researchers provide a one-off description of participants in the
methods section of research articles, and ignore the impact of
sampling decisions made during analysis (Barbour 2001).
However, if the researcher does not capture the flow of the
theoretical sampling process, the complexities involved in the
development of the theory may be lost. Theoretical sampling is
intertwined inextricably with the abstraction of description into
theory, and is crucial to discovering and refining categories and
their properties and suggesting relationships between concepts.
Ultimately, the theoretical sampling ‘flow’ of moving back and
forth between data collection and analysis poses a challenge to
researchers writing up grounded theory studies; it is often
difficult to convey the chaos of research within the structure of an
article or thesis. However, sampling theoretically is “more
difficult than simply collecting data from a preplanned set of
groups, since choice requires continuous thought, action and
search” (Glaser & Strauss 1967 p.52). Studies that produce an
artificially neat and static account of the grounded theory process
serve only to obscure this complexity (Barbour 2001). Novice
grounded theorists should be careful to write-up a grounded
theory study in a manner that best reflects the methodology.
Grounded theory researchers should avoid isolated, one-off, static
descriptions of participants but should instead be challenged to
integrate within their write up the progression, justification and
contribution of sampling decisions so as to mirror the complex
and iterative process of theory development.

Evaluating Credibility

Theoretical sampling is theoretically oriented, and will thus
be different for every theory. There is no definitive checklist for
ensuring credibility, and the reader should be careful when
applying conventional guidelines of trustworthiness in qualitative
research to grounded theory studies. For example, the emphasis
on thick description in qualitative research has been
demonstrated to be potentially antithetical to the inductive
nature of grounded theory; sampling should be theoretically
directed as opposed to variable oriented and only those
descriptive characteristics that have a proven contribution to
theoretical variation within the theory should be included in the
write up. The adequacy of a theoretical sample should be judged
on the process of theory generation. Glaser and Strauss (1967)
stated that an inadequate theoretical sample would be evident in
a theory that is lacking integration and has too many remaining
gaps. It would seem then that transparency is a universal
concern, common to both grounded theory and qualitative
research; the credibility of a theory, or any piece of research,
cannot be dissociated from the process by which it is generated.
The ‘mysticism’ arises in grounded theory research when the
researcher fails to describe adequately the complex and messy
process of analytic abstraction whereby theory is developed from
empirical data. In this sense, to ensure that a grounded theory
study has credibility there must be evidence that the final
theoretical product is actually ‘grounded’. This should be achieved
by making the process through which theory has been developed
explicit within the final write up, paying particular attention to
capturing the flow of theoretical sampling which will demonstrate
and explain the build up of theoretical insights into abstract
theory.

Conclusion

For the healthcare researcher, classic grounded theory offers
an inductive methodology with a distinctly practical purpose: to
provide a theory that has the potential to explain, interpret and
guide practice. However, the full potential of grounded theory can
only be realised through sound application of its distinct
methodological principles, most notably theoretical sampling.
Although grounded theory has evolved and diversified since its
inception, the emphasis on theoretical sampling as being
essential to the analytic abstraction of theory has remained
largely undisputed. Despite this apparent agreement, however, it
has been demonstrated that by pre-determining the type of data
sought or looking for a specific paradigm in the data, other
versions of ‘grounded theory’ seek only to subvert the inductive
nature of classic grounded theory. Furthermore, there is wide
evidence of inappropriate use and documentation of theoretical
sampling within healthcare literature, resulting from the
misconceptions regarding the methodological nature of classic
grounded theory. As a result, grounded theory studies have been
accused of mysticism, whereby codes and categories appear as if
out of nowhere. Novice and experienced grounded theory
researchers alike are thus encouraged to ‘demystify’ their
theoretical sampling processes, making explicit the steps taken to
build up theoretical insights. Researchers should capture the
complex flow of sampling for the purposes of theory development
by integrating key sampling decisions and justifications within
the write up of their studies. However, researchers should also be
wary of overly thick description of the sample; descriptive
characteristics may provide an adequate starting point however
these must not be awarded any assumed relevance until
validated or dismissed by the emerging theory. As a general
methodology, novice researchers should beware appraising
grounded theory on the basis of generic qualitative guidelines.
Novice researchers are encouraged to develop a sound
understanding of the theoretical sampling process in order to
ensure the credibility of one’s own studies, and to appraise that of
others’.

Authors

Jenna Breckenridge BSc(Hons), Ph.D. Candidate
Derek Jones, PhD, BA (Hons), Dip COT
School of Health Sciences
Queen Margaret University
Edinburgh, Scotland
E-Mail: jbreckenridge@qmu.ac.uk

References

Barbour, R. S. (2001) Checklists for Improving Rigour in
Qualitative Research: A Case of the Tail Wagging the
Dog? British Medical Journal, 322,1115-1117.

Boychuk-Duchsher, J. E. & Morgan, D. (2004) Grounded Theory:
Reflections of the Emergence vs. Forcing Debate. Journal
of Advanced Nursing
, 48, 605-612.

Charmaz, K. (2006) Constructing Grounded Theory: A Practical
Guide Through Qualitative Analysis
. London: Sage Pub.
Ltd.

Coyne, I. T. (1997) Sampling in Qualitative Research. Purposeful
and Theoretical Sampling; Merging or Clear Boundaries?
Journal of Advanced Nursing, 26, 623-630.

Curtin, M. & Fossey, E. (2007) Appraising the Trustworthiness of
Qualitative Studies: Guidelines for Occupational
Therapists. Australian Occupational Therapy Journal,
54, 88-94.

Dey, I. (2007) Grounding Categories. in: Bryant, A., Charmaz, K.
(Eds.). The Sage Handbook of Grounded Theory (pp.167-
190). London: Sage Pub. Ltd.

Draucker, C. B., Martsolf, D. S., Ross, R. & Rusk, T. B. (2007)
Theoretical Sampling and Category Development in
Grounded Theory. Qualitative Health Research, 17, 1137-
1148.

Glaser, B. G. (1978) Theoretical Sensitivity. California: Sociology
Press.

Glaser, B. G. (2002a) Constructivist Grounded Theory?
Qualitative Social Research, 5. Retrieved February 18,
2008, from http://www.qualitative-research.net/fqs/fqseng.
htm

Glaser, B. G. (2002b) Conceptualisation: On Theory and
Theorising Using Grounded Theory. International
Journal of Qualitative Methods
, 1, 1-31. Retrieved
February 18, 2008, from http://www.ualberta.ca/~ijqm/

Glaser, B. G. & Holton, J. (2004) Remodelling Grounded Theory.
Forum: Qualitative Social Research, 5. Retrieved
February 18, 2008, from http://www.qualitativeresearch.
net/fqs-texte/2-04/2-04glaser-e-htm

Glaser, B., Strauss, A. (1967) The Discovery of Grounded Theory:
Strategies for Qualitative Research
. New York: Aldine
Pub. Co.

Holton, J. A (2007) The Coding Process and Its Challenges. in:
Bryant, A., Charmaz, K. (Eds.) The Sage Handbook of
Grounded Theory
(pp.265-290). London: Sage Pub. Ltd.

Hood, J. C. (2007) Orthodoxy vs. Power: The Defining Traits of
Grounded Theory. in: Bryant, A., Charmaz, K. (Eds.) The
Sage Handbook of Grounded Theory
(pp.151-164).
London: Sage Pub. Ltd.

Melia, K. M. (1997) Producing ‘Plausible Stories’: Interviewing
Student Nurses. in: Miller, G., Dingwall, R. (Eds.) Context
and Method in Qualitative Research
(pp.26-36). London:
Sage Pub. Ltd.

Morse, J. M. (2008) “What’s your favourite colour?” Reporting
Irrelevant Demographics in Qualitative Research.
Qualitative Health Research, 18, 299-300.

Morse, J. M. (1991) Strategies for Sampling. in: Morse, J. M. (Ed.)
Qualitative Nursing Research: A Contemporary Dialogue
(pp.127-146). London: Sage Pub. Ltd.

Sandelowski, M. (1995) Sample Size in Qualitative Research.
Research in Nursing and Health, 18, 179-183.

Strauss, A. L. & Corbin, J. (1998) Basics of Qualitative Research:
Techniques and Procedures for Developing Grounded
Theory
(2nd ed.), London: Sage Pub. Ltd.

Walker, D. & Myrick, F. (2006) Grounded Theory: An Exploration
of Process and Procedure, Qualitative Health Research 16,
547-559.

Facebooktwittergoogle_plusredditpinterestlinkedinmail