Vous êtes sur la page 1sur 102

QUALITATIVE RESEARCH SERIES

Qualitative research series

Qualitative research: standards, challenges, and guidelines


Kirsti Malterud
Qualitative research methods could help us to improve our understanding of medicine. Rather than thinking of
qualitative and quantitative strategies as incompatible, they should be seen as complementary. Although procedures
for textual interpretation differ from those of statistical analysis, because of the different type of data used and
questions to be answered, the underlying principles are much the same. In this article I propose relevance, validity,
and reflexivity as overall standards for qualitative inquiry. I will discuss the specific challenges in relation to reflexivity,
transferability, and shared assumptions of interpretation, which are met by medical researchers who do this type of
research, and I will propose guidelines for qualitative inquiry.
A broad base of medical and scientific knowledge is needed
if medicine is to maintain its identity as a discipline founded
on scientific knowledge. However, interpretive action must
also be included in medical knowledge. In my first article,1 I
investigated the nature of clinical knowledge in medicine,
exposed some of the shortcomings of quantitative research
methods, and briefly introduced qualitative methods as an
approach for improved understanding. Here, I shall discuss
how scientific quality can be maintained when qualitative
research methods are applied. I present some overall
standards, describe specific challenges met when the
medical researcher uses qualitative research methods, and
subsequently propose guidelines for qualitative inquiry in
medical research. I do not intend to provide comprehensive
guidance for the inexperienced qualitative researcher, who
must be prepared to acquire basic skills of qualitative
research from the relevant literature. Some of the specific
terms that I use are presented in panel 1.

Standards
Qualitative research methods involve the systematic
collection, organisation, and interpretation of textual
material derived from talk or observation. It is used in the
exploration of meanings of social phenomena as
experienced by individuals themselves, in their natural
context.25 Qualitative research is still regarded with
scepticism by the medical community, accused of its
subjective nature and the absence of facts. Although the
adequacy of guidelines has been vigorously debated within
this cross-disciplinary field,6,7 scientific standards, criteria,
and checklists do exist.3,811 However, as Chapple and
Rogers12 point out, medical researchers often encounter
difficulties when they try to apply guidelines designed by
social scientists, which deal with issues important in their
own discipline, but which are not necessarily generically
valid as scientific standards.
Hamberg and colleagues,13 for example, claim that the
established criteria for scientific rigour in quantitative
research cannot be applied to qualitative studies. Referring
to Lincoln and Guba,2 they suggest alternative criteria:
credibility, dependability, confirmability, and transLancet 2001; 358: 48388
Section for General Practice, Department of Public Health and
Primary Health Care, University of Bergen, Ulriksdal 8C, N-5009
Bergen, Norway; and Department of General Practice and Research
Institute, University of Copenhagen, Denmark (Prof K Malterud MD)
(e-mail: kirsti.malterud@isf.uib.no)

THE LANCET Vol 358 August 11, 2001

ferability. They admit that these criteria correspond with


traditional ones in some ways, comparing credibility with
internal validity, confirmability with objectivity, and
transferability with generalisability.
Mays and Pope,7 however, maintain that qualitative
research can be assessed with reference to the same broad
criteria as quantitative research, albeit used in a different
way. Referring to Hammersley,14 they suggest that validity
and relevance are essential. Neither of these criteria are
straightforward to assess though, and each requires
judgments to be made. To improve validity, Mays and
Pope7 suggest procedures and principles such as
triangulation, respondent validation, clear detailing of
methods of data collection and analysis, reflexivity,
attention to negative cases, and fair dealing. Relevance can
be increased by the use of detailed reports and sampling
techniques. The importance of clinical relevance has also
been emphasised by Giacomini and Cook.15
I believe that qualitative research methods are founded
on an understanding of research as a systematic and
reflective process for development of knowledge that can
somehow be contested and shared, implying ambitions of
transferability beyond the study setting. Drawing on these
assumptions, the researcher must be prepared to use
strategies for: questioning findings and interpretations,
instead of taking them for granted; assessing their internal
and external validity, instead of judging them obvious or
universal; thinking about the effect of context and bias,
without believing that knowledge is untouched by the
human mind; and displaying and discussing the processes
of analysis, instead of believing that manuals grant
trustworthyness. Agreeing with Hammersley,14 and
Giacomini and Cook,15 I believe relevance and validity are
essential standards, but think of reflexivity as an equally
important measure, which should be added to the criteria.

Specific challenges
Although there are many similarities between qualitative
and quantitative research methods, some procedures are
very different, because of the different nature and
assumptions of the data and questions to be answered. The
effect of an investigator on a study, the principles and
consequences of sampling, and the process of organisation
and interpretation during analysis, all affect research, and
are closely related to different aspects of validity (panel 2).
Reflexivity
A researchers background and position will affect what
they choose to investigate, the angle of investigation, the

483

For personal use. Only reproduce with permission from The Lancet Publishing Group.

QUALITATIVE RESEARCH SERIES

Panel 1: Terms used in qualitative research


Term

Metaphor

Description

Reflexivity

The knowers mirror

An attitude of attending systematically to the context of knowledge construction,


especially to the effect of the researcher, at every step of the research process

Preconceptions

The researchers
backpack

Previous personal and professional experiences, prestudy beliefs about how things
are and what is to be investigated, motivation and qualifications for exploration of
the field, and perspectives and theoretical foundations related to education and
interests

Theoretical frame
of reference

The analysts reading


glasses

Theories, models, and notions applied for interpretation of the material and for
understanding a specific situation

Metapositions

The participating
observers sidetrack

Strategies for creating adequate distance from a study setting that you are
personally involved in

Transferability

External validity

The range and limitations for application of the study findings, beyond the context
in which the study was done

methods judged most adequate for this purpose, the


findings considered most appropriate, and the framing and
communication of conclusions. Contemporary theory of
knowledge acknowledges the effect of a researchers
position and perspectives, and disputes the belief of a
neutral observer.16 Haraway17 claims that the perspective of
the observer is always limited and determines what can be
seen. This notion applies even in laboratory science.18
Hence, in qualitative (and maybe also in quantitative)
inquiry, the question is neither whether the researcher
affects the process nor whether such an effect can be
prevented. This methodological point has been turned into
a commitment to reflexivity. The illusion of denying the
human touch is countered by establishing an agenda for
assessment of subjectivity. Objectivity, redefined by
Haraway,17 means to recognise that knowledge is partial
and situated, and to account adequately for the effects of
the positioned researcher. During all steps of the research
process, the effect of the researcher should be assessed, and,
later on, shared. Adequate accounts of these effects should
be presented in the publication, as the frame of discussions
of limitations and strengths of the study, and transferability
of findings.19 Bias, in the sense of undesirable or hidden
skewness, is thus accounted for, though not eliminated.
Subjectivity arises when the effect of the researcher is
ignored.
Dependent on positions and perspectives, different
researchers might therefore access different, although
equally valid, representations of the situation that is
studied. In qualitative research, these different ways of
approaching the same subject result in an increased
understanding of complex phenomena, not in a failure of
reliability. Multiple researchers might strengthen the design
of a studynot for the purpose of consensus or identical
readings, but to supplement and contest each others
statements. The single researcher will have to establish
other strategies for broad and critical reading. Validation by
consensus or repeatability is seldom adequate in qualitative
research.

Panel 2: Factors that affect research


Reflexivity

Share preconceptions
Establish metapositions

Transferability

Adequate and sufficiently varied sample


Consider whom and what the findings
concern

Interpretation
and analysis

484

Describe theoretical frame of reference


Transparent, systematic procedure

The investigator always enters a field of research with


certain opinions about what it is all about.20 Reflexivity
starts by identifying preconceptions brought into the
project by the researcher, representing previous personal
and professional experiences, prestudy beliefs about how
things are and what is to be investigated, motivation and
qualifications for exploration of the field, and perspectives
and theoretical foundations related to education and
interests. Miller,21 for instance, writes about the
inquisitiveness he felt towards his colleagues capacity to
combine efficiency and a biopsychosocial orientation,
whereas Gardner and Chapple22 introduce Gardners
distress as a general practitioner trying to relate to a patient
with angina who impeded referral for 6 years.22 In
qualitative study, researchers commonly claim that they
develop hypotheses, they do not test them. In a scientific
culture accustomed to specific procedures for hypothesis
testing, such claims are useful for rhetorical purposes, to
prevent expectations about identical procedures applied to
qualitative material. The researcher should not deny that
hypotheses exist. However, the qualitative researchers task
is to explain, and maybe question, the hypotheses as
ingredients of the preconceptions and as reflections,23 rather
than applying procedures for testing them.
Preconceptions are not the same as bias, unless the
researcher fails to mention them. If reflexivity is thoroughly
maintained, personal issues can be valuable sources for
relevant and specific research. However, the investigator
should take care not to confuse knowledge intuitively
present in advance, embedded in preconceptions, with
knowledge emerging from inquiry of systematically
obtained material. This situation can be avoided by
declaration of beliefs before the start of the study.
Reflexivity can also be maintained by looking at the data, or
its interpretation, for competing conclusions. In a study, in
which we asked patients to keep a diary, Stensland and
myself24 suggest strategies for obtaining metapositions as a
way of improving reflexivity, whereas Miller21 recommends
that data be taped and transcribed, therefore allowing
others not involved in the study to audit them. Based on a
review of 29 publications in which qualitative methods were
applied, Hoddinott and Pill25 concluded that important
contextual details were often missing, implying that critical
appraisal of the reports was hampered.
Transferability
The importance of sampling is closely related to validity.
Internal validity asks whether the study investigates what it
is meant to, whereas external validity asks in what contexts
the findings can be applied. The nature and extent of the

THE LANCET Vol 358 August 11, 2001

For personal use. Only reproduce with permission from The Lancet Publishing Group.

QUALITATIVE RESEARCH SERIES

data will ascertain which conclusions can be drawn about


what. The aim of research is to produce information that
can be shared and applied beyond the study setting. No
study, irrespective of the method used, can provide findings
that are universally transferable. The study design should
show a thorough consideration of what an adequate degree
of transferability would be, in view of the assumptions of
the research question, and present a relevant sampling
strategy. Sampling strategies might seem fundamentally
different for qualitative and quantitative inquiry. The key to
understanding how these different approaches still
accommodate scientific quality is to move beyond
procedures and to keep the principles of the research in
mind.
Purposeful or theoretical sampling are commonly done
to obtain qualitative material.26 Previous experience and
theoretical frameworks will indicate where to go for
resourcesie, Skelton and colleagues27 designed a study to
investigate the heterogeneity of practices and of a wide

range of practitioners, and Stensland and I24 adapted our


design to investigate patients with longstanding symptoms
without clinical findings, representing varying duration of
illness and symptom presentation.24 Sampling is usually
done in a stepwise way, including more data from one
group or another dependent on what extra material is
needed to answer the research question effectively. A
discussion about who and what the findings actually relate
to is a key component of external validation in a qualitative
study.23,28
The procedures described are fundamentally different
from those used to deal with prevalences, distributions, or
numerical differences, in which large representative or
random samples, allowing for calculations of probability
with subsequent inference to a defined population, are
required. In qualitative inquiry, the aim with respect to
external validity is to ascertain whether or not the study
hypothesis or results can be applied in other settings.
Presentation of contextual background material, such as

Panel 3: Guidelines for authors and reviewers of qualitative studies


Aim
Is the research question a relevant issue?
Is the aim sufficiently focused, and stated clearly?
Does the title of the article give a clear account of the aim?
Reflexivity
Are the researcher's motives, background, perspectives, and preliminary hypotheses presented, and is the effect of these issues
sufficiently dealt with?
Method and design
Are qualitative research methods suitable for exploration of the research question?
Has the best method been chosen with respect to the research question?
Data collection and sampling
Is the strategy for data collection clearly stated (usually purposive or theoretical, usually not random or representative)?
Are the reasons for this choice stated?
Has the best approach been chosen, in view of the research question?
Are the consequences of the chosen strategy discussed and compared with other options?
Are the characteristics of the sample presented in enough depth to understand the study site and context?
Theoretical framework
Are the perspectives and ideas used for data interpretation presented?
Is the framework adequate, in view of the aim of the study?
Does the author account for the role given to the theoretical framework during analysis?
Analysis
Are the principles and procedures for data organisation and analysis fully described, allowing the reader to understand what happened
to the raw material to arrive at the results?
Were the various categories identified from theory or preconceptions in advance, or were they developed from the data?
Which principles were followed to organise the presentation of the findings?
Are strategies used to validate results presented, such as cross-checks for rivalling explanations, member checks, or triangulation.
If such strategies are not described in this section, they should appear as validity discussions later in the report.
Findings
Are the findings relevant with respect to the aim of the study?
Do they provide new insight?
Is the presentation of the findings well organised and best suited to ensure that findings are drawn from systematic analysis of
material, rather than from preconceptions?
Are quotes used adequately to support and enrich the researcher's synopsis of the patterns identified by systematic analysis?
Discussion
Are questions about internal validity (what the study is actually about), external validity (to what other settings the findings or notions
can be applied), and reflexivity (the effects of the researcher on processes, interpretations, findings, and conclusions) addressed?
Has the design been scrutinised?
Are the shortcomings accounted for and discussed, without denying the responsibility of choices taken?
Have the findings been compared with appropriate theoretical and empirical references?
Are a few clear consequences of the study proposed?
Presentation
Is the report easy to understand and clearly contextualised?
Is it possible to distinguish between the voices of the informants and those of the researcher?
References
Are important and specific sources in the field covered, and have they been appropriately presented and applied in the text?

THE LANCET Vol 358 August 11, 2001

485

For personal use. Only reproduce with permission from The Lancet Publishing Group.

QUALITATIVE RESEARCH SERIES

demographics and study setting, is necessary if the reader is


to be able to ascertain for which situations the findings
might provide valid information. The pursuit of diversity
and contradictions in interpretive analysis of textual
material is not logically compatible with the standardisation
assumptions underlying probability statistics. Additionally,
the findings are not supposed to be valid for population
groups at large. Random sampling is therefore rarely a
relevant tool for validity in these studies
Good qualitative research does not exaggerate the extent
of the material. During analysis the researchers should have
a thorough knowledge of the study material, so that they are
aware of the content of the data and what they mean, and
so that they are able to ascertain what in the material is
relevant when trying to answer the research question.
Computer programs are useful for storing, ordering, and
retrieving information, but they cannot do the analysis
itself. Additionally, a large amount of material does not
actually guarantee transferability, and might result in a
superficial analysis, since the researchers are not able to test
reflexivity and look at counterhypotheses. The transcripts
from 15 patients and their four doctors, therefore, might be
more than sufficient.22 The nature of the research question
and the material, combined with the intention of external
validity, will determine the correct number of participants
for a study. One individual, as in a case study,29 might be
sufficient dependent on the topic and scope of the
investigation.21,23,30 The findings from a qualitative study are
not thought of as facts that are applicable to the population
at large, but rather as descriptions, notions, or theories
applicable within a specified setting.
Interpretation and analysis
A thorough, well prepared, and well documented analysis is
what distinguishes scientific approach from superficial
conjecture (panel 3). The researchers task is to organise,
compare, and validate alternative interpretations. Only
when the researcher can identify the systematic procedure
that has been followed in this process, can it be shared with
others.31 Declaring that qualitative analysis was done, or
stating that categories emerged when the material had been
read by one or more persons, is not sufficient to explain
how and why patterns were noticed.
Qualitative data represent large amounts of information,
and analysis implies abstraction and some degree of
generalisation. Components from the individual
informants history and expressions are used to gain
knowledge applicable to others. Analysis of qualitative data
involves decontextualisation and recontextualisation.32
Decontextualisation allows parts of the subject matter to be
lifted out and investigated more closely, together with other
elements across the material that tells about similar issues.
Recontextualisation will make sure that the patterns still
agree with the context from which they were collected, and
is important to prevent reductionism and to maintain the
connections between the field and the informants accounts
of reality.
The processes of systematic analysis of qualitative
data vary from project to project, dependent on the
research question, material, and choice of analytical
style.3 Miller and Crabtree33 present three styles of analysis,
according to the degree of predetermined or theoretically
founded categories for interpretation. With the
immersion/crystallisation (intuitive) analysis style, the
researcher organises data by examining the text thoroughly
and then crystallising out the most important aspects.21
With the editing (data-based) analysis style, the researcher
identifies units in the text, forming the basis for datadeveloped categories, which are used to reorganise the text

486

so that its meaning can be clearly seen.28 With the template


(theory-based) analysis style (not very frequently applied in
medical research), the text is organised according to preexisting theoretical or logical categories, to provide new
descriptions of previously known phenomena. A researcher
should always reveal the style of analysis used.
Interpretation is an integral part of qualitative inquiry.
The qualitative researcher might aim for induction, in the
sense of development of theory from data. However,
knowledge never emerges from data alone, but from the
relation between empirical substance and theoretical
models and notions.33 The theoretical framework can be
equated with the reading glasses worn by the researcher
when she or he asks questions about the material. Sharing
the type and role of framework is essential to maintain
communicative validity.5 A frequent shortcoming in report
writing is to omit information about whether the presented
categories represent empirical findings or if they were
identified in advance.
Neglect of the theoretical considerations does not
enhance the scientific quality of any study. Yet, the medical
researcher is not supposed to become a social scientist, even
when doing qualitative inquiry. Different degrees of
theoretical thoroughness are relevant for different purposes.
The medical researcher is advised to draw on theory from
other disciplines, yet to maintain the ambition of
constructing medical knowledge. Investigators should be
encouraged to declare that their readings or interpretations
have been supported, for instance, by models about self
efficacy,26 health belief,22 or proinflammatory cytokines,28
without being expected to permeate these thoroughly.
However, the task of transforming theory from other
disciplines so that it is applicable to medicine will require
some in depth research, as exemplified by Nessa,30 who has
developed a method for transcription of consultations from
pragmatics and textlinguistics.
Researchers who claim that they approach their material
inductively, without applying any theory for analysis, fail to
realise that their stance is unavoidably affected by theory.
This notion is particularly true for those working in
medicine, where the role of the theoretical framework is
seldom explicitly spelled out. Failure to acknowledge the
effect of theory might be a major threat to objectivity, since
notions and models used in interpretation of data are
always derived from a theory of some sort. Clarification and
declaration of the standpoints by a researcher, instead of
denial or hiding of the frame of reference, will enhance
intersubjectivity, in quantitative as well as qualitative
inquiry.
A medical researcher might find the task of condensing
their research to fit the limits of a journal article, without
compromising quality, difficult. To overcome this dilemma,
investigators, referees, and editors need to work together.
Because a range of procedures can be applied in qualitative
analysis, a transparent description of the path from data to
findings is necessary to convey what was done to the reader.
Clarification of the role of different data sources is an
important part of this description.21,24,28 To indicate that a
computer program was used for analysis is just as
insufficient as saying that SPSS was applied, without stating
the type of statistics involved. Furthermore, a researcher
cannot simply say that the material was coded for typical
patterns, resulting in some categories; the reader needs to
know the principles and choices underlying pattern
recognition and category foundation. The balance between
flexibility and rigidity is a demanding challenge in creative
qualitative analysis. Yet, reference to a previous, well
described procedure can satisfy the reader and save
precious words in a journal article. Unfortunately, word

THE LANCET Vol 358 August 11, 2001

For personal use. Only reproduce with permission from The Lancet Publishing Group.

QUALITATIVE RESEARCH SERIES

limits in medical journals might restrain development and


dissemination of innovative or complex qualitative research
methods.
Two basic methods of data analysis are commonly
referred to in the literature. Grounded theory, based on the
theory of social interactionism, is used to develop social
issues and theories grounded in qualitative data. Strauss
and Corbin35 provide specific procedures for analysis,
including open coding, axial coding, and selective coding.
Their constant comparative method can be applied to
approach a core category and a storyline as the main
outcomes of analysis. Giorgis analysis,36 based on
phenomenological philosophy, and modified by myself,31 is
suited for development of descriptions and notions related
to human experience.24,28 Giorgi recommends a four-step
analysis procedure: getting a total impression, identifying
meaning units, abstracting the contents of individual
meaning units, and summarising their importance.
Analysis might also be presented as a narrative.21,30,37 An
investigator often considers many factors before answering
the research question, and the reader should be aware of
them all. However, the more intuitively the analysis
procedure is accomplished, the harder it is to account for
what has been done. The beginner is therefore advised to
follow a path that has been trodden by others, even though
the more artistic potentials of analysis might then be traded
off for a more mechanical, but transparent, approach. The
experienced researcher, however, might move more freely
in the material without losing hold of the process that is to
be accounted for.23

differences cannot be inferred from this kind of material.


Correspondingly, the search for meaning and experience in
responses constructed by the researcher in advance, is a
risky business.
Accordingly, the principles of meta-analysis should be
thoroughly reconsidered when qualitative and quantitative
studies are analysed together. Complete integration is not a
realistic objective. In the context of medical research,
integration of methods invariably denotes treating the
qualitative study as if it were a quantitative one, recording
the material as variables, which are counted and
aggregated. Healthy and innovative meta-analysis should
develop methods for reasonable combination of findings
from qualitative and quantitative studies, acknowledging
and using the potential of the different nature of these
approaches. Interpretation of textual materials and
purposeful samples is different to the calculation of
numerical materials and random samples. Findings from
qualitative and quantitative studies can certainly be
aggregated and complemented by secondary analysis,
contributing to an extended approach to the phenomenon
in question, as well as a mutual validation. However, such
meta-analysis should be done on the results, and not by
accumulating and mixing quantitative and qualitative data,
which require fundamentally different procedures for
scientific analysis. When combining qualitative and
quantitative studies, the meta-analyst should be prepared to
handle contradictory findings, without having to discard
one and appoint the other as the gold standard.

Conclusions
Qualitative and quantitative methods
When qualitative and quantitative approaches are
combined, the methods are often applied in sequential
order. Semistructured interviews or observational data
might, for example, be used to explore hypotheses or
variables when planning a large epidemiological study,
resulting in enhanced sensitivity and accuracy of survey
questions and statistical strategy. In such instances,
qualitative studies might be thought of as precursors of
real science. However, qualitative studies can also be
added to quantitative ones, to gain a better understanding
of the meaning and implications of the findings. More
creative combinations are seen in triangulation.3 The idea
of triangulation originated from a craft used by land
surveyors, who increase the validity of a map by
incorporating measures from different angles. Multiple and
diverse observations can enrich the description of a
phenomenonie, an elephant looks very different when
seen from above or below. Someone reading a report might
gain a better understanding of what goes on in a medical
consultation if data from various sources, such as doctors
and patients,22 have been combined. The aim of
triangulation is to increase the understanding of complex
phenomena, not criteria-based validation, in which
agreement among different sources confirms validity.
Quantification of phenomena or categories can be done
to gain an overview of qualitative material, but the
application of such numbers should be done with caution.
Quasistatistical analysis of textual material, also termed
content analysis, has gained some popularity, and computer
programs are available to count the occurrence of specific
words or utterings in a text. However, the scientific logic of
statistics and transferability is far from accomplished in a
non-representative sample in which questions were not
asked in a standardised way to all participants. We do not
know to whom the findings can be transferred, and we do
not know the potential answers from informants who just
did not mention the issue. Prevalences, distributions, and

THE LANCET Vol 358 August 11, 2001

Medical research needs diversity. We need to prevent


methodological separatism and supremacy if the field of
medical knowledge is to be expanded, not just strengthened
or divided. Responsible application of qualitative research
methods is a promising approach to broader understanding
of clinical realities. No research method will ever be able to
describe peoples lives, minds, and realities completely
though, and medical doctors should be reminded that
scientific knowledge is not always the most important or
relevant type of information when dealing with people.
References
1
2
3
4
5
6
7

8
9
10
11

12

13

Malterud K. The art and science of clinical knowledge: evidence beyond


measures and numbers. Lancet 2001; 358: 397400.
Lincoln YS, Guba EG. Naturalistic inquiry. Newbury Park: Sage
Publications, 1985.
Miles MB, Huberman AM. Qualitative data analysis: an expanded
sourcebook, 2nd edn. Thousand Oaks: Sage Publications, 1994.
Britten N, Jones R, Murphy E, Stacy R. Qualitative research methods in
general practice and primary care. Fam Pract 1995; 12: 10414.
Kvale S. InterViews: an introduction to qualitative research writing.
Thousand Oaks: Sage Publications, 1996.
Lincoln YS. Emerging criteria for quality in qualitative and interpretive
research. Qual Inq 1995; 1: 27589.
Mays N, Pope C. Quality in qualitative health research. In: Pope CP,
Mays N, eds. Qualitative research in health care, 2nd edn. London:
BMJ Books, 2000: 89101.
Blaxter M. Criteria for evaluation of qualitative research. Med Sociol
News 1996; 22: 6871.
Elder NC, Miller WL. Reading and evaluating qualitative research
studies. J Fam Pract 1995; 41: 27985.
Mays N, Pope C. Rigour and qualitative research. BMJ 1995; 311:
10912.
Giacomini MK, Cook DJ. Users guides to the medical literature,
XXIII: qualitative research in health careA, are the results of the study
valid? JAMA 2000; 284: 35762.
Chapple A, Rogers A. Explicit guidelines for qualitative research: a step
in the right direction, a defence of the soft option, or a form of
sociological imperialism? Fam Pract 1998; 15: 55661.
Hamberg K, Johansson E, Lindgren G, Westman G. Scientific rigour in
qualitative research: examples from a study of womens health in family
practice. Fam Pract 1994; 11: 17681.

487

For personal use. Only reproduce with permission from The Lancet Publishing Group.

QUALITATIVE RESEARCH SERIES

14 Hammersley M. Reading ethnographic research. New York: Longman,


1990.
15 Giacomini MK, Cook DJ. Users guide to the medical literature, XXIII:
qualitative research in health careB, what are the results and how do
they help me care for my patients? JAMA 2000; 284: 47882.
16 Nagel T. The view from nowhere. New York: Oxford University Press,
1986.
17 Haraway D. Situated knowledges: the science question in feminism and
the privilege of partial perspective. In Haraway D. Simians, cyborgs, and
women: the reinvention of nature. New York: Routledge, 1991:
183201.
18 Latour B, Woolgar S. Laboratory life: the construction of scientific facts.
Princeton: Princeton University Press, 1986.
19 van Maanen J. Tales of the field. Chicago; University of Chicago Press,
1988.
20 Giorgi A.The role of observation and control in laboratory and field
research settings. Phenomenology Pedagogy 1986; 4: 2228.
21 Miller WL. Routine, ceremony, or drama: an exploratory study of the
primary care clinical encounter. J Fam Pract 1992; 34: 28996.
22 Gardner K, Chapple A. Barriers to referral in patients with angina:
qualitative study. BMJ 1999; 319: 41821.
23 Malterud K, Hollnagel H. Encouraging the strengths of women
patients: a case study from general practice on empowering dialogues.
Scand J Public Health 1999; 27: 25459.
24 Stensland P, Malterud K. New gateways to dialogue in general practice:
development of an illness diary to expand communication. Scand J Prim
Health Care 1997; 15: 17579.
25 Hoddinott P, Pill R. A review of recently published qualitative research
in general practice: more methodological questions than answers? Fam
Pract 1997; 14: 31319.

26 Kuzel AJ. Sampling in qualitative inquiry. In: Crabtree BF,


Miller WL, eds. Doing qualitative research, 2nd edn. Thousand Oaks:
Sage Publications, 1999: 3345.
27 Skelton AM, Murphy EA, Murphy RJL, ODowd TC. General
practitioner perceptions of low back patients. Fam Pract 1995; 12:
4448.
28 Sderlund A, Skoge AM, Malterud K. I could not lift my arm holding
the fork...: living with chronic fatigue syndrome. Scand J Prim Health
Care 2000; 18: 16569.
29 Yin RK. Case study research: design and methods. Newbury Park: Sage
Publications, 198489.
30 Nessa J. From a medical consultation to a written text. Scand J Prim
Health Care 1995; 13: 8392.
31 Malterud K. Shared understanding of the qualitative research
process: guidelines for the medical researcher. Fam Pract 1993; 10:
20106.
32 Tesch R. Qualitative research: analysis types and software tools. New
York: Falmer Press, 1991.
33 Miller WL, Crabtree BF. Clinical research: a multimethod typology
and qualitative roadmap. In: Crabtree BF, Miller WL, eds. Doing
qualitative research, 2nd edn. Thousand Oaks: Sage Publications, 1999:
330.
34 Morse JM. The power of induction. Qual Health Res 1992; 2: 36.
35 Strauss A, Corbin J. Basics of qualitative research: grounded theory
procedures and techniques. Newbury Park: Sage Publications, 1990.
36 Giorgi A. Sketch of a psychological phenomenological method. In:
Giorgi A, ed. Phenomenology and psychological research. Pittsburgh:
Duquesne University Press, 1985: 822.
37 Riessman C. Narrative analysis. Newbury Park: Sage Publications,
1993.

Uses of error: Surprises in diagnosis


Working through the night back in the mid-1960s, I was trying to break what we
thought was a lupus flare in a Hispanic woman who had entered the hospital 2 weeks
before with fever of unknown origin. She had been worked up for various bacterial
and viral diseases with the usual chest radiograph, blood, urine, and cerebrospinal
fluid microscopy and cultures. No infectious agent was found. The leading diagnosis
was a flare-up from her lupus erythematosus. High doses of steroid were given and
although her fever subsided she still had weakness and signs of her autoimmune
disease during her course at the hospital. That night she was especially compromised
with difficulty in breathing and high fever. Despite high steroid doses and close
attention by the staff, she died.
As a group, the house staff was not satisfied with the diagnosis of lupus
erythematosus on the death certificate. We thought some other process must have
taken place, but that we had ruled out all the obvious possible causes. She had a
slightly enlarged liver and spleen, again believed secondary to her autoimmune
disease. We were able to do a liver biopsy in an attempt to uncover some unexpected
disease.
As often happens, she died on a Friday and we had the weekend to wait in
anticipation. Finally, the pathology report came back: Mycobacterium tuberculosis
throughout the liver. The surprise of this obvious diagnosis was a lesson to us all. Of
course, with high-dose steroids, tuberculosis should have been high on our list, but
she had not given us any indication of tuberculosis. Her lungs had been clear on
admission, and the tuberculin test was negative, probably secondary to her steroid
dose.
As a lesson in medicine, she was presented at our weekly clinical pathological
conference in which surprises in diagnosis were routinely considered. We had all
learned in medical school about the reactivation of mycobacterial infection during
immunosuppression, but engrossed in treating one disease, we failed to consider
what we knew. This patient was entitled to more than one diagnosis and miliary
tuberculosis, to which she most probably succumbed, could have been approached
with appropriate therapy.
Jay A Levy
Department of Medicine, University of California, San Francisco, CA 94143-1270, USA

488

THE LANCET Vol 358 August 11, 2001

For personal use. Only reproduce with permission from The Lancet Publishing Group.

T86981045

Kirsti Malterud, 2001, Qualitative research: standards, challenges, and guidelines. Lancet, 358: 48388.

Guideline
:
1. ?
2. ?
3. ?
:
1. ()
?
:
1. ?
2. ?

1. ?
2. ?
3. ?
4. ?
5. ?

1. ?
2. ?
3. ?

1.
?
2.
?
3. ?
4.
?
5.


1. ?
2. (insight)?
3. (
)?

T86981045

4. ?

1. ()
()?
2. ?
3. ?
4. ?
5. ?

1. ?
2. ?

1.
?

228

InterViews

not be overlooked that the implicit, or unconscious, meanings attrib_


uted to interviewees and patients may often simply be the explicit and
conscious theories of the expert interpreter_
Eco (1990 , 1992) has addressed the vicissitudes of interpretation
in academic texts and in his novels. The Name of the Rose (1984) can
be read as a parody of the modern meaning hunters; as a critique of
the modern quest for true and objective meanings, of an insane
passion for truth" expressed in the intellectual dogmatism of the
scholastic disputes at the universiry of Paris; as well as of the empiricist
protagomst detective" searching for the objective truth while attempting to solve a mystery th,\t turns out to be very much of his own
making.
In his later novel Foucau/t's Pendu/um (1989) , the caricatures are
turned toward the relativism of the New Age , with its unlimited interpretations where everything can mean everything, as in the following
passage on the interpretation of quantitative measurements:

Truths?" Aglie laughed... Still, amid a11 the nOsense there are
some unimpeachable truths. G.entlemen, would you follow me to the
window?"
He threw open the shutters dramatically and pointed. At the corner
of the narrow street and the broad avenue, stood a little wooden kiosk,
where, presumably, lottery tickets were sold
Gent1emen ," he S31 invite you to go and measure that kiosk. You
wil1 see that the length of the CQunter is one hundred and forty-nine
centimeters1n other words, one hundred-billionth of the distance
between the earth and the sun. The height at the rear, one hundred and
seventy-six centimeters, divided by the width of the window, fifty-six
centimeters, is 3.14. The height at the front is nineteen decimeters , equal,
in other words , to the number of years of the Greek lunar cycle. The
sum of the heights of the two front corners and the two rear corners is
one hundred and ninety times two plus one hundred and seventy-six
times two , which equ a1s seven hundred and thi-two the date of the
victory at Poitiers. The thickness of the counter is 3.10 centimeters, and
the width of the cornice of the window is 8.8 centime Replacing the
numbers before the decimals by the corresponding letters of the
alphabet, we obtain C for ten and H for eig or Cl0HS , which is the
formula for naphthalene."
Fantastic," 1 said. You did all these measurements?" (Eco, p. 288).

13
.
The Social Construction
of Validity
1 now turn to the issue of how to get beyond the extremes of a
subjective relativism w l1 ere everything can mean everything, and an
absolutist quest for the one and only true , objective meaning.
Verification of knowledge is commonly discussed in the social
sciences in relation to the concepts of reliabili validi and generalizability. The main emphasis in this chapter will be on validation,
treating the interdependence of philosophical understandings of truth,
social science concepts of validity, and the practical issues of verifying
interview knowledge. Classical conceptions of truth will be included
as well as a postmodern approach leading to validity as soal construction. The ensuing practical consequences for interview research
involve an emphasis on the qualiry of the craftsmanship of research
and on communicative and pragmatic forms of validation

The Trinity of GeneraliZ ab i1ity,


Reliab i1ity, and Validity
In modern social science the concepts of generalizability , reliabili
and validity have reached the status of a scientific holy trinity. They
appear to belong to some abstract realm in a sanctuary of science far
removed from the interactions of the everyday world , and to be
worshipped with respect by all true believers in science

229

230

InterViews

As an introduction to the multiple contexts and discourses of


verification and the social construction of knowledge , 1 will start with
a history of my own encounters with the concept of validity. As a
student of psychology in Norway in the 1960s, 1 read heavy texts On
the importance of validity, reliability, and genelizability in scientific
research. 1 tried to memorize the definitions of predictive validitv.
concurrent validi content validity, and face val and struggled
to understand the concept of construct validity. The very terms
validity and reliability did not belong to the Norwegian vernacular
but were foreign English-Latin terms. The
p sy
chometr d
of validity appeared
a bstract and esoteric, as if belonging to some
distant philosophical universe together with Kant's transcendental a
prioris and the like.
As a student 1 dared to ask some natural scientists on campus about
these fundamental scientific concepts, and was somewhat bewildered
to find that the very terms of the methodological holy trinity of
psychological science were often unfamiliar to natural scientists. The
cdncepts were , however , very real to us students of psychology;
generalizabili validity, and reliability were frequently used as examination topics to differentiate between students who had , and those
who had not , pledged a l1 egiance to the scientific trinity of psychology
When later traveling in the United States 1 learned other meanings
for the terms validity and reliability; for example, when told while
cashing a check in the supermarket that my European driver's license
was not valid as identif in an academic discussion that my
argument was not valid. Or that the information about the used car 1
was looking at was not reliable, the car dealer was known to be an
unreliable person. Here the terms valid and reliable belong to the
vernacular , important to the ongoing interactions of everyday life.
When 1became engaged in qualitative research , the positivist trinity
emerged again , now employed by mainstream researchers to disqnalify
qualitative research. The stimulus qualitative research interview"
appeared automatical1y to trigger conditioned responses like: The
results are not reliable , they are produced by leading interVlew ques
tions"The interview
findings cannot be generalized, there are 100
few in1
1e

r
Vl
ews

ub
je

Ct8"; and
base on subjective inte
rpretations. "

The Social Construction o[ Validity

231

Some qualitative researchers have a different attitude toward questions of validity, reliability , and generalizability. These are simply
ignored or dismissed as some opprsive positivist concepts that
hamper a creative and emancipatory qualitative research. Other qualiive researchers-Lincoln and Guba (1985) , for instance-have
gone beyond the relativism of a rampant antipositivism and have
reclaimed ordinary language terms to discuss the truth value of their
findings , using concepts such as trustworthiness , credibility, dependability, and confirmability.
From a postmodern perspective issues of reliabili validity, and
generalizability are sometimes discarded as leftovers from a modernist
correspondence theory of truth. There are multiple ways of knowing
and mUltiple truths , and the concept of validity indicates a firm
boundary line between truth and nontruth. In co hereto Lather
(1995) , from a feminist post-structural frame valorizing practice,
add es validity as aI} incitement to discourse , a fertile obsession ,
and atte ll1pts to reinscribe validity in ways that use the postmodern
problematic to loosen the master code of positivism.
1 will return to external critiques of the trustworthine of interview
findings in the book's conclusion, Chapter 15. In the present chapter
1will attempt to conceptualize generalizability, reliabili and validity
in ways appropriate to qualitative research. The discussion represen
a rather moderate postmodernism; although rejecting the notion of
an objective universal truth , it accepts the possibility of specific local,
personal, and community forms of truth , with a focus on daily life and
local narrative (Kvale, 1992; Rosenau , 1992). The present approach
18 not to ject the concep of reliability, generalizability, and validity,
but to reconceptualize them in forms relevant to interview research.
The understanding of verification starts in the liv~d world and daily
language where issues of reliable observations, of generalization from
one case to another , of valid arguments , are part of everyday so al
mteractlO.

Generalizability
A persistent questio posed to interview studies is hher..
results are neralizabl!;. In everyday life w generalize moT

232

InterViews

spontaneously. From our experience with 0 situation or person we


anticipate new instances, we form expectations of what will happen
in other similar situations or with similar persons. 5cientific knowledq
!ays claim to generalizabiliiY;Y!Sl..Y<:.rnons. the mof
sU.hehot.J:rutLcould be
genelly. 4.E~tiI!g.lJ.Jllamstic yiew~ imolies th
lque--eh-phenomenhas....its....Q
structure Jogic. Within psychology, universal laws of behavio

a;e b sought by natural science-oriented schools such as behaviorism, whereas the uniquene of the individual person has dominated
in humanistic psychology. In a ppstmodern approach the quest for
universal knowledge , as well as the cult of the individually unique , is
replaced by an emphasis on the heterogeneity and contextuality of
knowledge , with a shift from generalization to ntextualization.
Forms of Generalizability. The issue f qualitative generalization
has been treated particularly in relation to case studies. Stake (1994)
provides this definition: Qualitative case study is characterized by the
main researcher spending substantial time, on site, personally in
contact with activities and operations of the case, reflecting, revising
meanings of what is going on" (p. 242). Three forms of generalizability will be outlined based on Stake's discussion of generalization from
case studies-naturalistic, statistical , and alytic.
Naturalistic generalization rests on personal experience: It dev e\ ops
for the person as a function of experience; it derives from tacit
knowledge of how things are and leads to expections rather than
formal predictions; it may become verbalized, thus passing from tacit
knowledge to explicit propositional knowledge.
Statistical generalization is formal and explicit: It is based on subjects se\ ected at random from a population. With the use of inferential
statistics the confidence level of generalizing from the selected sample
to the population at large can be stated in probability coefficients.
When the interviewees are selected at random and the interview
findings quantified , the findings may be subjected to statistical generalization. Thus for the correlation found between talkativity and grade
point average it was possible to state that there was only 1/1 ,000
probability that this was a chance finding limited to the 30 randomly
chosen pupils of the grade study (Chapter 12, Questions Posed to an
Interview Text).

The Socia/ Construction ofValidity

233

More often, interview subjects are not selected at random but by


other criteria, such as typicality or extremeness, or simply by accessibility. For example , an interview sample of women who have turned
to a help center for victims of violence are a self.selected and not a
random sample from the population. Their strong motivation for help
may lead to valuable knowledge on the nature of being subjected to
violence. The findings of the se \f-selected sample cannot, however, be
statistically generalized to the population at large.
Analytical generalization involves a reasoned judgment about the
extent to which the findings from one study can be used as a guide to
what might occur in another situation. It is based on an analysis of the
similarities and differences of the two situations. In contrast to spontaneous naturalistic generalization, the researcher here bases the generalization claims on an assertationallogic. There are several forms of
assertational logic , such as the legal form of argumentation in court
and arguments for generalization based on theory. By specifying the
supporting evidence and making the arguments explicit, the researcher can allow readers to judge the soundness of the generalization
claim (see also Yin , 1994 , on inductive generalization).
In her article, Generalizing From Single Case Studies" in system
evaluation, Kennedy (1979) argues for establishing rules for drawing
inferences about the generality of qualitative findings from a case
study, rules of inference that reasonable people can agree on. Whereas
the scientist tends to study specific cases in order to draw inferences
about the general case, the practitioner draws on knowledge of the
general case to form interpretations of and actions in the specific case.
As one point of departure , Kennedy turns to practical situations in the
legal and the clinical fi e\ ds.
In case law it is the most analogous preceding case, the one with
the most attributmilar to the actual case, that is selected as the
most relevant precedent. The validity of the generalization hinges on
the extent to which the attributes compared are relevant , which again
rests upon rich , dense , thick descriptions of the case. Kennedy outlines
criteria for relevant attributes of comparison in legal and clinical cases,
the latter instance encompassing precision of description , longitudinal
information, and multidisciplinary assessment.
In case law, the court dedes whether a previous cas

234

InterViews

Thus it is the receiver of the information who determines the applicability of a finding to a new situation. . . . Like generalizations in law, clinical
generalizations are the responsibility of the receiver of information
rather than the original generator of inforrnation, and the evaluator must
be careful to provide sufficient information to make such generalizations
possible. (Kennedy, 1979, p. 672)
Reseaher and Re ader Geralization. There is an issue here of who
should conduct the analytical generalization from the qualitative
research case-the researcher or the reader and the user? How much
should the researcher formalize and argue generalizations or leave the
generalizing to the reader? In science, it has commonly been the
researcher who builds up and argues for the generality of his or her
findings-through statistical procedures or by an assertational logic.
For the legal and the clinical cases discussed by Kennedy, it is the judge
or the clinician who makes the judgment of whether a pievious case
was sufficiently analogous to be used as a precedent for the present
case. In both instances it is paramount that sufficient evidence is
provided by the researcher for the analytic generalizations to be made
An example of a reader generalization that can be mentioned is Freud's
therapeutic case stories, where his descriptions and analyses have been
so vivid and convincing that readers today still generalize many of the
findings to current cases.

The Socia/ Consction ofValidity

235

whereby Marx's analysis of wage labor became increasingly generalizable to the situation of workers at large.
A third target of generalization is what could belocating situations that we believe are ideal and exceptional and studying them to
see what goes on there. As examples, Schofield mentions school classes
with unusual intellectual gains and also well-functioning racially
desegregated schools. In constructivist and postmodern approaches
the emphasis on the could be" is extended from preconceived ideals
to more open forms. Donmoyer (1990) thus advocates the use of case
studies to teach readers to envisage possibilities, to expand and enrich
the repertoire of social constructions available to practitioners and
others. We may here add the interest in ethnographic studi cases
demonstrating the rich varieties of human behavior, also indicating
possible ranges for our own society. Gergen (1992) depicts the construction of new worlds as one potential of a postmodern psychology.
Rather than telling it like it is ," the challenge is to tell it as it may
become." A "generative" theory is designed to unseat conventional
thought and thereby open new and desirable alternatives for thought
and action. Rather than mapping only what or predicting future
cultural trends , research becomes one means of transforming culture.

Reliability and Validity of Interviews


Targets of Generalization. Schofield (1990) has suggested three
targets for generalization. The first is studying what is-attempting to
establish the typical , the common, the ordinary. One seeks to mi
mize the fit between the research case and what takes place more
broadly in a society. A second rget is what may be-here the aim of
genelizing is not what is , but what may be. Schofield mentions a
study of the use of mputers in school that did not select average
representative schools, but schools at the leading edge of integrating
computers in teaching. This was done on the assumption that the most
advanced cases might provide findings generalizable to the future role
of computers in schools. A historical example may be added here-at
the time when Marx analyzed the situation of the wage laborers and
the contradictions of the use versus the exchange value of labor, wage
laborers made up only a small percentage of the working population.
Decades later, wage labor became the dominating form of labor,

Throughout this book 1 have emphasized that issues of verification


do not belong to some separate stage of an investigat but should
be addressed throughout the entire research process. As an introduction to conceptual issues of validity and t uth , some concrete issues of
the reliability and validity of interview inquiries from previous chapters will be briefly recapilated.
Reliability. Reliability pertains to the consistency of the research
findings. Issues of reliability during interviewing, transcribing , and
analyzing have been treated in the previous chapters. Interviewer
reliability was in particular discussed in relation to leading questions ,
which-when they are not a deliberate part of an interviewing technique-may inadvertently influence the answers , such as in the example of different wordings of a question about car speeds leading to

236

InterViews

different answers (Chapter 8, Leading Questions). 1nterviewer reliability in the grade study was discussed on the basis of the categorizations of the subjec answers (Chapter 11 , Control of An alysis). Under
transcription of interviews, an example was given of the intersubjec
tive reliability of the transcripts when the same passage was typed by
two different persons (Chapter 9, Transcription Reliability and Validity). During categorization of the grading interviews, percentages were
reported for the intersubjective agreement between two coders for the
same interviews (Chapter 11 , Control of An alysis). Though increasing
the reliability of the interview findings is desirable rder to counteract haphazard subjectivi a. strong emphasis on reliability may
counteract creative innovations and variability.

The Socia/ Construction of Validity

237

Box 13.1
Validation at Seven Stages

1. Thematizing. The validity of an investigation rests on


the soundness of the theoretical presuppositions of a study
and on the logic of the derivations from theory to the
research questions of the study.

2. Designing. The validity of the knowledge produced


depends on the adequacy of the design and the methods
used for the subject matter and purpose of the study. From
an ethical perspective, a valid research design involves
beneficence-producing knowledge beneficial to the human situation while minimizing harmful consequences.

Validity. Al though validation is treated in this chapter as a separate


stage , it concerns all seven stages of an interview investigation. 1n the
present approach , the emphasis on validation is moved from inspection at the end of the production line to quality control throughout
the stages of knowledge production.
Box 13.1 gives an overview of validity issues throughout an interview investigation. Before turning to conceptual issues of validity,
including validation as social construction, a brief outline of generalization by qualitative studies will be given.

3. Interviewing. Validity here per ns to the trustworthiness of the subject's reports and the quality of the interviewing itself, which should include a careful questioning
as to the meaning of what is said and a continual checking
of the information obtained as a validation in situ.

Validity in Modern and Postmodern Contexts

4. Tnscribing. The question of what constitutes a valid


translation from oral to written language is involved in the
choice of linguiic style for the transcript.

Ascertaining validity involves issues of truth and knowledge. 1 will


first discuss some meanings of validity, then include classical conceptions of truth , and thereafter discuss postmodern conceptions of
knowledge. The practical implications for interview research are then
treated with respect to validity as craftsmanship in research , as communication and action.
1n ordinary language dictionari validity refers to the truth and
correctness of a statement. A vali4 argument is sound, well grounded,
justifiable, strong, and convincing. A valid inference is correctly
derived from its premises. 1n social science textbooks one finds both
a narrow and a broad definition of validity. 1n a positivist approach,
scientific validity became restricted to measurements: for instance,

5. Analyzing. This has to do with whether the questions


put to an interview text are valid and whether the logic of
the interpretations issound.
6. Validating. This entails areflected judgment as to what
forms of validation are relevant to a specific study, the
application of the concrete procedures of validation, and a
decision on what the appropriate community is for a dialogue on validi
7. Reporting. This involves the question of whether a
given report is a valid account of the main findings of a
study, as well as the role of the reade of the report in
validating the results.

238

InterViews

Validity is often defined by asking the question: Are you measuriug


what you think you are measuring?" (Kerlinger, 1979, p. 138). Qualitatlve arch is then iuvalid if it does not result in numbers. In a
broader concept, validity pertains to the degree that a method investigates what it is intended to investigate the extent to which our
observations indeed reflect the pheuomena or variables of interest
to us" (Pervin, 1984, p. 48). Within this wider conceptiou of validity , qualitative research can, in principle , lead to valid scientific
knowledge.
Textbook presentations have been based on positivist epistemological aumptions with a correspondence theory of truth. The standard
definitions of validity have been taken from the criteria developed for
psychological tests as formalized by Cronbach and Meehl in 1955. In
psychology, validity became linked to psychometrics, where the concurrent and predictive validity of the psychological tests were declared
in correlation coefficients, indicating correspondence between test
results and some external criteria. These psychometric tests , such as
intelligence tests, have frequently been applied to predict school
success. The external criterion was here simple-grade point average
in later schooling. With a further questioning about what the school
grades measure , the issue becomes more complex. Grades have been
found to predict later grades in school, but success after graduation to
a lesser extent. The issue of predictive validity is here not merely an
empirical issue, but raises such normative questions as what should
the criteria of success be-position in the occupational hierarchy,
income, contributions to the community?
The issue of what is valid knowledge involves the philosophical
question of what is truth. Within philosophy, three classical criteria
of truth are discerned-correspondence, coherence, and pragmatic
utility. The correspondence criterion of truth concerns whether a
knowledge statement corresponds to the objective world. The coherence criterion refers to the consistency and internal logic of a statement. An d the pragmatc criterion relates the truth of a knowledge
statement to its practical consequences.
Al though the three criteria of truth do not necessarily exclude each
other, they have each obtained strong positions in different philosophical traditions. The correspondence criterion has been central
within a positivist social science
,

The Socia/ Construction of Validity

111'

239

expressed as its degree of correspondence with an objective reality.


The coherence criterion has been strong in mathematics and hermeneutics. The pragmatic criterion has prevailed in pragmatism and to
a certain extent in Marxist philosophy. The three truth criteria can be
regarded as abstractions from a unity, where a comprehensive verification of qualitative research findings wilI involve observation, conversation , and interaction
The belief in an objective world has been the basis of a modernist
undersnding of truth and validity. In a positivist philosophy, knowledge became a reflection of reality: There is only one correct view of
this independent external world, and there is ideally a one-to-one
correspondence between elements in the real world and our knowledge of this world. In a postmodern era , the foundations of true and
valid knowledge in a medieval absolute God or a modern objective
reality have dissolved. The conception of knowledge as a mirror of
reality is replaced by knowledge as a social construction of reality.
Truth is constituted through a dialogue; valid knowledge claims
emerge as conflicting interpretations and action possibilities are discussed and negotiated among the members of a community.
In science the decisive point is the conversation in the community
of researchers about the relation among the methods , the findin
and the nature of the phenomena investigated. The move from knowledge-as-observation to knowledge-as-conversation was ilIustted in
a recent television program on the development of the natural sciences. After showing the newest technical advances in microscopes for
cell studies and giant telescopes for the investigation of space, the
camera suddenlyshifted to a room with elegant 18th-century furniture. The transition was accompanied by a voice saying something to
the effect that: It is not by the techniques of these instruments and the
resulting observations that the truth of the new sentific knowledge
is determined, but through discussions about the observations among
the scientists, such as in this room of the British Royal Society of
Sciences.
The social construction of valid knowledge is brought out in the
concept of construct validity, which was originally inoduced by the
psychometricians Cronbach and Meehl for psychological tests. It
pertains to the measurement of theoretical constructs-such as intelligence and authoritarianism-by different measures; constr

240

InterViews

ity involves correlations with other measures of the construct and


logical analysis of their relationships. Cronbach (1971) later extended
the concept of construct validiry to qualitative summaries as well as
numerical scores; it 18 an open proce in which to validate is to
investigate-validation is more than cooboration; t 1S a proce for
developing under interpretations of observations" (p. 443).
Cherryholmes (1988; see also Tschudi, 1989) has argued that
construct validity is a discursive and rhetorical concept. A construct
and its measurement are validated when the discourse about their
relationships is persuasive to the communiry of researchers. A constructive conception of validity goes beyond the original discourse of
psychological testing and experimental design , and opens in Cherry
holmes's analysis to multiple discourses, such as phenomenological,
interpretative, critical, and deconstructive analyses. This radicalization of construct validity brings it close to a postmodern emphasis on
the social construction of knowledge.
Some implications of the above discussion for validation of qualitative research will now be discussed. First, when giving up a correspondence theory of truth as the basis for understanding validiry, there
is, following Popper, a change in emphasis from verification to falsification. The quest for absolute , certain knowledge is replaced by a
conception of defensible knowledge claims. Validation becomes the
issue of choosing among competing and falsifiable interpretations, of
examining and providing arguments for the relative credibiliry of
alternative knowledge claims (Polkinghorne, 1983). Validation here
comes to rest on the quality of craship in research.
Second , a modern belief in knowledge as a mirror of realiry recedes
and a social construction of reali with coherence and pragmatic
criteria of truth , comes to the foreground. Method as a guarantee of
truth dissolves; with a social construction of reality the emphasis is on
the discourse of the communiry. Communication of knowledge becomes significant, with esthetics and rhetorics entering into a scientific
discourse.
Third, with a modern legitimation mania receding, there is an
emphasis upon a pragmatic proof through action. The legitimation of
knowledge through external justification by appeals to some grand
systems, or meta-narratives, and the modern fundamentalism of secufing knowledge on so

The

Socia/ Construction ofValidity

241

Justification of knowledge is replace by application , knowledge


becomes the ability to perform effective actions. Criteria of efficiency
and their desirability become pivotal, raising ethical issues of right
action. Values do not belong to a realm separated from scientific
knowledge , but permeate the creation and application of knowledge.
Implications of the above discussion for interview research will now
be addressed in relation to validation as craftsmanship, as communication, and as action. This does not lead to new , fixed criteria replacing the psychometric concepts of validi nor does it secure unambiguous knowledge. Rather, it extends the frames of reference for asking
about the validity of knowledge in social research- Post-modern
social sci~nce presumes methods that multiply paradox, inventing ever
more elaborate repertoires of questions, each of which encourages an
infiniry of answers, rather than methods that settle on solutions"
(Rosenau , 1992, p. 117).

Validity as Quality of Craftsmanship


1 will here attempt to demystify the concept of validi to bring it
back from philosophical abstractions to the everyday practice' of
scientific research. With an alternative concept of validity-going
from correspondence with an objective reality to defensible knowledge claims-validity is ascertained by examining the sources of
invalidiry. The stronger the falsification attempts a proposition has
survived, the more val the more trustworthy the knowledge. Validation comes to depend on the quality of craftsmanship during
investigation, continually checking, questioning, and theoretically
interpreting the findings.
The concept of validity as quality of craftsmanship is not limited to
a postmodern approach , but becomes pivotal with a postmodern dismissal of an objective realiry against which knowledge is to be measured. The craftsmanship and credibility of the researcher becomes
essential. Based on the qualiry of his or her past research in the area ,
the credibiliry of the researcher is an important aspect of fellow researchers ascribing validiry to the findings reported. Validiry is not only
a matter of the methods used; the person of the researcher (Salner,
1989) , including his or her moral integriry (Smith , 1990) , is critical

242

InterViews

for evaluation of the quality of the scientific knowledge produced.


Three aspects of validation as investigation will now be outlinedchecking , questionin and theorizing the knowledge produced.
To Validate Is to Check. The researcher adopts a critical outlook on

the analysis , states explicitly his or her perspective on the subject


matter studied and the controls applied to counter selective perceptions and biased interpretations, and in general plays the devil's
advocate toward his or her own findings.
Various modes of checking the findings have been suggested by
writers on qualitative research. An investigative concept of validation
is inherent in the grounded theory approach of Glaser and Strauss
(1967). Validation is here not some final verification or product
control; verification is built into the research process with continual
checks on the credibili plausibil and trustworthiness of the
findings. Miles and Huberman (1 994) emphasize that there are no
canons or infallible decision-making rules for establishing the validity
of qualitative research. Their approach is to analyze the many sources
of potential biases that might invalidate qualitative observations and
interpretations; they outline in detail tactics for testing and confirming
qualitative findings. These tactics include: checking for representativeness and for researcher effects, triangulating, weighing the
evidence , checking the meaning of outl using extreme cases,
following up on surprises, looking for negative evidence, making
if-then tests , ruling out spurious relations , replicating a finding , checking out rival explanations, and getting feedback from informants
(p.263).
Runyan (1981) discussed the validation of multiple interpretations
in psychobiography in relation to the episode of van Gogh cutting off
his left ear and giving it to a prostitute. More than a dozen explanations of this act have been proposed in the literature, ranging from
inspiration by newspaper accounts of Jack the Ri pper, to visits to
bullfights in Arles , to aggression turned inward and a reawakening of
Oedipal themes. Runyan discusses in detail the credibility and strength
of the different interpretations. This includes checking the empirical
evidence for and against an interpretation, examining the theoretical
coherence , and critically evaluating and comparing the relative plausibility of the different interpretations given for the same act. 1n general,

The Socil Construction ofValidity

243

the more attempts at falsification an interpretation has survived, the


stronger it stands.
To Validate Is to Question. When ascertaining valid the questions of what" and why" need to be answered before the question
of how": The content and purpose of an investigation precedes the
method. Discussing the question Do photographs tell the truth?"
Becker (1 979) makes the general question Is it true?" specific in 1s
this photograph telling the truth about what?" An d to decide what a
picture is telling us the truth about, he suggests that we should ask
ourselves what questions it might be answering.
A common critique of research interviews is that their findings are
not valid because the subjects' reports may be false. This is a possibility
that needs to be checked in each specific case (see Dean & Whyte ,
1969). The issue of validity again depends on the what" of the
researcher's questions. In hermeneutical interpretations, the questions
posed to a text become a11-important. 1n the grading study, the
primacy of the question posed to an interview statement was demonstrated by the interpretations of pupils' statements about competition,
talkativi and wheedling (Chapter 12, Questions Posed to an 1nterview Text). Different qutions posed to interview texts led to diffi: rent answers. Thus one type of question led to an experiential reading
of the pupils' statements. An other type of question led to a veridical
reading, regarding the interviewees as w t nesses or informants. The
questioning also involved a symptomatic reading, focusing on the
interviewees themselves and their reasons for making a given statement. The forms of validation differ for the different questions to the
interview texts. 1n the grade study they varied from a critical follow-up
in the interview of the pupils' statements, to statistical analysis to
verify a pupil's postulated connection between talkativity and grades ,
to the coherence of interpretations about the production and the
consequences of beliefs about grading.
Richardson (1994) has taken issue with the geometrical concept of
triangulation , which was applied above in the validation of a veridical
reading of a pu's postulate of a connection between talkativity and
grades (Chapter 12, Questions Posed to an 1nterview Text). Ri chardson rejects the use of a rig two-dimensional triangle as a
central image for validity for postmodern texts , because it contains

244

InterViews

asumptions

of a fixed point or object that can be triangulated:


Rather, the central image is the crystal, which combines symmetry
and substance with an infinite variery of shap substances, transmutations , multidimensionalities, and angles of approach" (p. 522). She
then outlines how crystallization by means of postmodern mixedgenre texts provides us with a deepened, complex, and partial understanding of the topic. The multiple questions to, and readings of, the
pupils' statements about grades may be seen as crystallizations opening
to continual transformations of the meaning of grades.
To Validate Is to Theorize. Validity is not only an issue of method.
Pursuing the methodological issues of validation generates theoretical
questions about the nature of the phenomena investigated. Deciding
whether a method investigates what it intends to investigate involves
a theoretical conception of what is investigated. In the terms of
grounded theory, verifying interpretations is an intrinsic part of the
generation of theory.
The inconclusive results in the grade srudy of the attempts at an
informant-triangulation of a pupil's belief in a connection between
talkativiry and grades need not merely indicate a problem of method;
it also raises theoretical questions about the social contruction of
school realiry. Pupils and teachers may live in different social realities
with regard to which pupil behaviors lead to good grades. It is possible
that pupils, in a kind of superstitious" behavior, believe in a connection where there is none; or it may be that teachers overlook or deny
a relation that actually exists. Ambiguiry of the teacher's bases for
grading , and contradictory beliefs by pupils and teachers about which
behaviors lead to good grades, appear to be essential aspects of the
social reality of schoo l. The complexities of validating qualitative research need not be due to an inherent weakness in qualitative metho
but may on the contrary on their extraordinary power to picture
and to question the complexity of the social reality investigated.
Communicative Validity
Communicative validiry involves testing the validiry of knowledge
claims in a dialogue. Valid knowledge is constituted when conflicting
knowledge claims are argued in a dialogue: What is a valid observation

The Socia/ Construction ofValidity

245

is decided through the argumentation of the participants in a discourse. ln a hermeneutical approach to meaningful action as a text,
Ricoeur (1971) rejects the position that all interpretations of a text
are equal; the logic of validation allows us to move between the two
limits of dogmatism and skepticism. lnvoking the hermeneutical circle
and criteria of falsiili he describes validation as an argumentative discipline comparable to the juridical procedures of legal interpretation. Validation is based on a logic of uncertainty and of qualitative probabili where it is always possible to argue for or against
an interpretation, to confront interpretions and to arbitrate between
them.
A communicative approach to validity is found in several approaches in the social sciences. ln psychoanalysis the validity of an
interpretation is worked out in a dialogue between patient and th
pist. It is also implied in evaluation studies of social systems; House
(1980) has thus emph~sized that in system evaluation, research does
not mainly concern predicting events, but rather whether the audience
of a report can see new relations and answer new but relevant
questions. Cronbach (1980) has advocated a discursive approach
where validiry rests on public discussion. The interpretation of a test
is going to remain open and unsettled, the more so because of the role
that values play in action based on tests; the aim for a research report
is to advance sensible discussion-and,The more we learn, and the
franker we are with ourselves and our clientele , the more valid the use
of tests will become" (p. 107). ln a discussion of narrative research,
Mishler (1990) has conceprualized validation as the social construction of knowledge. Valid knowledge claims are established in a
discourse through which the results of a srudy come to be viewed as
sufficiently trustworthy for other investigators to rely upon in their
own work.
When conversation is the ultimate context within which knowledge
is to be nnderstood, as argued by Rorry (Chapter 2 , lnterviews in
Three Conversations) , the narure of the discourse becomes essential.
There is today a danger that a conception of truth as dialogue and
communicative validation may become empty global and positive
undifferentiated terms , without the necessary conceprual and theoretical differentiations worked out. Some specific questions concerning
the how, why, and who of communication will now be raised.

246

InterViews

How. Communication can involve persuasion through rational


discourse or through populist demagogy. The forms of persuasion
about the truth of knowledge claims will be different in the harsh
logical argumentation of a philosophical dialogue , in the juridical
proceedings and legal interpretations in a courtroom, in a narrative
capturing an audience , and in a humanistic therapy encounter based
on positive feelings and reciprocal sympathy.
Philosophical discourses , such as the dialogues of Socrates, are
characterized by a rational argumentation. The participants are
obliged to test statements about the truth and falsity of propositions
on the basis of argued points of view, and the best argument wins.
This discourse is ideally a form of argumentation where no social
exertion of power takes place, the only form of power being the force
of the better argument.
Why. The question here concerns the purpose of a discourse aboU!
knowledge. What are the aims and criteria of arriving at true
knowledge? Habermas's (1971) discourse theory implies a consensual
theory of uth: The ideal discourse aims at universally valid truths as
an ideal. Eisner (1991) has advocated qualitative research as art, based
on connoisseurship and criticism, accepting the personal, literary, and
even poetic as valid sources of knowledge. The aim is here consensus:
Consensual validation base agreement among competent others
that the description , interpretation, evaluation, and thematics of an
educational situation are right" (p. 11. From a postmodern perspective , Lyotard (1984) has argued that consensus is only a stage in a
discussi and not its goal , which he posi paralogy-to create
new ideas, new differentiations, new rules for the discourse. To
Lyotard, discourse is a game between adversaries rather than a dialogue between partners.

ue

Who. The concept of communicative validity raises the question of


who communicates with whom. Who is a legitimate partner in a dialogue about true knowledge? Three interpretative communities were
brought in by the validation of the interviews of grading. The member
of the interpretative community validating an interpretation could be
the subject interviewed, thegeneral public interpreting within a critical
common sense understanding analogous to a jury, or the scientic
community of scholars possessing methodical and theoretical compe-

The Socia/ Construction ofValidi

247

tence in the specific area. Taking a lead from the use of reflecting
teams" in psychotherapy, where the one-way mirror is reversed and
the family in treatment views the therapeutic team's discussions of
their interpersonal interaction (An dersen , 1987), we might also reverse the direction in research and have the subjects listen to and
comment on the researchers' conversations about their interviews.
Validation through negotiations of the community of scholars is
nothing new; in the natural sciences the acceptance of the scientific
community has been the last , ultimate criterion for ascertaining the
truth of a proposition. What is relatively new in qualitative research
in the soal sciences is the emphasis on truth as negotiated in a local
context, with extension of the interpretative community to include
the subjects investigated and the lay public. Communicative validation
approximates an educational endeavor where truth is developed in a
communicative process, with both researcher and subjects learning
and changing through the dialogue.
A heavy reliance on intersubjective validation may, however , also
imply a lack of work on the part of the researcher and a lack of
confidence in his or her interpretations, with an unwillingne to take
responsibility for the interpretations. There may be a general populist
trend when leaving the validation of interpretations to the readers; as
m reaer response validation, with an abdication to the ideology of a
consumer 80 ety:

Power and Truth. Different professional communities may construct knowledge different and conflicts may arise about which
professions have the right to decide what is valid knowledge within a
field , such as health , for example. Furthermore, there is the specific
issue of who decides who is a competent and legitimate member of
the interpretative community. The selection of members of the community to make decisions about issues of truth and value is considered
crucial for the results in many cases, such as in the selection of
members of a jury, or of a committee to examine a doctoral candidate,
or of an academic appointment committee.
Haberm's consensus theory of truth is based on the ideal of a
dominance-free dialogue , which is a delibete abstraction from the
webs of power relationships within real-life discourses , and again in
contrast with Lyotard's postmodern understanding of a scientific con-

248

InterViews

versation as a game of power. More generally, scientists are not Purchased to find truth , but to augment power: The games of scientific
language become the games of the rich, in which whoever is wealthiest
has the best chance of being right. An equation between wealth.
efficiency and truth is thus established" (Lyotard, 1984, p. 45).

Pragmatic Validity
Pragmatic validation is verification in the literal sense- to make
true." To pragmatists, truth is watever assists us to take actions that
produce the desired results. Kn owledge is action rather than observation , the effectiveness of our knowledge beliefs is demonstrated by the
effectiveness of our action. In the pragmatic validation of a knowledge
claim, justification is replaced by application. Marx stated in his
second thesis on Feuerbach that the question of whether human
thought can lead to objective truth is not a theoretical but a practical
one. Man must prove the truth , that is, the reality and power of his
thinking in practice. An d his 11 th thesis is more pointed; the philosophers have only interpreted the world differently, what matters is
changing the world.
A pragmatic concept of validity goes farther than communication;
it represents a stronger knowledge claim than an agreement through
a dialogue. Pragmatic validation rests on observations and interpretations , with a commitment to act on the interpretations.Actions
speak louder than words." With the emphasis on instigating change,
a pragmatic knowledge interest may counteract a tendency of social
constructionism to circle around in endless interpretations and a
plunge of postmodern analyses into infinite deconstructions.
A pragmatical knowledge interest in helping pn change is
intrinsic to the therapeutic interview, where communication of interpretations serves to instigate changes in the patient. For naturalistic
inquiry, Li ncoln and Guba (1985) have gone farther than consensual
validation and pointed to action-oriented quality criteria for qualitative research , such as an inquiry enhancing the level of understanding of the participants and their ability to take action , empowering
them to take increased control of their lives. Action arch goes from
descriptions of social conditions to actions that can change the very

The Socia/ Construction ofValidity

249

conditions investigated. Also, system evaluation goes beyond the


corresponence criterion to include pragmatic validity: The ultimate
tests of the credibility of an evaluation report is the response of decision makers and information users to that report" (Patton, 1980,
p.339).
We may discern between two types of pragmatic validationwhether a knowledge statement is accompanied by action , or whether
it instigates changes of action. In the first case, validation of a subject's
verbal statement is based on supporting action that accompanies the
statement. This concerns going beyond mere lip service to a belief, to
following it up with action. Thus in investigations of racial prejudice,
comprehensive inquiries go beyond a subject's mere verbal statements
against racial segregation and investigate whether the statements are
also accompanied by appropriate supportive actions.
The second, stronger, form of pragmatic validation concerns
whether interventions based on the researcher's knowledge may instigate actual changes in behavior. Freud did not rely on the patient's
self-understanding and verbal communication to validate therapeutic
interpretations; he regarded neither the patient' s yes" nor no" to
his interpretations as sufficient confirmation or disconfirmat the
yes" or no" could be the result of suggestion as well as of resistance
in the therapeutic pro. Freud recommended more indirect forms
of valia~ion such as observing the patient's reactions to an interpre
tation , for example in the form of changes in the patient's free
associations , dreams , recall of forgotten memories, and alteration of
neurotic symptoms (Freud, 1963 , p. 279). Spence (1982) has followed
up on the emphasis on the pragmatic effects of interpretations: Narrative truth is constructed in the therapeutic encounter, it carries the
conviction of a good story, and it is to be judged by its aesthetic value
and by the curative effect of its rhetorical force.
In collaborative action research , investigators and subjects together
develop knowledge of a social situation and then apply this knowledge
through new actions in the situat thus tting the validity of the
knowledge in praxis. Reason (1994) describes a study of health
workers that was based on participatory inquiry with a systematic
testing of theory in live-action contexts. The topic was stress that came
from hidden agendas in their work situatio such as suspicions of
drug tak

250

InterViews

visited. The coresearchers first developed knowledge through discussions among themselves , by role playing, and thereafter by raising their
concerns directly with their client families. Reason discusses the
validity in this cooperative inquiry, and emphasizes the need to get
beyond a mere consensus collusion where the researchers might band
together as a group in defense of their anxieties , which may be
overcome by a continual interaction between action and reflection
throughout the participatory inquiry.
How. The forms of pragmatic validation vary: There can be a
patien t' s reactions to the psychoanalyst' s interpretation of his or her

dreams, or a client's responses to a behavior therapist's interventions


to change the reinforcement contingencies of his or her problem
behavior. There are the reactions of an audience to a system evaluation
report , and the cooperative interaction of researcher and subjects in
action research.

Why. A scientific discourse is, in principle, indefinite; there is no


requirement of immediate action; new arguments that could alter or
invalidate earlier knowledge can always appear. In contrast to the
uncoerced consensus of the scientific discourse , practical contexts may
require actions to be undertaken and decisions to be made that involve
a coercion to consensus. This includes the proceedings of a jury, the
negotiations of a dissertation committee, decisions about therapeutic
interventions, and decisions about institutional changes in action
research.
Apgmatic approach implies that truth is whatever assists us to
take actions that produce the desired results. Deciding what the
desired results are involves values and ethics. The moral normative
aspect of validation is recognized in system evaluation, where the
validity of an evaluation depends upon whether the evaluation is true ,
credible, and normatively correct" (House , 1980, p. 255). The importance of values in validation follows through a change of emphasis in
social research from primarily mapping the social world with respect
to what to changing the focus to what could be. Thus Gergen's
(1992) postmodern conception of generative theory (see Generalizability, above) involves research that opens new possibilities of
thought and action as a means of transforming culture.

The Socia/ Construction ofValidity

251

Who. The question of who" involves the researcher and the users
of the knowledge produced. Patton (1980) emphasizes the credibility
of the researcher as an important criterion of whether a research
report is accepted or not as a basis for action. The question of who"
also involves ethical and political issues. Who is to decide the direction
of change? There may be personal resistance to change in a therapy
as well as conflicting vested interests in the outcome of an action study.
Thus, regarding audience validation in system evaluation, who are the
stakeholders that will be included in the decisive audience: the funding
agency, the leaders of the system evaluated, the employees, or the
clients of the system?
Power and Truth. Pragmatic validation raises the issue of power and
truth in social research: Where is the power to decide what the desired
results of a study will be, or the direction of change; what values are
to constitute the basis for action? An d , more generally, where is the
power to de de what kinds of truth seeking are to be pursued, what
research questions are worth funding? Following Foucault we should
here beware of localizing power to specific persons and their int.entions , and instead analyze the netlike organization and multiple fields
of power-knowledge dynamics.

f
Validity of the Validity Question
1 have argued here for integrating validation into the crafmanship
of research, and for extending the concept of validation from observation to also include communication about, and pragmatic effects of,
knowledge claims. The understanding of validity as craftsmanship, as
communication and action , does not replace the importance of precise
observations and logical argumentation, but includes broader conceptions of the nature of truth in social research. The conversational and
pragmatic aspects of knowledge have within a positivist tradition been
regarded as irrelevant, or secondary, to obtaining objective observations; in a postmodern conception of knowledge the very conversation
about , and the application of, knowledge become essential aspec of
the construction of a social world. Rather than providing fixed criteria, communicative and pragmatic validation refer to extended ways
of posing the question of validity in social research.

252

InterViews

1 have further attempted to demystify the concept of validi


main n that verification of information and interpretations is a
normal actJ vlty m the interactions of daily life. Even so, a pervasive
attention to validation can be counterproductive and lead to a general
invalidation. Rather than let the product, the knowledge c1aim , speak
for itseIf, validation can involve a legitimation mania that may further
a corrosion of validity-the more one validates , the greater the need
for further validation. Such a counterfactuality of strong and repeated
emphasis on the truth of a Statement may be expressed in the folk
saymg,Beware when they swear they are teIIing the truth."
Ideally, thc qualIty of the craftsmanship results in products with
knowledge c1a11
ms
right
t hat
t hey so to say, carry the validation with them, like a stron~
piece of art. In such cases, the research procedures would be t
en
ta

nd tI re
s
1

o
nVln

iC
m

ga
s
tr
ue
be

au11

and god. Appeals to external certification , or offial validity stamps of approval, then become secondarv
Valid resch would in sense be research t makes questions of
vali.dity superfluous.

14
'.i

Improving Interview Reports


When the understanding of validation and generalization is extended
to include communication with readers, the writing of repor takes
on a key position in an interview inquiry. Reporting is not simply
re-presenting the views of the interviewees, accompanied by the
researcher's viewpoints in the form of interpretations. The interview
report is itseIf a social construction in which the author's choice of
writing style and literary devices provide a specific view on the
subjec lived world. The writing process is one aspect of the social
conuction of the knowledge gained from the interviews, and the
report becomes the basis for the research community to ascertain the
validity of the knowledge reported. The current focus on conversation
and rhetorics in social research , as weII as what is termed a crisis of
representation, leads to an emphasis on the presentation of research
findings.
Interview reports are often boring to read. Some ways of improving
standard modes of reporting interviews will be outlined and some
ethical issues of reporting interviews pointed out. FinaIIy, after discussing writing as a social construction, modes of enriching interview
reports are suggested.

Boring Interview Reports


Some three thousand years ago , Odysseus returned to Greece from
his research inquiry in distant countries. Homer's oral tale of the

T76991048

The Social Construction of Validity


1. :
To conceptualize generalizability, reliability, and validity in ways appropriate to
qualitative research, and to reconceptualize them in forms relevant to interview
research.
2. (Generalizability): ?
-

Three forms of generalizability based on Stakes discussion of generalization


from case studies can be used in critique for Literature review.
a. Naturalistic generalization: rested on personal experiences: It develops for
the person as a function of experience; it derives from tacit knowledge of how
things are and leads to expectations rather than formal predictions; it may
become verbalized, thus passing from tacit knowledge to explicit propositional
knowledge.

b. Statistic generalization: When the interviewees are selected at random and


the interview findings quantified, the findings may be subjected to statistical
generalization.
c. Analytical generalization:
()
(situations)
:

:
(supporting evidence)(Yin, 1994)
(argument)?(Yin, 1994)?
(precision of description)(Kennedy, 1979)?
(longitudinal information) (Kennedy, 1979)?
(multidisciplinary assessment)?
3.
- Issues of verification() do not belong to some separate stages of an
investigation, but should be addressed throughout the entire research process.

Reliability pertains to the consistency of the research findings.


:

(): (Interviewer
reliability)? (leading questions)
intersubjective
reliability?
1

T76991048

Whether the percentages were reported for the intersubjective agreement between
two coders for the same interviews
4. Validity as quality of Craftmanship()
-

(To validate is to check).

Miles and Huberman(1994) outline in detail tactits for testing and confirming
qualitative findings, including checking for
(Checking for representativeness and

researcher effects)
(triangulation)
(weighing the evidence)
Outliers (meaning of outliers)
(extreme cases)
(following up on surprices)
(looking for negative evidence)
(if-then tests)
(ruling out spurious relations)
(replicating a finding)
(rival)

(getting feedback from informants)


Runyan (1981) discussed in detail the credibility and strength of the different

interpretations().

(Checking the empirical evidence for and


against an interpretation)
(Examining the theoretical coherence.)
(Critically

evaluating and comparing the relative plausibility of the different interpretations


given for the same act.)
-

(To validate is to Question)

A common critique of research interviews is that their findings are not valid
because the subjects reports may be false. Different questions posed to interview
texts led to different answers. The forms of validation differ for the different questions
to the interview texts. For example, a critical follow-up to a statistical analysis and the
coherence of the interpretations,
Richardson(1994) proposed the concept of (triangulation)
-

(To validate is to Theorize)


2

T76991048

Whether involving a theoretical conception of what is investigated.

5. (Communicative Validity)
Communicative validity involves testing the validity of knowledge claims in a
dialogue. What is a valid observation is decided through the argumentation of the
participants in a discourse.

Critique criteria
House (1980) stated that in system evaluation, research does not mainly concern
predicting events, but rather whether the audience of a report can see new
relations and answer new but relevant questions (
).

Valid knowledge claims are established in a discourse through which the results
of a study come to be viewed as sufficiently trustworthy for other
investigators to rely upon in their own work(
?)

How. Communication can involve persuasion through rational discourse or


through populist demagogy().
Why. The purpose of a discourse() about true knowledge.

Who. Who communicates to whom. Who is a legitimate partner in a dialogue


about the true knowledge? The subject interviewed? The general public or the
scientific community of scholars?
Power and truth. The selection of members of the community to make decisions
about issues of truth and value is considered crucial for the results in many cases.

We might also reverse the direction in research and have the subjects listen to
and comment on researchers conversations about their interviews.

6. Pragmatic Validity:
The effectiveness of our knowledge beliefs is demonstrated by the effectiveness
of our action. In pragmatic validity, justification is replaced by application().
Pragmatic validation rests on observations and interpretations, with a commitment to
act on the interpretations.

Lincoln and Guba(1985) have gone farther than consensual validation and
pointed to action-oriented quality criteria for qualitative research, such as
An inquiry enhancing the level of understanding of the participants(
) and their ability to take action(), empowering them to
take increased control of their lives().

Two types of pragmatic validation- whether a knowledge statement is


3

T76991048

accompanied by action(), or whether it instigates


changes of action().

Who. The question of who involves the researcher and the users of the
knowledge produced. Patton(1980) emphasizes the credibility of the researcher
as an important criterion of whether a research report is accepted or not as a basis
for action.
Power and Truth. Where is the power to decide what the desired results of a
study will be, the direction of change; what values are to constitute the basis for
action?

Daniel T. L. Shek
X. Y. Han
Vera M. Y. Tang

The materials in this paper have been included in the following publications:

Shek, D. T. L., Tang, M. Y., & Han, X. Y. (2005). Evaluation of Evaluation Studies Using
Qualitative Research Methods in the Social Work Literature (1990-2003): Evidence That
Constitutes a Wake-Up Call. Research on Social Work Practice, 15, 180-194.

2005
1990-2003
241 70-100

qualitative

evaluation 1990 2003 Social Work


Abstracts
13

2
3
4
5

6
7

Bryman1988
123

4
5
6Denzin Lincoln1998

1
2
3
45
6

2004
1
quantitative
492
qualitative 1,238 Brun1997
1982 1992 54
Dellgran Hojer2001
1979 1998 89 14
36Roberts1989

2003
14 14 7

Thyer1989
p.309

Lather, 1986; Lincoln & Guba, 2000; Seale, 2002


:
Schwandt, 2000Bloland, 1995
Schwandt1996farewell to
criteriology
p.70

Sokal & Bricmont, 1998

Atherton & Bolland, 2002; Rubin, 2000


accountability

Huberman & Miles, 1994


Patton (2002

positivistic paradigm

constructivist paradigm

artistic and evocative


paradigm
critical paradigm

LeCompte
Goetz1982internal reliability

external reliability

internal validity
external validity
triangulation
Tschudi, 1989
Guba Lincoln1981
credibility

fittingness

auditability

confirmability

Sandelowski1986

1
2

3
4
5
67
Huberman Miles1994

5
1

4
5

Anastas, 2004;
Drisko, 1997Drisko1997
Drisko1997

Drisko1997

Anastas2004
1
2

Patton, 1990; Denzin & Lincoln, 2000

general
Patton, 1990specific

1
Drisko1997

Patton1990

2
3

Drisko1997

Lincoln & Guba, 2000; Ryan, 1998

Steier1991

p.2
Rosenau1992

p.114
4
Huberman &
Miles, 1994

5
Drisko1997

reliability
validity

intra-rater reliabilityinter-rater
reliability 6
7
Anastas2004

communalLincoln1998
community
Rizzo, Corsaro Bates1992
peer checking
member checking

Sandelowski1986

p.34
audit trail, Huberman & Miles, 1994
9

Anastas2004

10

negative cases 11
Huberman
Miles (1994
Gambrill, 1999; Gomory, 2000a, 2000bLincoln & Guba, 2000;
Ryan, 1998
12Drisko1997

11990-2003
qualitativeevaluation

Social Work Abstracts 1990 2003 2003


12 31 75
28
28 1

2
12 28

1:

2:

3:

4:

5:

6:

7:

8:

9:

10:

11:

12:

1228

2
1N=22
N=8
2
3
13

10

44
95
3
6
67

8
9285
auditabilityaudit trail

13
101811
8
12

1
1990 2003

11

generalspecific

Patton1990

2 3 9

LeCompte Goetz 1982

4 5

Janesick1998

12

p.42

truth valueconsistency
Rubin2000

pp.12-13

Salcido Cota1995

p.44

nave realism

13

4
2002
1979 30

p.63
1980 2003
180

2000 (MSW)(MSSM)

200220

p. 66

14

Cautious Use

Critical Use
Gambrill, 1999, 2004; Gomory, 2000a, 2000b
Steier, 1991
Gambrill1999

Shek, Lam Tsoi2004

Criteria Formulation
Thyer1989
p.309
Guba Lincoln1994
p.114

Hammersley1990
realitytruth

15

truthp.61Hammersley 1992
truthrelevance
Seale, 2002
Smith Deemer2000
p.878

anything goes

p.878

p.894

Curriculum Development

(Shek, Lam & Tsoi, 2004)

Chinese Qualitative Research Database

Social Work Abstracts

16

Conducting High Quality Studies

Correcting Myths About Qualitative Research

17

nave realism

18

References
Anastas, J.W. (2004). Quality in qualitative evaluation: Issues and possible answers. Research
on Social Work Practice, 14(1), 57-65.
Atherton, C. R., & Bolland, K. A. (2002). Postmodernism: A dangerous illusion for social
work. International Social Work, 45, 421-433.
Batchelor, J., Gould, N., & Wright, J. (1999). Family centers: A focus for the children in need
debate. Child and Family Social Work, 4, 197-208.
Bloland, H. G. (1995). Postmodernism and higher education. Journal of Higher Education,
66(5), 521-559.
Bronstein, L. R., & Kelly, T. B. (1998). A multidimensional approach to evaluating
school-linked services: A school of social work and county public school partnership.
Social Work in Education, 20(3), 152-164.
Brun, C. (1997). The process and implications of doing qualitative research: An analysis of
54 doctoral dissertations. Journal of Sociology and Social Welfare, 24(4), 95-112.
Bryman, A. (1988). Quantity and quality in social research. London: Unwin Hyman.
Cigno, K., & Gore, J. (1999). A seamless service: Meeting the needs of children with
disabilities through a multi-agency approach. Child and Family Social Work, 4,
325-335.
Davis, D., Ray, J., & Sayles, C. (1995). Ropes Course Training for youth in a rural setting:
At first I thought it was going to be boring. Child and Adolescent Social Work
Journal, 12(6), 445-463.
De Anda, D. (2001). A qualitative evaluation of a mentor program for at-risk youth: The
participants perspective. Child and Adolescent Social Work Journal, 18(2), 97-117.
Dellgran, S., & Hojer, P. (2001). Mainstream is contextual: Swedish social work research
dissertations and theses. Social Work Research, 25(4), 243-252.
Denzin, K., & Lincoln, Y.S. (1998). The landscape of qualitative research. Thousand Oaks,
CA: Sage.
Denzin, N. K., & Lincoln, Y. S. (Eds.). (2000). Handbook of qualitative research. Thousand
Oaks, Calif.: Sage.
Derezotes, D. (1995). Evaluation of the Late Nite Basketball Project. Child and Adolescent
Social Work Journal, 12(1), 33-50.

19

Drisko, T.W. (1997). Strengthening qualitative studies and reports: Standards to promote
academic integrity. Journal of Social Work Education, 33(1), 185-197.
Dvell, F., & Jordan, B. (2001). How low can you go? Dilemmas of social work with
asylum seekers in London. Journal of Social Work Research and Evaluation, 2,
189-205.
Elks, M. A., & Kirkhart, K. E. (1993). Evaluating effectiveness from the practitioner
perspective. Social Work, 38(5), 554-563.
Erera, P. I. (1997). Empathy training for helping professionals: model and evaluation. Journal
of Social Work Education, 33(2), 245-260.
Gambrill, E. (2004). The future of evidence-based social work practice. In Thyer, B. and Kazi,
M. A. F. (Eds.), International Perspectives on Evidence-based Practice in Social Work
(pp. 215-234). London: Venture Press.
Gambrill, E. (1999). Evidence-based practice: An alternative to authority-based practice.
Families in Society, 80, 341-350.
Gardner, F. (2003). Critical reflection in community-based evaluation. Qualitative Social
Work, 2(2), 197-212.
Gellis, Z. D. (2001). Using a participatory research approach to mobilize immigrant minority
family caregivers. Journal of Social Work Research, 2, 267-282.
Goicoechea-Balbona, A., Barnaby, C., Ellis, I., & Foxworth, V. (2000). AIDS: the
development of a gender appropriate research intervention. Social Work in Health Care,
30(3), 19-37.
Gomory, T. (2000a). A fallibilistic response to Thyer's theory of theory-free empirical
research in social work practice. Journal of Social Work Education, 37, 26-50.
Gomory, T. (2000b). Critical realism (Gomory's blurry theory) or positivism (Thyer's
theoretical myopia): Which is the prescription for social work research? Journal of
Social Work Education, 37, 67-78.
Guba, E. G., & Lincoln, Y. S. (1981). Effective evaluation. San Francisco: Jossey-Bass.
Guba, E. G., & Lincoln, Y. S. (1994). Competing paradigms in qualitative research. In Denzin,
N.K. and Lincoln, Y.S. (Eds.), Handbook of qualitative research (pp.105-117).
Thousand Oaks, CA; Sage.
Hammersley, M. (1990). Reading ethnographic research: A critical guide. London: Longman.

20

Hammersley, M. (1992). Whats wrong with ethnography: Methodological explorations.


London: Routledge.
Heath, D. T. (1998). Qualitative analysis of private mediation: benefits for families in public
child welfare agencies. Children and Youth Services Review, 20(7), 605-627.
Hill, M., Dillane, J., Bannister, J., & Scott, S. (2002). Everybody needs good neighbours: An
evaluation of an intensive project for families facing eviction. Child and Family Social
Work, 7, 79-89.
Huberman, A. M., & Miles, M. B. (1994). Qualitative data analysis: An expanded sourcebook.
Thousand Oaks, Calif.: Sage.
Illback, R. J., & Kalafat, J. (1995). Initial evaluation of a school-based integrated service
program: Kentucky Family Resource and Youth Services Centers. Special Services in
the Schools, 10, 139-163.
Janesick, V.J. (1998). The dance of qualitative research design. In Denzin, N.K. and Lincoln,
Y.S. (Eds.), Strategies of qualitative inquiry (pp.35-55). Thousand Oaks, Calif.: Sage.
Lather, P. (1986). Issues of validity in openly ideological research: Between a rock and a soft
place. Interchange, 17(4), 63-84.
LeCompte, M.D., & Goetz, J.P. (1982). Problems of reliability and validity in ethnographic
research. Review of Educational Research, 52(1), 31-60.
Ligon, J., Markward, M. J., & Yegidis, B. L. (1999). Comparing student evaluations of
distance learning and standard classroom courses in graduate social work education.
Journal of Teaching in Social Work, 19, 21-29.
Lincoln, Y. S. (1998). From understanding to action: New imperatives, new criteria, new
methods for interpretive researchers. Theory and Research in Social Education, 26(1),
12-29.
Lincoln, Y. S., & Guba, E. G. (2000). Paradigmatic controversies, contradictions and
emerging confluences. In Denzin, N.K. and Lincoln, Y.S. (Eds.), Handbook of
qualitative research (p.163-188). Thousand Oaks, Calif.: Sage.
Mowbray, C. T., & Bybee, D. (1998). The importance of context in understanding
homelessness and mental illness: Lessons learned from a research demonstration project.
Research on Social Work Practice, 8, 172-199.
Netting, F. E., & Williams, F. G. (2000). Expanding the boundaries of primary care for elderly
people. Health and Social Work, 25, 233-242.

21

Patton, M. Q. (1990). Qualitative evaluation and research methods. Newbury Park, Calif.:
Sage.
Patton, M. Q. (2002). Two decades of developments in qualitative inquiry. Qualitative Social
Work, 1(3), 261-283.
Platt, D. (2001). Refocusing childrens services: Evaluation of an initial assessment process.
Child and Family Social Work, 6, 139-148.
Rehr, H., & Epstein, I. (1993). Evaluating the Mount Sinai Leadership Enhancement Program:
a developmental perspective. Social Work in Health Care, 18, 79-99.
Ringma, C., & Brown, C. (1991). Hermeneutics and the social sciences: An evaluation of the
function of hermeneutics in a consumer disability study. Journal of Sociology and
Social Welfare, 18, 57-73.
Rivard, J. C., Johnsen, M. C., Morrissey, J. P., & Starrett, B. E. (1999). The dynamics of
interagency collaboration: how linkages develop for child welfare and juvenile justice
sectors in a system of care demonstration. Journal of Social Service Research, 25,
61-82.
Rizzo, T. A., Corsaro, W. A., & Bates, J. E. (1992). Ethnographic methods and interpretive
analysis: Expanding the methodological options of psychologists. Developmental
Review, 12, 101-123.
Roberts, C.A. (1989). Research methods taught and utilized in social work. Journal of Social
Service Research, 13(1), 65-86.
Rosenau, P. M. (1992). Post-modernism and the social sciences insights, inroads and
intrusions. Princeton, NJ: Princeton University Press.
Rubin, A. (2000). Editorial: Social work research at the turn of the millennium: Progress and
challenges. Research on Social Work Practice, 10, 9-14.
Ryan, K. (1998). Advantages and challenges of using inclusive evaluation approaches in
evaluation practice. American Journal of Evaluation, 19(1), 101-122.
Salcido, R. M., & Cota, V. (1995). Cross-cultural training for child welfare workers when
working with Mexican-American clients. Journal of Continuing Social Work Education,
6, 39-46.
Sandelowski, M. (1986). The problem of vigor in qualitative research. Advances in Nursing
Science, 8(3), 27-37.
Schwandt, T. A. (1996). Farewell to criteriology. Qualitative Inquiry, 2(1), 58-72.

22

Schwandt, T. A. (2000). Three epistemological stances for qualitative inquiry: Interpretivism,


hermeneutics, and social constructionism. In Denzin, N.K. and Lincoln, Y.S. (Eds.),
Handbook of qualitative research (p.189-213). Thousand Oaks, Calif.: Sage.
Seale, C. (2002). Quality issues in qualitative inquiry. Qualitative Social Work, 1, 97-110.
Shek, D. T. L., Lam, M. C., and Tsoi, K. W. (2004). Evidence-based practice in Hong Kong.
In Thyer, B. and Kazi, M. A. F. (Eds.), International Perspectives on Evidence-based
Practice in Social Work (pp. 167-181).

London: Venture Press.

Smith, J. K., & Deemer, D. K. (2000). The problem of criteria in the age of relativism. In
Denzin, N.K. and Lincoln, Y.S. (Eds.), Handbook of qualitative research (pp.877-896).
Thousand Oaks, CA; Sage.
Sokal, A., & Bricmont, J. (1998). Fashionable nonsense: Postmodern intellectuals sense of
science. New York: Picador USA.
Steier, F. (Ed.). (1991). Research and reflexivity. London: Sage.
Stout, K. D. (1991). A continuum of male controls and violence against women: A teaching
model. Journal of Social Work Education, 27, 305-319.
Sun, A. (1998). Organizing an international continuing education program: A preliminary
study based on participants perspectives. Journal of Continuing Social Work Education,
6(4), 23-29.
Thyer, B.A. (1989). First principles of practice research. British Journal of Social Work, 19,
309-323.
Tschudi, F. (1989). Do qualitative and quantitative methods require different approaches to
validity? In Kvale, S. (Ed.), Issues of validity in qualitative research (p.109-134). Lund,
Sweden: Studentlitteratur.
Vera, M. I. (1990). Effects of divorce groups on individual adjustment: A multiple
methodology approach. Social Work Research and Abstracts, 26, 11-20.
Walsh-Bowers, R., & Basso, R. (1999). Improving early adolescents peer relations through
classroom creative drama: an integrated approach. Social Work in Education, 21, 23-32.
Whipple, E. E., Grettenberger, S. E., & Flynn, M. S. (1997). Doctoral research education in a
social context: Development and implementation of a model outreach course. Journal of
Teaching in Social Work, 14, 3-26.
2002262-66

23

1. 1 5

1990 10

11 6 5

Ringma C. 1991
& Brown

C.

24
.

Stout K.D. 1991

I25 I

Vera M.I.

II II
12

Elks M.A. 1993


& Kirkhart

K.E.

17

11 6

Rehr H. & 1993 Mount Sinai

Epstein I.

160 3 1

2
3

2 14 2

3 39 .

24

1. ()

1995
Davis
D. Ray

J. & Sayles

C.

266

Derezotes
D.

1995

Illback R.J 1995


& Kalafat

J.

1 6

3
45

151
1 3
226 2

328

1 370
1 25555

2 22

Salcido
R.M. &
Cota V.

1995

436 9
.

53 240
61
23
36

25

1. ()

10

Erera P.I.

1997
ETP

I
.

ETP

22
18

4
5

II 4

III 4

29
20

11

1997
Whipple
E.E.

Grettenberg
er S.E. &
Flynn
M.S.

(1)
(2)

10

12

Mowbray
C.T. &
Bybee D.

1998

163

(1)
(2)
Client
Level Assessment Measure

(3)
(4)
(5) 12
12

26

1. ()

13

Heath D.T. 1998

126 1 Key

Informant Survey

4
5

( 1)

245 2
5 23
/
/
12
3
1

14

Bronstein 1998
L.R. &

Kelly T.B.
-

19 11
6

2 2

30
4

15

Sun A

1998

22

27

1. ()

16

17

Walsh
-Bowers
R. &
Basso R.

1999

Batchelor 1999
J. Gould

N. & Wright

J.

I 15

12 12
7

1
I
2 5
5
3
4

II 15

17
11
7

II1 10
2 10

3
4

5
I

1 1
36

29 2

30
2 2

3 3

28

4
5

1. ()

18

Cigno K. 1999
& Gore J.

151
4

iii
/ 3
iii
iv
3

4
5

275 1

362 275 5
47
15
6 2
19

Ligon J. 1999
Markward

M.J. &

Yegidis B.

L.

1 14 120
122

2 38 27

20

Rivard
J.C.
Johnsen
M.C.
Morrissey
J.P. &
Starrett
B.E.

1999

1
2

3
6

63

74

21

Goicoechea 2000
-Balbona
HIV
A.

Barnaby

C. Ellis
I. &
Foxworth
V.

9 HIV

29

1. ()

22

Netting
F.E. &
Williams
F.G.

2000 9

1105
40 32
23

289
44 24
21

56 1

23

De Anda
D.

2001

mentor

18 2 1~
2 2 ~

24

Dvell F.
& Jordan
B.

2001

4 4
2
2

25

Gellis
Z.D.

2001

35

29
20

30

1. ()

26

Platt D.

2001


36
22
41 47

47 10
10
8

27

Hill M. 2002
Dillane

J.
Bannister
J. & Scott
S.

19

23
53 20 53 20 //
//
24 10 24 10 //
//
3 3

4 4
51 126
63 70 51 63 70

5 5

28

Gardner F. 2003

Sharon Action Rural


Access Project



Sharon Action

0.96

1.00

0.96

0.96

0.96

0.82

31

2. 6 12
6

10

11

12

audit

trail

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

0.96

0.96

1.00

1.00

0.96

1.00

0.96

0.96

0.86

0.96

32

99(2)

2011/05/11

2005
19902003
241 70100

Patton2002

1. positivisticparadigm

LeCompte Goetz1982
1. internalreliability

2. externalreliability

3.
4.
5.

2.

internalvalidity
externalvalidity

Tschudi1989triangulation

constructivistparadigm

Guba Lincoln1981
1. credibility
2. fittingness

3. auditability

4. confirmability

Sandelowski1986
12
3
45

67
1

99(2)

2011/05/11

3.

4.

artisticandevocativeparadigm

criticalparadigm

Huberman Miles1994
5

1.
2.
3.
4.
5.

Drisko1997

1.

2. 1

3.

Drisko1997

99(2)

4.
5.
6.

5.
6.
7.

2011/05/11

Anastas2004
1.

2.

3.
4.

99(2)

2011/05/11

general
Patton,1990specific

Huberman&Miles,1994
Drisko1997

intraraterreliability

interraterreliability

99(2)

2011/05/11


peer checking
communal
memberchecking

audittrail

Sandelowski1986

Huberman&Miles,1994

Anastas2004


10

11



negativecases

12

Gambrill,1999;Gomory,2000a,2000b
Lincoln&Guba,2000;Ryan,1998

Nursing and Health Sciences (1999), 1, 147153

Research Article

Challenges of the grounded theory approach to


a novice researcher
Kaisa Backman, rn, mnsc and Helvi A. Kyngs, rn, phd
Department of Nursing and Health Administration, University of Oulu, Oulu University Hospital,
Oulu, Finland

Abstract

The grounded theory approach has been used in nursing research since 1970. The latest
methodological books describe the research process in detail. However, there are many problems
involved in the grounded theory approach, which especially need to be considered by a novice
researcher. One of these problems is the question of how deeply and widely the researcher
should familiarize her- or himself with the research topic before the empirical study. The
problems also include the need to focus the research problem and to choose the sampling
method. Data analysis is a multistage process, which demands from the researcher both
sensitivity and time to work out the findings which emerge from the data. In this paper, the
grounded theory approach is described as a process undertaken by the novice researcher.
The purpose of this paper is to discuss the challenges of the grounded theory approach and the
problems encountered by a researcher using the method for the first time.

Key words

grounded theory approach, inductive research, method of nursing research, research method.

INTRODUCTION
The grounded theory approach is both a way to do
qualitative research and a way to create inductive
theory. The approach was developed by the sociologists Glaser and Strauss in the USA. Their first
book Discovery of Grounded Theory was published in
1967. Awareness of the fact that the grounded theory
approach comes from sociology helps one to understand better some of its steps and processes. The
grounded theory approach has traditionally been
part of the postpositivist inquiry paradigm, but
it approaches the constructivist inquiry paradigm
(Annells, 1996). The approach has been further
developed by Glaser (1978), Chenitz and Swanson
(1986) and Strauss and Corbin (1990). Using the
grounded theory approach, it is possible to study
the meanings of events for people. This is based on

Correspondence address: Kaisa Backman, Department of Nursing and


Health Administration, University of Oulu, Box 5000, 90401 Oulu,
Finland. Email: kaisa.backman@oulu.fi
Received 3 March 1999; revised 20 May 1999; accepted 7 June 1999.

the assumption that meanings must be shared and this


sharing is accomplished via a common language and
socialization (Chenitz & Swanson, 1986). Grounded
theory research addresses social processes composed
of meanings, which are meant to be clarified and
made public (Glaser, 1978; Chenitz & Swanson, 1986;
Smith & Biley, 1997).
The grounded theory approach has been used in
nursing research since 1970. The studies have focused
especially on nursing practice and nursing education
(May, 1979; Duffy, 1983; Janhonen, 1992; Brandriet,
1994). The reason for the choice of the approach
has usually been that it is a way to create theory
(Glaser, 1978; Baker et al., 1992; Olshansky, 1996;
Bailey, 1997).
A theory is composed of concepts and their mutual
connections. It is a systematic structure which describes, explains and/or controls phenomena (Walker
& Avant, 1988; Meleis, 1991). According to Glaser and
Strauss (1967), grounded theory is inductively derived
from the study of the phenomenon it represents. It
may be either substantive or formal. A substantive
theory is relevant to the people concerned and is
readily modifiable (Glaser, 1978). A formal theory

148

is developed further than a substantive theory. It


meets the criteria of fit, relevance and easy modification (Strauss & Corbin, 1990). Relevance means
that the grounded theory approach allows core
problems and processes to emerge. To aid the effort
to arrive at the core categories, Glaser and Strauss
(1967) have developed the notion of Basic Social
Process (BSP), which explains a considerable portion of the action in an area and relates to most
categories of lesser weight used to make the theory
workable.
In studies utilizing the grounded theory approach,
the discovered grounded theory is generally called a
theory (Chen, 1996; Frenn, 1996) or a model (Wilde
et al., 1994; Kyngs & Hentinen, 1995; Backman &
Hentinen, 1997). In the scientific sense, a model may
be used to define or describe something and to
specify relationships and processes, while a theory is
a systematically related set of statements, including
law-like generations, which is empirically testable
(Lancaster & Lancaster, 1981). It is also possible to
use two or more methodological approaches to
generate theory. For example, Sandelowski (1995a)
formulated a theory about the transition of infertile
couples to parenthood based on three studies, of
which one utilized grounded theory.
Studies where grounded theory is created are called
the grounded theory approach (Strauss & Corbin,
1990), grounded theory research (Glaser, 1978) or the
grounded theory method (Chenitz & Swanson, 1986).
The approach of research could be called the paradigm of inquiry within which the method resides.
It implies certain philosophical backgrounds which
direct the research process. In that way the grounded
theory can be seen as an approach (Smith & Biley,
1997).
When a person reads for the first time about the
grounded theory approach, it may seem quite clear
and straightforward. Especially the new methodological books (i.e. Strauss & Corbin, 1990) describe
the analysis process in notable detail. According to
Glaser (1978), however, grounded theory research is a
many-faceted process. It requires time and theoretical
sensitivity to move from the data to the theory and
back. There are many problems involved in the
grounded theory approach, which especially need to
be considered by a novice researcher. In this paper,
the grounded theory approach is described as a process undertaken by the researcher. The purpose of
this paper is to discuss the challenges of the grounded
theory approach and the problems encountered by a
researcher using the method for the first time. No
research examples are given, and ethical considerations are also omitted.

K. Backman and H. Kyngs

GROUNDED THEORY APPROACH AS


A RESEARCHERS PROCESS
According to the traditional research practice, a
research process is imitated by identifying the
phenomena and by naming the research problems.
After the data have been collected and analysed,
the results and conclusions are reported. When the
grounded theory approach is used, however, the
research does not necessarily follow the chronological
stages of the traditional research process. Data collection, data analysis and the formulation of grounded
theory often take place at the same time. This may
be problematic for the researcher, because it may
cause difficulties in shaping the research process as a
whole.

Reading the literature and formulating the


study questions
A grounded theory is inductively derived from the
process of study. One does not begin with a hypothetical theory and, then prove it. Rather, one begins
by collecting the data in the field first. Then the
researcher starts analysing the data and generating a
theory. (Strauss & Corbin, 1990.) According to Glaser
(1978), the presupposition of an inductive approach
is that the researcher has as few preconceived ideas
about the research phenomena as possible. There is
the risk of a biased interpretation of the data, if one is
too imbued with concepts from the literature. This
detachment may, however, be quite difficult for a
novice researcher because reading the literature
usually helps to clear up ones thoughts and to narrow
down the topic of research.
The grounded theory approach may also be used
when there is already some knowledge about the
research phenomenon, but a new point of view is
sought (Strauss & Corbin, 1990; Bailey, 1997; Smith
& Biley, 1997). In that case, the researcher needs
to get familiar with the previous knowledge so as to
be able to outline the research phenomenon. The
problem is that previous knowledge may direct the
research and make it more difficult to find a
new point of view (Olshansky, 1996.) This requires
bracketing. To bracket, the researcher must identify
and suspend what he/she already knows about the
experience being studied and approach the data without preconceptions. This could be particularly difficult
to a novice researcher, because he/she has little experience about the emotions involved in data collection and analysis in qualitative research. However,
it has also been said that, from the grounded theory
perspective, the researcher is a social being who also
creates and recreates social processes. Therefore,

Grounded theory approach

previous experiences are also data. No effort is made


to put aside ideas or assumptions about the situation
being studied. On the contrary, the researcher uses
them in order to understand better the processes
being observed (Baker et al., 1992). Sandelowski
(1993) points out that although a theoretical orientation may not be explicitly stated in a qualitative
project, or may even be denied, it is always implicit in
the way a problem is presented, in the way the literature is reviewed and, most importantly, in the
selection and description of the method itself.
According to Strauss and Corbin (1990), the literature
review also implicitly displays the authors theoretical
sensitivity toward the target phenomenon, and the
qualitative method itself is based on a priori assumptions about and interpretations of inquiry.
The research questions in grounded theory research are statements that identify the phenomenon
to be studied. The questions are formulated so that
they give the researcher the flexibility and freedom
to explore the phenomenon in depth (Glaser, 1978;
Strauss & Corbin, 1990). Concepts which strictly
narrow down the research questions easily direct the
study deductively. On the other hand, clean-cut and
well-defined concepts make it easier for a novice
researcher to maintain the logic of the study. If the
research questions are very flexible and the researcher begins data collection by interviewing, the
choice of suitable themes may also be problematic.
During the first interviews the researcher may wonder
whether the essential matters are discussed or not.
According to Glaser (1978), the researcher is not able
to know beforehand what the essential matters are
and the research question may even change during
data collection.

Data collection
The data are generally collected by using interviews,
observations, diaries or other written documents
or a combination of some methods (Glaser, 1978;
Chenitz & Swanson, 1986; Strauss & Corbin, 1990).
The grounded theory literature emphasizes (Glaser,
1978; Chenitz & Swanson, 1986; Strauss & Corbin,
1990) the need to combine many data collection
methods. In nursing studies, however, mainly only
interview data are used. In that case, the research
does not necessarily account for the social structural
influences on the experiences of the respondents
(Benoliel, 1996).
Theoretical sampling refers to a data collection
process which aims to create theory, which means that
coding and analysing serve as data collection in the
next phase. When theoretical sampling is used, data

149

collection is started by concentrating on the social


processes that appear inherently the most interesting.
After preliminary data collection and analyses, the
sampling can be made more selective, because preliminary hypotheses which describe the phenomenon
emerge from the data and promote the data collection further (Glaser, 1978).
According to Glaser (1978), theoretical sampling
should be used in the grounded theory approach
(also Glaser & Strauss, 1967; Strauss & Corbin, 1990;
Olshansky, 1996). Nevertheless, many researchers
decide to use a selective sample. In that case, the subjects are mainly chosen before the data collection,
and the data are collected within a certain time.
Selective sampling is defended on the grounds that it
is impossible to address the whole research phenomenon. Using the selective sampling, the researcher is
able to make his/her choices based on his/her interest
and the possible restrictions placed on his/her research work (Schatzman & Strauss, 1973). A programme of research employing grounded theory
typically begins with a selective sampling strategy
aimed at phenomenological variation and then proceeds to theoretical sampling (Sandelowski, 1995b).
If the data are collected within a certain time and
not analysed simultaneously, it is difficult to determine the theoretical shape and to recognize the
saturation. Simultaneous collection and analysis of
data and the emergency theoretical structure help to
orient further data collection. It helps to find key
words and key persons, to outline the research phenomenon and to organize the process in an attempt to
control the study (Glaser & Strauss, 1967; Glaser,
1978). This is important for a novice researcher. Using
theoretical sampling, the researcher may, however,
make conclusions that are too firm, based on his/her
preliminary analyses, and this may influence too
strongly the further data collection and the emerging
theory.

Data analysis
Data analysis is like a discussion between the actual
data, the created theory, the memos and the researcher. Such discussion takes place when the data
are broken down, conceptualized and put back
together in new ways. The data give rise to the codes
and the categories which combine the codes. The
categories and hypotheses must be verified against
the data by comparing the categories with each
other, with the data and with the researchers conclusions (Glaser & Strauss, 1967; Glaser, 1978;
Chenitz & Swanson, 1986; Strauss & Corbin, 1990).
If the researcher has preliminary assumptions or

150

knowledge about the research phenomenon, then that


could have an effect on the analysis. For a novice researcher, however, it is useful to have some previous
knowledge, because it can help the analysis. In that
case the grounded theory is not created exclusively
based on actual data, but it may also be based on
previous knowledge about the research phenomenon.
Some authors (Chenitz & Swanson, 1986; Smith &
Biley, 1997) point out that the previous knowledge
should be used in a phase when the researcher compares her/his theory with the previous theories.
The data analysis can also be seen as the researchers process. According to Glaser (1978), the
process has three phases. The first phase is called
input. In that phase the data move as part of the
researchers thinking. In the second phase the data
are in the researchers mind. He/she has a lot of different ideas concerning the theory, but nothing seems
clear. This is called a drugless trip. The last phase,
called saturation, is the most important for theory
development. In this phase the researcher writes
down the results of the analysis and makes his/her
conclusions.
Data analysis is started during the data collection.
In this phase the researcher identifies the research
phenomenon. The process continues while coding the
data. During the analysis, categories are identified and
developed in terms of their properties and dimensions
through a process involving the generation of basic
categories to describe features of the data and constant comparisons between cases, instances and categories. Similar events and incidents are grouped
together into categories. (Glaser, 1978; Chenitz &
Swanson, 1986; Strauss & Corbin, 1990).
In the input phase it may be difficult to understand
how the codes are connected to others. This may
cause a feeling that the data do not describe the
research phenomenon. According to Glaser (1978),
this is a consequence of the fact that the researcher
is processing the data unconsciously and is unable yet
to identify and write down codes and categories.
Generally, the researcher wants to discuss the organization of the data with someone else. Discussion can
help the researcher to continue the analyses, but
he/she may also feel confused, because other peoples
opinions may be so different from hers/his and it can
require energy to continue the work. Glaser (1978)
presents as a solution that the researcher should read
the data again and again and continue coding even
if he/she is very unsure about the analysis. His
advice is that the researcher should write down
his/her thoughts and discuss them later with other
people. A novice researcher may experience this
phase as very chaotic. He/she has a lot of ideas about

K. Backman and H. Kyngs

the construction of theory, but he/she is unable to


make discoveries yet.
In the second phase of the drugless trip, the
researcher gradually identifies the grounded theory.
He/she studies the connections between the categories. One method to find them is axial coding. It is
a complex process of inductive and deductive thinking involving several steps. It puts the data back
together in new ways by making connections between
a category and its subcategories (Strauss & Corbin,
1990). The focus of axial coding is on specifying
a category in the context in which it had appeared.
Axial coding can be done in accordance with the 18
paradigms presented by Glaser (1978). One of these
paradigms is the 6 C paradigm model (Chenitz &
Swanson, 1986), where the categories are studied in
terms of their contexts, consequences, causes, conditions, covariance and contingents (Glaser, 1978;
Chenitz & Swanson, 1986; Strauss & Corbin, 1990;
Munchall & Oiler, 1993). When doing axial coding
by the 6 C paradigm, it is difficult to decide the differences between the causes and conditions. According to Mackie (1985), the cause of some action can
also be its condition. Furthermore, the definition of
covariance in axial coding may be difficult to understand. How can we indicate that some categories appear at the same time? Axial coding is a laborious
and time-consuming process. During that process the
researcher concretely discusses with his/her data, but
on the other hand, the process is very abstract and
takes place deep in the researchers mind and partly
unconsciously. The researcher must tolerate the
uncertainty of not finding connections between the
categories or feeling that they describe the research
phenomenon inadequately (Glaser, 1978). On the
other hand, many things take place fast or appear to
do so in grounded theory. Categories may emerge
before a few interviews are over, and one or two can
be overriding or consist of core contents that seem
to wrap up the whole study. The researcher should,
however, put the brake on these premature forays
because the generating process takes time (Glaser,
1978).
Gradually, the researcher gets to know what kind
of theory is emerging. The problem can be that the
researcher has so many ideas about the theory that
he/she misses the essential points. Glaser (1978)
advises that in this phase the researcher should write
down all his/her ideas. To discover the final categories,
the researcher returns to the data again and again and
makes sure that the categories are based on data
and have connections (Strauss & Corbin, 1990). Data
collection, analyses and theory stand in reciprocal
relationship with each other.

Grounded theory approach

The goal of grounded theory is to generate a theory


that accounts for a pattern of behaviour which is
relevant for those involved. The generation of theory
takes place around a core category (Glaser, 1978). A
novice researcher may find it difficult to look for
a core variable when coding his/her data, because
switching ones focus from studying a unit to studying
a process is painful. It takes time and much coding
and analysis to verify a core category through saturation, relevance and workability.
Basic Social Processes are one type of core
category. Basic Social Processes implies change over
time, and two types were identified by Glaser (1978)
as of interest to social scientists: (i) basic psychological process (BSPP) occurring to individuals and/or
groups; and (ii) basic social structural process (BSSP)
referring to changes in social structural arrangements.
A major concern for the inves-tigator is to determine
how contextual features of the environment influence
the direction and form of the identified social process
(Benoliel, 1996).
While analysing, the researcher should write theoretical memos, which include the ideas that he/she has
during the process. The memos help to develop the
characteristics of categories and to integrate them, to
create hypotheses and theory (Chenitz & Swanson,
1986; Strauss & Corbin, 1990; Anderson, 1991; Baker
et al., 1992). The researcher may be tired and disappointed in this phase. The discovered theory may
seem simple, and the researcher may not want to read
his/her texts. According to Glaser (1978), it is good
to take a pause and to discuss the theory with other
people and write more memos before writing the final
results.

The presentation of the results and


their reliability
One key element in the successful dissemination
and utilization of qualitative findings is a well-written
research report. In contrast to quantitative research,
there is no single style for reporting the findings
of qualitative research. Qualitative researchers must
select from an array of representational styles and
formats those that best fit their research purposes,
methods, and data. The purpose of a grounded theory
study is to emphasize the researchers theoretical reformulation of the data, while the data themselves
only appear to support the theory. Data are only used
to show how a theory was constructed, and that it
was indeed constructed from this data (Sandelowski,
1998a).
Glaser and Strauss (1967) point out that the results
section of a report should start by presenting first the

151

discovered grounded theory and the categories. The


presentation of the results is a challenge for the
researcher, and it has been done in many different
ways. According to Sandelowski (1998a), there are
three ways of presenting the findings from grounded
theory studies: (i) coding families; (ii) typology family;
and (iii) the use of strategy while coding a family.
The researcher develops his/her own style to organize
the results of his/her study. The results are discovered
via a multiphase process. The researcher is able to
describe part of this process accurately, but part
of the process has taken place unconsciously in the
researchers mind, and this part of the process is
difficult to write down. This may cause problems to
the readers, because they are unable to follow the
way in which the results have been discovered and
to understand the connection between the data and
the results. For this reason, the readers may consider
the discovered theory unreliable.
According to Glaser (1978), there are two essential
factors to evaluate in grounded theory. The theory
must have fit and relevance and it must work.
Fit means that the categories of the theory have a
connection to the data. It is not permissible to force
the data to the categories which the researcher has
discovered. The connection between the theory and
the categories has to be confirmed by continuous
comparative analysis, which means that the researcher
has to compare, during the whole process, first the
substantive codes and their properties, later the categories and their contents, and finally the discovered
theory with other theories. Work means the theorys
ability to explain the phenomenon and to predict
and interpret actions which are connected to the phenomenon. Especially the working of the theory is
difficult to establish. The discovered grounded theory
may need to be tested before it can be used for
predicting.
The purpose of the grounded theory approach is to
create a theory which has connections with the data.
The instructions for the analysis process emphasize
that the connection with data should be maintained
throughout the whole process. This requirement may
prevent the researcher from conceptualizing the data
and from formulating abstract categories and discovering theory. The problem is how the researcher is
able to disengage from the data to create theory. If
he/she is unable to do that, he/she may discover a
theory which is naive, concrete and written by using
the same terms as in the data. In that case the discovered theory may be simplistic and ill-constructed.
One possibility to validate the results is to use an
expert, someone outside the project, with special
knowledge, who warrants the study as good and

152

true. Sandelowski (1998b), however, points out that


an outsider-expert can only serve as an advisor,
trouble-shooter, and peer-debriefer. They cannot
evaluate the interpretation made by the researcher.

DISCUSSION
There are many methodological books and articles
concerning the grounded theory approach in nursing
science, but also in other sciences (Glaser & Strauss,
1967; Glaser, 1978; Chenitz & Swanson, 1986;
Charmaz, 1990; Strauss & Corbin, 1990; Anderson,
1991; Hutchinson, 1993; Keddy et al., 1996; Wilson
& Hutchinson, 1996). The essential point for a
novice researcher is whether he/she is committed to
certain methodological books and their instructions
or tries to apply several different instructions and
views. There are different ways of discovering the
grounded theory, and this may confuse a novice
researcher during the complicated and difficult research process.
The use of the grounded theory approach requires
a novice to commit him/herself to a time-consuming
and long process. To discover the grounded theory
assumes a dialogue between the data and the theory.
For a novice researcher, applying the grounded
theory approach is more or less a compromise
between the demands of the approach and the
resources which he/she has available.
The literature does not contain many descriptions
about the grounded theory approach as a researchers
process. It often surprises the researcher by its
challenges. During the past few years the different
ways of using the grounded theory approach have
also been discussed (Annells, 1996; Melia, 1996).
In their book titled Basics of Qualitative Research,
Strauss and Corbin (1990) describe systematically
how to analyse the data. The method can be a good
tool for a novice, but it may also hinder the way to
create inductive theory. Burns (1989) considers the
basis of evaluating qualitative research and emphasizes both flexibility and accuracy. Probably the
creation of theory by the means of the data and the
simultaneous observance of instructions is the biggest
challenge of the grounded theory approach.

REFERENCES
Anderson MA. Use of grounded theory methodology in a
descriptive research design. ABNF J. 1991; 2: 2832.
Annells M. Grounded theory method: philosophical perspectives, paradigm of inquiry and postmodernism.
Qualitative Hlth Res. 1996; 6: 379393.

K. Backman and H. Kyngs

Backman K, Hentinen M. Model for the self-care of homedwelling elderly. J. Advanced Nurs. 1999; 30 (in press).
Bailey P. Finding your way around qualitative methods in
nursing research. J. Advanced Nurs. 1997; 25: 1822.
Baker C, Wuest J, Stern N. Method slurring: The grounded
theory/phenomenology example. J. Advanced Nurs. 1992;
17: 13551360.
Benoliel JQ. Grounded theory and nursing knowledge.
Qualitative Hlth Res. 1996; 6: 406428.
Brandriet LM. Gerontological nursing: application of ethnography and grounded theory. J. Gerontol. Nurs. 1994; 20:
3340.
Burns N. Standards for qualitative research. Nurs. Sci. Q
1989; 2: 4452.
Charmaz K. Discovering chronic illness: using grounded
theory. Social Sci. Med. 1990; 30: 11611172.
Chen YL. Conformity with nature: a theory of Chinese
American elders health promotion and illness prevention
processes. Adv. Nurs. Sci. 1996; 19: 1726.
Chenitz WC, Swanson JM. From Practice to Grounded
Theory. Qualitative Research in Nursing. Massachusetts:
Addison-Wesley, 1986.
Duffy M. Transcending options: creating a milieu for practising high level wellness. PhD Thesis. Salt Lake City. The
University of Utah, 1983 (unpubl.).
Frenn M. Older adults experience of health promotion: a
theory for nursing practice. Public Hlth Nurs. 1996; 13:
6571.
Glaser BG. Theoretical Sensitivity. San Francisco, USA: The
Sociology Press, 1978.
Glaser BG. Strauss AL. The Discovery of Grounded Theory:
Strategies for Qualitative Research. Chicago: Aldine, 1967.
Hutchinson SA. Qualitative Approaches in Nursing
Research. Grounded Theory: the Method. Nln Publications
1993; 192535: 180212.
Janhonen S. The Core of Nursing as Seen by Nurse Teachers
in Finland, Norway and Sweden. Acta Universitatis
Ouluensis. Series D. Medica 245. Oulu: University of Oulu
Printing Center. 1992.
Keddy B, Sims SL, Stern PN. Grounded theory as a feminist
research methodology. J. Advanced Nurs. 1996; 23: 448
453.
Kyngs H, Hentinen M. Meaning attached to compliance
with self-care, and conditions for compliance among
young diabetics. J. Advanced Nurs. 1995; 21: 729736.
Lancaster W, Lancaster J. Models and model building in
nursing. Adv. Nurs. Sci. 1981; 3: 3142.
Mackie JL. Causes and conditions. Am. Philosoph. Quart.
1995; 2: 245264.
May K. Management of detachment and involvement in
pregnancy by first-time expectant fathers. PhD Thesis. San
Francisco: The University of California, 1979.
Meleis AF. Theoretical Nursing: Development and Progress.
Philadelphia: J.B. Lippincott Company, 1991.
Melia KM. Rediscovering Glaser. Qualitative Hlth Res.
1996; 6: 368377.
Munchall PL, Oiler C. Nursing Research. A Qualitative
Perspective. New York: USA: National League for
Nursing Press, 1993.

Grounded theory approach

Olshansky EF. Theoretical issues in building a grounded


theory: Application of an example of a program of
research on infertility. Qualitative Hlth Res. 1996; 6: 394
405.
Sandelowski M. Theory unmasked: The uses and guises of
theory in qualitative research. Res. Nurs. Hlth 1993; 16:
213218.
Sandelowski M. A theory of the transition to parenthood of
infertile couples. Res. Nurs. Hlth 1995a; 18: 123132.
Sandelowski M. Sample size in qualitative research. Res.
Nurs. Hlth 1995b; 18: 179183.
Sandelowski M. Writing a good read: Strategies for representing qualitative data. Res. Nurs. Hlth 1998a; 21: 375
382.
Sandelowski M. The call to experts in qualitative research.
Res. Nurs. Hlth 1998b; 21: 467471.

153

Schatzman L, Strauss AL. Field Research. Strategies for a


Natural Sociology. New Jersey: Prentice-Hall Englewood
Cliffs, 1973.
Smith K, Biley F. Understanding grounded theory: principles and evaluation. Nurse Researcher 1997; 4: 17
30.
Strauss A, Corbin J. Basics of Qualitative Research.
Grounded Theory Procedures and Techniques. Newbury
Park, USA: Stage Publishing Company, 1990.
Walker L, Avant K. Strategies for Theory Construction in
Nursing. Norwalk: Appleton & Lange, 1988.
Wilde B, Larsson G, Larsson M, Starrin B. Development of
a patient-centred questionnaire based on a grounded
theory model. Scand. J. Caring Sci. 1994; 8: 3748.
Wilson HS, Hutchinson SA. Methodological mistakes in
grounded theory. Nurs. Res. 1996; 45: 122124.

99

T26974070

2011 5 11

Challenges of the grounded theory approach to a novice researcher


1.

2.


1.

theoretical sensitivity
priori assumptions

2.


Backman, K., & Kyngs, H. A. (1999). Challenges of the grounded theory approach to a novice
researcher. Nursing and Health Sciences, 1, 147153.

1.

2. theoretical
sampling

selective sampling


1.

2.

3.


1. Glaser1978input
drugless trip1saturation

drugless trip

(1)

(2)
axial coding

Backman, K., & Kyngs, H. A. (1999). Challenges of the grounded theory approach to a novice
researcher. Nursing and Health Sciences, 1, 147153.

1.

2.
(1)

(2)

basic social process


basic
psychological processbasic social
structural process

theoretical memos


1.

2.

3.

Backman, K., & Kyngs, H. A. (1999). Challenges of the grounded theory approach to a novice
researcher. Nursing and Health Sciences, 1, 147153.

1. coding
familiestypology family
the use of strategy while coding a family
2. fit and
relevance
work
(1) fit and relevance
continuous comparative analysis

(2) work

3.

4.

drugless trip

Glaser1978,
1998

http://lonelydissertator.blogspot.com/2009/07/classic-grounded-theory-community-and.html

Backman, K., & Kyngs, H. A. (1999). Challenges of the grounded theory approach to a novice
researcher. Nursing and Health Sciences, 1, 147153.

International Journal for Quality in Health Care; Volume 19, Number 6: pp. 349 357
Advance Access Publication: 14 September 2007

10.1093/intqhc/mzm042

Consolidated criteria for reporting


qualitative research (COREQ): a 32-item
checklist for interviews and focus groups
ALLISON TONG1,2, PETER SAINSBURY1,3 AND JONATHAN CRAIG1,2
1
School of Public Health, University of Sydney, NSW 2006, Australia, 2Centre for Kidney Research, The Childrens Hospital at Westmead,
NSW 2145, Australia, and 3Population Health, Sydney South West Area Health Service, NSW 2170, Australia

Abstract
Background. Qualitative research explores complex phenomena encountered by clinicians, health care providers, policy
makers and consumers. Although partial checklists are available, no consolidated reporting framework exists for any type of
qualitative design.
Objective. To develop a checklist for explicit and comprehensive reporting of qualitative studies (indepth interviews and
focus groups).

Results. Items most frequently included in the checklists related to sampling method, setting for data collection, method of data
collection, respondent validation of ndings, method of recording data, description of the derivation of themes and inclusion of
supporting quotations. We grouped all items into three domains: (i) research team and reexivity, (ii) study design and (iii) data
analysis and reporting.
Conclusions. The criteria included in COREQ, a 32-item checklist, can help researchers to report important aspects of the
research team, study methods, context of the study, ndings, analysis and interpretations.
Keywords: focus groups, interviews, qualitative research, research design

Qualitative research explores complex phenomena encountered


by clinicians, health care providers, policy makers and consumers in health care. Poorly designed studies and inadequate
reporting can lead to inappropriate application of qualitative
research in decision-making, health care, health policy and
future research.
Formal reporting guidelines have been developed for randomized controlled trials (CONSORT) [1], diagnostic test
studies (STARD), meta-analysis of RCTs (QUOROM) [2],
observational studies (STROBE) [3] and meta-analyses of
observational studies (MOOSE) [4]. These aim to improve
the quality of reporting these study types and allow readers to
better understand the design, conduct, analysis and ndings of
published studies. This process allows users of published
research to be more fuller informed when they critically
appraise studies relevant to each checklist and decide upon
applicability of research ndings to their local settings. Empiric
studies have shown that the use of the CONSORT statement
is associated with improvements in the quality of reports of

randomized controlled trials [5]. Systematic reviews of qualitative research almost always show that key aspects of study
design are not reported, and so there is a clear need for a
CONSORT-equivalent for qualitative research [6].
The Uniform Requirements for Manuscripts Submitted to
Biomedical Journals published by the International Committee
of Medical Journal Editors (ICMJE) do not provide reporting
guidelines for qualitative studies. Of all the mainstream biomedical journals (Fig. 1), only the British Medical Journal (BMJ)
has criteria for reviewing qualitative research. However, the
guidelines for authors specically record that the checklist is
not routinely used. In addition, the checklist is not comprehensive and does not provide specic guidance to assess some
of the criteria. Although checklists for critical appraisal are
available for qualitative research, there is no widely endorsed
reporting framework for any type of qualitative research [7].
We have developed a formal reporting checklist for
in-depth interviews and focus groups, the most common
methods for data collection in qualitative health research.

Address reprint requests to: Allison Tong, Centre for Kidney Research, The Childrens Hospital at Westmead, NSW 2145,
Australia. Tel: 61-2-9845-1482; Fax: 61-2-9845-1491; E-mail: allisont@health.usyd.edu.au, allisont@chw.edu.au
International Journal for Quality in Health Care vol. 19 no. 6
# The Author 2007. Published by Oxford University Press on behalf of International Society for Quality in Health Care; all rights reserved

349

Downloaded from intqhc.oxfordjournals.org at national cheng kung unversity on May 3, 2011

Methods. We performed a comprehensive search in Cochrane and Campbell Protocols, Medline, CINAHL, systematic reviews
of qualitative studies, author or reviewer guidelines of major medical journals and reference lists of relevant publications for
existing checklists used to assess qualitative studies. Seventy-six items from 22 checklists were compiled into a comprehensive
list. All items were grouped into three domains: (i) research team and reexivity, (ii) study design and (iii) data analysis and
reporting. Duplicate items and those that were ambiguous, too broadly dened and impractical to assess were removed.

A. Tong et al.

Downloaded from intqhc.oxfordjournals.org at national cheng kung unversity on May 3, 2011

Figure 1 Development of the COREQ Checklist. *References [26, 27], References [6, 28 32], Author and reviewer
guidelines provided by BMJ, JAMA, Lancet, Annals of Internal Medicine, NEJM.
These two methods are particularly useful for eliciting
patient and consumer priorities and needs to improve the
quality of health care [8]. The checklist aims to promote
complete and transparent reporting among researchers and
indirectly improve the rigor, comprehensiveness and credibility of interview and focus-group studies.

350

Basic definitions
Qualitative studies use non-quantitative methods to contribute new knowledge and to provide new perspectives in
health care. Although qualitative research encompasses a
broad range of study methods, most qualitative research

Consolidated criteria for reporting qualitative research

publications in health care describe the use of interviews and


focus groups [8].
Interviews
In-depth and semi-structured interviews explore the experiences of participants and the meanings they attribute to
them. Researchers encourage participants to talk about issues
pertinent to the research question by asking open-ended
questions, usually in one-to-one interviews. The interviewer
might re-word, re-order or clarify the questions to further
investigate topics introduced by the respondent. In qualitative
health research, in-depth interviews are often used to study
the experiences and meanings of disease, and to explore personal and sensitive themes. They can also help to identify
potentially modiable factors for improving health care [9].
Focus groups

Methods
Development of a checklist
Search strategy. We performed a comprehensive search for
published checklists used to assess or review qualitative
studies, and guidelines for reporting qualitative studies in:
Medline (1966Week 1 April 2006), CINAHL (1982
Week 3 April 2006), Cochrane and Campbell protocols,
systematic reviews of qualitative studies, author or reviewer
guidelines of major medical journals and reference lists of
relevant publications. We identied the terms used to index
the relevant articles already in our possession and performed
a broad search using those search terms. The electronic
databases were searched using terms and text words for
research (standards), health services research (standards) and
qualitative studies (evaluation). Duplicate checklists and
detailed instructions for conducting and analysing qualitative
studies were excluded.
Data extraction. From each of the included publications, we
extracted all criteria for assessing or reporting qualitative
studies. Seventy-six items from 22 checklists were compiled
into a comprehensive list. We recorded the frequency of each
item across all the publications. Items most frequently
included in the checklists related to sampling method, setting
for data collection, method of data collection, respondent

COREQ: content and rationale


(see Tables 1)
Domain 1: research team and reflexivity
(i) Personal characteristics: Qualitative researchers closely
engage with the research process and participants and are
therefore unable to completely avoid personal bias. Instead
researchers should recognize and clarify for readers their
identity, credentials, occupation, gender, experience and training. Subsequently this improves the credibility of the ndings
by giving readers the ability to assess how these factors
might have inuenced the researchers observations and
interpretations [13 15].
(ii) Relationship with participants: The relationship and
extent of interaction between the researcher and their participants should be described as it can have an effect on the
participants responses and also on the researchers understanding of the phenomena [16]. For example, a clinician
researcher may have a deep understanding of patients issues
but their involvement in patient care may inhibit frank discussion with patient participants when patients believe that
their responses will affect their treatment. For transparency,
the investigator should identify and state their assumptions
and personal interests in the research topic.
Domain 2: study design
(i) Theoretical framework: Researchers should clarify the
theoretical frameworks underpinning their study so readers
can understand how the researchers explored their research
questions and aims. Theoretical frameworks in qualitative
research include: grounded theory, to build theories from the
data; ethnography, to understand the culture of groups with
shared characteristics; phenomenology, to describe the
meaning and signicance of experiences; discourse analysis,
to analyse linguistic expression; and content analysis, to systematically organize data into a structured format [10].

351

Downloaded from intqhc.oxfordjournals.org at national cheng kung unversity on May 3, 2011

Focus groups are semi-structured discussions with groups of


4 12 people that aim to explore a specic set of issues [10].
Moderators often commence the focus group by asking
broad questions about the topic of interest, before asking the
focal questions. Although participants individually answer the
facilitators questions, they are encouraged to talk and interact
with each other [11]. This technique is built on the notion
that the group interaction encourages respondents to explore
and clarify individual and shared perspectives [12]. Focus
groups are used to explore views on health issues, programs,
interventions and research.

validation of ndings, method of recording data, description


of the derivation of themes and inclusion of supporting
quotations. We grouped all items into three domains: (i)
research team and reexivity, (ii) study design and (iii) data
analysis and reporting. (see Tables 2 4)
Within each domain we simplied all relevant items by
removing duplicates and those that were ambiguous, too
broadly dened, not specic to qualitative research, or
impractical to assess. Where necessary, the remaining items
were rephrased for clarity. Based upon consensus among the
authors, two new items that were considered relevant for
reporting qualitative research were added. The two new items
were identifying the authors who conducted the interview or
focus group and reporting the presence of non-participants
during the interview or focus group. The COREQ checklist
for explicit and comprehensive reporting of qualitative
studies consists of 32 criteria, with a descriptor to supplement each item (Table 1).

A. Tong et al.

Table 1 Consolidated criteria for reporting qualitative studies (COREQ): 32-item checklist
No Item

Guide questions/description

.............................................................................................................................................................................

(ii) Participant selection: Researchers should report how


participants were selected. Usually purposive sampling is
used which involves selecting participants who share particular characteristics and have the potential to provide rich, relevant and diverse data pertinent to the research question

352

[13, 17]. Convenience sampling is less optimal because it


may fail to capture important perspectives from difcultto-reach people [16]. Rigorous attempts to recruit participants
and reasons for non-participation should be stated to reduce
the likelihood of making unsupported statements [18].

Downloaded from intqhc.oxfordjournals.org at national cheng kung unversity on May 3, 2011

Domain 1: Research team and reexivity


Personal Characteristics
1. Interviewer/facilitator
Which author/s conducted the interview or focus group?
2. Credentials
What were the researchers credentials? E.g. PhD, MD
3. Occupation
What was their occupation at the time of the study?
4. Gender
Was the researcher male or female?
5. Experience and training
What experience or training did the researcher have?
Relationship with participants
6. Relationship established
Was a relationship established prior to study commencement?
7. Participant knowledge of the
What did the participants know about the researcher? e.g. personal goals, reasons for doing the
interviewer
research
8. Interviewer characteristics
What characteristics were reported about the interviewer/facilitator? e.g. Bias, assumptions,
reasons and interests in the research topic
Domain 2: study design
Theoretical framework
9. Methodological orientation and What methodological orientation was stated to underpin the study? e.g. grounded theory,
Theory
discourse analysis, ethnography, phenomenology, content analysis
Participant selection
10. Sampling
How were participants selected? e.g. purposive, convenience, consecutive, snowball
11. Method of approach
How were participants approached? e.g. face-to-face, telephone, mail, email
12. Sample size
How many participants were in the study?
13. Non-participation
How many people refused to participate or dropped out? Reasons?
Setting
14. Setting of data collection
Where was the data collected? e.g. home, clinic, workplace
15. Presence of non-participants
Was anyone else present besides the participants and researchers?
16. Description of sample
What are the important characteristics of the sample? e.g. demographic data, date
Data collection
17. Interview guide
Were questions, prompts, guides provided by the authors? Was it pilot tested?
18. Repeat interviews
Were repeat interviews carried out? If yes, how many?
19. Audio/visual recording
Did the research use audio or visual recording to collect the data?
20. Field notes
Were eld notes made during and/or after the interview or focus group?
21. Duration
What was the duration of the interviews or focus group?
22. Data saturation
Was data saturation discussed?
23. Transcripts returned
Were transcripts returned to participants for comment and/or correction?
Domain 3: analysis and ndingsz
Data analysis
24. Number of data coders
How many data coders coded the data?
25. Description of the coding tree Did authors provide a description of the coding tree?
26. Derivation of themes
Were themes identied in advance or derived from the data?
27. Software
What software, if applicable, was used to manage the data?
28. Participant checking
Did participants provide feedback on the ndings?
Reporting
29. Quotations presented
Were participant quotations presented to illustrate the themes / ndings? Was each
quotation identied? e.g. participant number
30. Data and ndings consistent
Was there consistency between the data presented and the ndings?
31. Clarity of major themes
Were major themes clearly presented in the ndings?
32. Clarity of minor themes
Is there a description of diverse cases or discussion of minor themes?

Table 2 Items included in 22 published checklists: Research team and reexivity domain
Item

References

............................................................................................................................................................................
a
a
b
b
b

[26]

[27]

[6]

[28]

[32]

[13] [15] [14] [17] [33] [34] [35] [16] [19] [36] [7] [37] [23] [38] [39] [22] BMJ

............................................................................................................................................................................................................................................

Other publications, bSystematic review of qualitative studies; BMJ, British Medical Journaleditors checklist for appraising qualitative research); , item included in the checklist.

353

Consolidated criteria for reporting qualitative research

Research team and reexivity


Nature of relationship between the
researcher and participants
Examination of role, bias, inuence
Description of role
Identity of the interviewer
Continued and prolonged engagement
Response to events
Prior assumptions and experience
Professional status
Journal, record of personal experience
Effects of research on researcher
Qualications
Training of the interviewer/facilitator
Expertise demonstrated
Perception of research at inception
Age
Gender
Social class
Reasons for conducting study
Sufcient contact
Too close to participants
Empathy
Distance between researcher and participants
Background
Familiarity with setting

Downloaded from intqhc.oxfordjournals.org at national cheng kung unversity on May 3, 2011

A. Tong et al.

354

Table 3 Items included in 22 published checklists: Study design


Item

References

.................................................................................................................................................................
a
a
b
b
b

[26] [27] [6] [28]

[32] [13] [15] [14] [17] [33] [34] [35] [16] [19] [36] [7] [37] [23] [38] [39] [22] BMJ

............................................................................................................................................................................................................................................

Study design
Methodological orientation, ontological or
epistemological basis
Samplingconvenience, purposive
Setting
Characteristics and description of sample
Reasons for participant selection
Non-participation
Inclusion and exclusion, criteria
Identity of the person responsible for recruitment
Sample size
Method of approach
Description of explanation of research to participants
Level and type of participation
Method of data collection, e.g. focus group,
in-depth interview
Audio and visual recording
Transcripts
Setting and location
Saturation of data
Use of a topic guide, tools, questions
Field notes
Changes and modications
Duration of interview, focus group
Sensitive to participant language and views
Number of interviews, focus groups
Time span
Time and resources available to the study

Other publications, bSystematic review of qualitative studies; BMJ, British Medical Journaleditors checklist for appraising qualitative research; , item included in the checklist.

Downloaded from intqhc.oxfordjournals.org at national cheng kung unversity on May 3, 2011

Table 4 Items included in 22 published checklists: Analysis and reporting


Item

References

.......................................................................................................................................................................
a
a
b
b
b

[26] [27] [6]

[28]

[32]

[13] [15] [14] [17] [33] [34] [35] [16] [19] [36] [7] [37] [23] [38] [39] [22] BMJ

............................................................................................................................................................................................................................................

Other publications, bSystematic review of qualitative studies; BMJ, British Medical Journaleditors checklist for appraising qualitative research, , item included in the checklist.

355

Consolidated criteria for reporting qualitative research

Respondent validation
Limitations and generalizability
Triangulation
Original data, quotation
Derivation of themes explicit
Contradictory, diverse, negative cases
Number of data analysts
In-depth description of analysis
Sufcient supporting data presented
Data, interpretation and conclusions
linked and integrated
Retain context of data
Explicit ndings, presented clearly
Outside checks
Software used
Discussion both for and against the
researchers arguments
Development of theories, explanations
Numerical data
Coding tree or coding system
Inter-observer reliability
Sufcient insight into meaning/perceptions
of participants
Reasons for selection of data to support ndings
New insight
Results interpreted in credible, innovative way
Eliminate other theories
Range of views
Distinguish between researcher and
participant voices
Proportion of data taken into account

Downloaded from intqhc.oxfordjournals.org at national cheng kung unversity on May 3, 2011

A. Tong et al.

Domain 3: analysis and findings


(i) Data analysis: Specifying the use of multiple coders or
other methods of researcher triangulation can indicate a
broader and more complex understanding of the phenomenon. The credibility of the ndings can be assessed if the
process of coding (selecting signicant sections from participant statements), and the derivation and identication of
themes are made explicit. Descriptions of coding and
memoing demonstrate how the researchers perceived, examined and developed their understanding of the data [17, 19].
Researchers sometimes use software packages to assist with
storage, searching and coding of qualitative data. In addition,
obtaining feedback from participants on the research ndings
adds validity to the researchers interpretations by ensuring
that the participants own meanings and perspectives are
represented and not curtailed by the researchers own agenda
and knowledge [23].
(ii) Reporting: If supporting quotations are provided,
researchers should include quotations from different

356

participants to add transparency and trustworthiness to their


ndings and interpretations of the data [17]. Readers should
be able to assess the consistency between the data presented
and the study ndings, including the both major and minor
themes. Summary ndings, interpretations and theories generated should be clearly presented in qualitative research
publications.

Discussion
The COREQ checklist was developed to promote explicit
and comprehensive reporting of qualitative studies (interviews and focus groups). The checklist consists of items
specic to reporting qualitative studies and precludes generic
criteria that are applicable to all types of research reports.
COREQ is a comprehensive checklist that covers necessary
components of study design, which should be reported. The
criteria included in the checklist can help researchers to
report important aspects of the research team, study
methods, context of the study, ndings, analysis and
interpretations.
At present, we acknowledge there is no empiric basis that
shows that the introduction of COREQ will improve the
quality of reporting of qualitative research. However this is
no different than when CONSORT, QUOROM and other
reporting checklists were introduced. Subsequent research
has shown that these checklists have improved the quality of
reporting of study types relevant to each checklist [5, 25],
and we believe that the effect of COREQ is likely to be
similar. Despite differences in the objectives and methods of
quantitative and qualitative methods, the underlying aim of
transparency in research methods and, at the least, the theoretical possibility of the reader being able to duplicate the
study methods should be the aims of both methodological
approaches. There is a perception among research funding
agencies, clinicians and policy makers, that qualitative
research is second class research. Initiatives like COREQ
are designed to encourage improvement in the quality of
reporting of qualitative studies, which will indirectly lead to
improved conduct, and greater recognition of qualitative
research as inherently equal scientic endeavor compared
with quantitative research that is used to assess the quality
and safety of health care. We invite readers to comment on
COREQ to improve the checklist.

References
1. Moher D, Schulz KF, Altman D. The CONSORT statement:
revised recommendations for improving the quality of reports
of parallel-group randomized trials. JAMA 2001;285:1987 91.
2. Moher D, Cook DJ, Eastwood S et al. Improving the quality of
reports of meta-analyses of randomised controlled trials: the
QUOROM statement. Quality of Reporting of Meta-analyses.
Lancet 1999;354:1896 900.

Downloaded from intqhc.oxfordjournals.org at national cheng kung unversity on May 3, 2011

Researchers should report the sample size of their study to


enable readers to assess the diversity of perspectives included.
(iii) Setting: Researchers should describe the context in
which the data were collected because it illuminates why participants responded in a particular way. For instance, participants might be more reserved and feel disempowered talking
in a hospital setting. The presence of non-participants during
interviews or focus groups should be reported as this can
also affect the opinions expressed by participants. For
example, parent interviewees might be reluctant to talk on
sensitive topics if their children are present. Participant
characteristics, such as basic demographic data, should be
reported so readers can consider the relevance of the ndings and interpretations to their own situation. This also
allows readers to assess whether perspectives from different
groups were explored and compared, such as patients and
health care providers [13, 19].
(iv) Data collection: The questions and prompts used in
data collection should be provided to enhance the readers
understanding of the researchers focus and to give readers the
ability to assess whether participants were encouraged to
openly convey their viewpoints. Researchers should also report
whether repeat interviews were conducted as this can inuence
the rapport developed between the researcher and participants
and affect the richness of data obtained. The method of
recording the participants words should be reported.
Generally, audio recording and transcription more accurately
reect the participants views than contemporaneous
researcher notes, more so if participants checked their own
transcript for accuracy [1921]. Reasons for not audio recording should be provided. In addition, eld notes maintain contextual details and non-verbal expressions for data analysis and
interpretation [19, 22]. Duration of the interview or focus
group should be reported as this affects the amount of data
obtained. Researchers should also clarify whether participants
were recruited until no new relevant knowledge was being
obtained from new participants (data saturation) [23, 24].

Consolidated criteria for reporting qualitative research

3. STROBE Statement: Strengthening the reporting of observational studies in epidemiology. http://www.strobe-statement.


org/Checkliste.html

23. Popay J, Rogers A, Williams G. Rationale and standards for the


systematic review of qualitative literature in health services
research. Qual Health Res 1998;8:341 51.

4. Stroup DF, Berlin JA, Morton SC et al. Meta-analysis of observational studies in epidemiology: a proposal for reporting.
Meta-analysis Of Observational Studies in Epidemiology
(MOOSE) group. JAMA 2000;283:2008 12.

24. Blumer H. Critiques of Research, in the Social Sciences.


New Brunswick, NJ: Transaction Books, 1979.

5. Moher D, Jones A, Lepage L. Use of the CONSORT


Statement and quality of reports of randomized trials. A comparative before-and-after evaluation. JAMA 2001;285:1992 5.
6. Mills E, Jadad AR, Ross C et al. Systematic review of qualitative
studies exploring parental beliefs and attitudes toward childhood vaccination identied common barriers to vaccination.
J Clin Epidemiol 2005;58:1081 8.
7. Kna KA, Howard NJ. Interpreting and reporting qualitative
research. Res Nurs Health 1984;7:7 14.
8. Sofaer S. Qualitative research methods. Int J Qual Health Care
2002;14:329 36.

10. Liamputtong P, Ezzy D. Qualitative Research Methods. 2nd edn.


Melbourne, Victoria: Oxford University Press, 2005.
11. Krueger RA, Casey MA. Focus Groups. A Practical Guide for
Applied Research. Thousand Oaks CA: Sage Publications, 2000.
12. Morgan DL. Focus Groups as Qualitative Research. Newbury Park,
California: Sage, 1988.
13. Giacomini MK, Cook DJ. Users guides to the medical literature XXIII. Qualitative research in health care. A. Are the
results of the study valid? JAMA 2000;284:357 62.
14. Malterud K. Qualitative research:standards challenges guidelines. Lancet 2001;358:483 8.
15. Mays N, Pope C. Qualitative research in health care: assessing
quality in qualitative research. BMJ 2000;320:502.
16. Elder NC, William L. Reading and evaluating qualitative
research studies. J Fam Pract 1995;41:279 85.
17. Cote L, Turgeon J. Appraising qualitative research articles in
medicine and medical education. Med Teach 2005;27:715.
18. Altheide D, Johnson J. Criteria for assessing interpretive validity
in qualitative research. In Denzin N, Lincoln Y (eds). Handbook
of Qualitative Research. Thousand Oaks CA: Sage Publications,
1994.
19. Fossey E, Harvey C, McDermott F et al. Understanding and evaluating qualitative research. Aust N Z J Psychiatry 2002;36:71732.
20. Seale C, Silverman S. Ensuring rigour in qualitative research.
Eur J Public Health 1997;7:379 84.
21. Scheff T. Single case analysis in the health sciences. Eur J Public
Health 1995;5:724.
22. Bluff R. Evaluating qualitative research. Br J Midwifery
1997;5:232 5.

26. Critical Skills Appraisal Programme (CASP) 10 Questions to


help you make sense of qualitative research: Milton Keynes
Primary Care Trust, 2002.
27. Spencer L, Ritchie J, Lewis J et al. Quality in Qualitative Evaluation:
A Framework for Assessing Research Evidence. London: Cabinet
Ofce. Government Chief Social Researchers Ofce, 2003.
28. Campbell R, Pound P, Pope C et al. Evaluating meta-ethnography:
a synthesis of qualitative research on lay experience of diabetes
and diabetes care. Soc Sci Med 2003;56:67184.
29. Feder GS, Hutson M, Ramsay I et al. Women exposed to intimate partner violence: expectations and experiences when they
encounter health care professionals: a meta-analysis of qualitative studies. Arch Intern Med 2006;166:22 37.
30. Pound P, Britten N, Morgan M et al. Resisting medicines: a synthesis of qualitative studies of medicine taking. Soc Sci Med
2005;61:133 55.
31. Smith LK, Pope C, Botha JL. Patients help-seeking experiences
and delay in cancer presentation: a qualitative synthesis. Lancet
2005;366:825 31.
32. Walter FM, Emery J, Braithwaite D et al. Lay understanding of
familial risk of common chronic diseases: a systematic review
and synthesis of qualitative research. Ann Fam Med 2004;
2:583 94.
33. Inui TS, Frankel RM. Evaluating the quality of qualitative
research: a proposal pro-term. J Gen Intern Med 1991;6:485 6.
34. Boulton M, Fitzpatrick R, Swinburn C. Qualitative research in
health care: II A structured review and evaluation of studies.
J Eval Clin Pract 1996;2:171 9.
35. Dixon-Woods M, Shaw RL, Agarwal S et al. The problem of
appraising qualitative research. Qual Saf Health Care
2004;13:223 5.
36. Hoddinott P, Pill R. A review of recently published qualitative
research in general practice. More methodological questions
than answers? Fam Pract 1997;14:3139.
37. Kuzel AJ, Engel JD, Addison RB et al. Desirable features of
qualitative research. Fam Pract Res J 1994;14:369 78.
38. Treloar C, Champness S, Simpson PL et al. Critical appraisal
checklist for qualitative research studies. Indian J Pediatr
2000;67:347 51.
39. Cesario S, Morin K, Santa-Donato A. Evaluating the level of
evidence in qualitative research. J Obstet Gynecol Neonatal Nurs
2001;31:708 14.
Accepted for publication 7 July 2007

357

Downloaded from intqhc.oxfordjournals.org at national cheng kung unversity on May 3, 2011

9. Wright EB, Holcombe C, Salmon P. Doctors communication


of trust, care, and respect in breast cancer: qualitative study.
BMJ 2004;328:864 8.

25. Delaney A, Bagshaw SM, Ferland A et al. A systematic evaluation of the quality of meta-anlyses in the critical care literature.
Crit Care 2005;9:575 82.

99(2) t86981095

Consolidatedcriteriaforreportingqualitativeresearch(COREQ):a32itemchecklist
forinterviewsandfocusgroup
InternationaljournalforQualityinHealthCare,Vol.19,No.6,pp.349357,2007

(checklist)

:
()

:
CochraneandCampbellProtocolsMedlineCINAHL
72 22
:(1) (2) (3)
(reporting)

Table 1 COREQ 32-item checklist


No

Item

Guide Questions/description

Domain 1:

1. /facilitator
?
2.
PhDMD?
3.

4.
?
5.
?

6.
?
7. ?e.g.
8.
?

99(2) t86981095

Domain 2 :

9.

10.
11.
12.
13.
(setting)
14.
15.
16.

17.
18.
19. /
20.
21.
22.
23. Transcript returned

?..
?
?
?
??
?
?
?
??
??
?
??
?
?

Domain 3:

24.
?
25. Coding tree
Coding Tree?
26.
?
27.
()
28.
?

29.
/?
?
30. ?
31.
?
32.
?

Evaluative Criteria for Qualitative


Research in Health Care: Controversies
and Recommendations
Deborah J. Cohen, PhD
Benjamin F. Crabtree, PhD
Department of Family Medicine, Research
Division, University of Medicine and
Dentistry, Robert Wood Johnson Medical
School, Somerset, New Jersey

ABSTRACT
PURPOSE We wanted to review and synthesize published criteria for good qualitative research and develop a cogent set of evaluative criteria.
METHODS We identified published journal articles discussing criteria for rigorous research using standard search strategies then examined reference sections
of relevant journal articles to identify books and book chapters on this topic. A
cross-publication content analysis allowed us to identify criteria and understand
the beliefs that shape them.
RESULTS Seven criteria for good qualitative research emerged: (1) carrying out
ethical research; (2) importance of the research; (3) clarity and coherence of the
research report; (4) use of appropriate and rigorous methods; (5) importance of
reflexivity or attending to researcher bias; (6) importance of establishing validity
or credibility; and (7) importance of verification or reliability. General agreement
was observed across publications on the first 4 quality dimensions. On the last
3, important divergent perspectives were observed in how these criteria should
be applied to qualitative research, with differences based on the paradigm
embraced by the authors.
CONCLUSION Qualitative research is not a unified field. Most manuscript and

grant reviewers are not qualitative experts and are likely to embrace a generic
set of criteria rather than those relevant to the particular qualitative approach
proposed or reported. Reviewers and researchers need to be aware of this tendency and educate health care researchers about the criteria appropriate for
evaluating qualitative research from within the theoretical and methodological
framework from which it emerges.
Ann Fam Med 2008;6:331-339. DOI: 10.1370/afm.818.

INTRODUCTION

Conicts of interest: none reported

CORRESPONDING AUTHOR

Deborah Cohen, PhD


Research Division
Department of Family Medicine
University of Medicine and Dentistry
Robert Wood Johnson Medical School
1 Worlds Fair Dr
Somerset, NJ 08873
cohendj@umdnj.edu

ntil the 1960s, the scientic methodwhich involves hypothesis


testing through controlled experimentationwas the predominant approach to research in the natural, physical, and social
sciences. In the social sciences, proponents of qualitative research argued
that the scientic method was not an appropriate model for studying people (eg, Cicourel,1 Schutz,2,3 and Garnkel4), and such methods as observation and interviewing would lead to a better understanding of social
life in its naturally occurring, uncontrolled form. Biomedical and clinical
research, with deep historical roots in quantitative methods, particularly
observational epidemiology5 and clinical trials,6 was on the periphery of
this debate. It was not until the late 1960s and 1970s that anthropologists
and sociologists began introducing qualitative research methods into the
health care eld.4,7-14
Since that time, qualitative research methods have been increasingly
used in clinical and health care research. Today, both journals (eg, Qualita-

ANNALS O F FAMILY MEDICINE

WWW.A N N FA MME D.O R G

331

VO L. 6, N O. 4

J ULY/AUG UST 2008

C R I T E R I A F O R Q U A L I TAT I V E R E S E A R C H

tive Health Research) and books are dedicated to qualitative methods in health care,15-17 and a vast literature
describes basic approaches of qualitative research,18,19
as well as specic information on focus groups,20-23
qualitative content analysis,24 observation and ethnography,25-27 interviewing,28-32 studying stories33,34
and conversation,35-37 doing case study,38,39 and action
research.40,41 Publications describe strategies for sampling,42-45 analyzing, reporting,45-49 and combining qualitative and quantitative methods50; and a growing body
of health care research reports ndings from studies
using in-depth interviews,51-54 focus groups,55-57 observation,58-60 and a range of mixed-methods designs.61-63
As part of a project to evaluate health care
improvements, we identied a need to help health care
researchers, particularly those with limited experience in qualitative research, evaluate and understand
qualitative methodologies. Our goals were to review
and synthesize published criteria for good qualitative
research and develop a cogent set of evaluative criteria
that would be helpful to researchers, reviewers, editors,
and funding agencies. In what follows, we identify the
standards of good qualitative research articulated in
the health care literature and describe the lessons we
learned as part of this process.

METHODS
A series of database searches were conducted to
identify published journal articles, books, and book
chapters offering criteria for evaluating and identifying
rigorous qualitative research.
Data Collection and Management
With the assistance of a librarian, a search was conducted in December 2005 using the Institute for Science (ISI) Web of Science database, which indexes
a wide range of journals and publications from 1980
to the present. Supplemental Appendix 1, available
online-only at http://www.annfammed.org/cgi/
content/full/6/4/331/DC1, describes our search
strategy. This search yielded a preliminary database
of 4,499 publications. Citation information, abstracts,
and the number of times the article was cited by other
authors were exported to a Microsoft Excel le and an
Endnote database.
After manually reviewing the Excel database, we
found and removed a large number of irrelevant publications in the physical and environmental sciences
(eg, forestry, observational studies of crystals), and
further sorted the remaining publications to identify
publications in health care. Among this subset, we read
abstracts and further sorted publications into (1) publications about qualitative methods, and (2) original research
ANNALS O F FAMILY MEDICINE

using qualitative methods. For the purposes of this analysis, we reviewed in detail only publications in the rst
category. We read each publication in this group and
further subdivided the group into publications that (1)
articulated criteria for evaluating qualitative research, (2)
addressed techniques for doing a particular qualitative
method (eg, interviewing, focus groups), or (3) described
a qualitative research strategy (eg, sampling, analysis).
Subsequent analyses focused on the rst category;
however, among publications in the second category, a
number of articles addressed the issue of quality in, for
example, case study,39 interviewing,28 focus groups,22,64,65
discourse,66 and narrative67,68 research that we excluded
as outside the scope of our analysis.
Books and book chapters could not be searched
in the same way because a database cataloging these
materials did not exist. Additionally, few books on
qualitative methods are written specically for health
care researchers, so we would not be able to determine
whether a book was or was not contributing to the
discourse in this eld. To overcome these challenges,
we used a snowball technique, identifying and examining books and book chapters cited in the journal
articles retrieved. Through this process, a number of
additional relevant journal articles were identied as
frequently cited but published in nonhealth care or
nonindexed journals (eg, online journals). These articles were included in our analysis.
Analysis
We read journal articles and book chapters and prepared notes recording the evaluative criteria that
author(s) posited and the world view or belief system
in which criteria were embedded, if available. When
criteria were attributed to another work, this information was noted. Books were reviewed and analyzed
differently. We read an introductory chapter or two to
understand the authors beliefs about research and prepared summary notes. Because most books contained a
section discussing evaluative criteria, we identied and
read this section, and prepared notes in the manner
described above for journal articles and book chapters.
An early observation was that not all publications
offered explicit criteria. Publications offering explicit
evaluative criteria were treated as a group. Publications
by the same author were analyzed and determined to
be sufciently similar to cluster. We examined evaluative criteria across publications, listing similar criteria in thematic clusters (eg, importance of research,
conducting ethically sound research), identifying the
central principle or theme of the cluster, and reviewing
and rening clusters. Publications that discussed evaluative criteria for qualitative research but did not offer
explicit criteria were analyzed separately.

WWW.A N N FA MME D.O R G

332

VO L. 6, N O. 4

J ULY/AUG UST 2008

C R I T E R I A F O R Q U A L I TAT I V E R E S E A R C H

Preliminary ndings were synthesized into a Web site for the


Robert Wood Johnson Foundation
(http://www.qualres.org). This
Web site was reviewed by Mary
Dixon-Woods, PhD, a health care
researcher with extensive expertise in qualitative research, whose
feedback regarding the implications of endorsing or positing a
unied set of evaluative criteria
encouraged our reection and
inuenced this report.

Table 1. Common Paradigms in Health Care Research


Paradigm

Assumptions

Positivism

There is a real world of objects apart from people


Researchers can know this reality and use symbols to accurately describe, represent and explain this reality
Researchers can compare their claims against this objective reality. This allows
for prediction, control, and empirical verification
There are real-world objects apart from people

Realism

Researchers can only know reality from their perspective of it


We cannot separate ourselves from what we know; however, objectivity is an
ideal researchers strive for through careful sampling and specific techniques

Interpretivism

It is possible to evaluate the extent to which objectivity or truth is attained.


This can be evaluated by a community of scholars and those who are studied
Reality as we know it is constructed intersubjectively. Meaning and understanding are developed socially and experientially
We cannot separate ourselves from what we know. Who we are and how we
understand the world are linked

RESULTS

Researchers values are inherent in all phases of research. Truth is negotiated

We identied 29 journal artithrough dialogue


cles19,26,45,69-94 and 16 books or
Findings or knowledge claims are created as an investigation proceeds and
emerge through dialogue and negotiations of meanings among community
book chapters95-110 that offered
members (both scholars and the community at large)
explicit criteria for evaluating the
All interpretations are located in a particular context, setting, and moment
quality of qualitative research.
Supplemental Appendix 2, available online-only at http://www.annfammed.
important when it was pragmatically and theoretically
org/cgi/content/full/6/4/331/DC1, contains a table
useful and advanced the current knowledge base.* Clarlisting citation information and criteria posited in
ity and coherence of the research report were criteria
these works. An additional 29 publications were idenemphasizing that the report itself should be concise and
tied that did not offer explicit criteria but informed
provide a clear and adequate description of the research
discourse on this topic and our analysis.111-139
question, background and contextual material, study
Seven evaluative criteria were identied: (1) carrydesign (eg, study participants, how they were chosen,
ing out ethical research; (2) importance of the research; how data are collected and analyzed), and rationale for
(3) clarity and coherence of the research report; (4)
methodological choices. Description of the data should
use of appropriate and rigorous methods; (5) imporbe unexaggerated, and the relationship between data
tance of reexivity or attending to researcher bias; (6)
and interpretation should be understandable.
importance of establishing validity or credibility; and
(7) importance of verication or reliability. There was
Researcher Bias
general agreement observed across publications on the
The majority of publications discussed issues of
rst 4 quality dimensions; however, on the last 3 criresearcher bias, recognizing researchers preconcepteria, disagreement was observed in how the concepts
tions, motivations, and ways of seeing shape the qualiof researcher bias, validity, and reliability should be
tative research process. (It should be noted there is
applied to qualitative research. Differences in perspecample evidence to suggest researcher motivations and
tives were grounded in paradigm debates regarding the
preconceptions shape all research.)140 One perspective
nature of knowledge and reality, with some arguing
(interpretivist) viewed researcher subjectivity as somefrom an interpretivist perspective and others from a
thing used actively and creatively through the research
more pragmatic realist perspective. Three major paraprocess rather than as a problem of bias.72 A hallmark
digms and their implications are described in Table 1.
of good research was understanding and reporting relevant preconceptions through reexive processing (ie,
Fundamental Criteria
reective journal-keeping). A second perspective (realIt was widely agreed that qualitative research should
ist) viewed researcher bias as a problem affecting the
be ethical, be important, be clearly and coherently
trustworthiness, truthfulness, or validity of the account.
articulated, and use appropriate and rigorous methods.
In addition to understanding researchers motivations
Conducting ethically sound research involved carrying
and preconceptions, value and rigor were enhanced by
out research in a way that was respectful,69 humane,95
* References 26, 69, 70, 73, 77, 80, 94, 95, 98, 106.
and honest,77 and that embodied the values of empathy, References 19, 26, 69, 70, 73, 75, 77, 84, 85, 87, 95, 107.
collaboration, and service.77,84 Research was considered
References 19, 69, 70, 72, 73, 77, 80-82, 87, 94, 103, 105.
ANNALS O F FAMILY MEDICINE

WWW.A N N FA MME D.O R G

333

VO L. 6, N O. 4

J ULY/AUG UST 2008

C R I T E R I A F O R Q U A L I TAT I V E R E S E A R C H

controlling bias through techniques


to verify and conrm ndings, as discussed in more detail below.* Thus,
whereas all publications agreed that
researcher bias was an important consideration, the approach for managing
bias was quite different depending on
the paradigm grounding the work.

Table 2. Verification Techniques Used in Qualitative Research


Technique

Definition

Triangulation

Using multiple data sources in an investigation to produce


understanding
The process of exposing oneself to a disinterested peer in a
manner paralleling an analytical session and for the purpose of
exploring aspects of the inquiry that might otherwise remain
only implicit within the inquirers mind
Auditing involves having a researcher not involved in the research
process examine both the process and product of the research
study. The purpose is to evaluate the accuracy and evaluate
whether the findings, interpretations, and conclusions are supported by the data
Data, analytic categories, interpretations, and conclusions are
tested with members of those groups from whom the data
were originally obtained. This can be done both formally and
informally, as opportunities for member checks may arise during the normal course of observation and conversation

Peer review/
debriefing

External audits/
auditing

Validity
A number of publications framed the
concept of validity in the context
Member checking
of quantitative research, where it
typically refers to the best available
approximation to the truth or falsity
of propositions.142(p37) Internal validity
refers to truth about claims made regarding the relationship between 2 variables. External validity refers to
the extent to which we can generalize ndings. Across
publications, different ideas emerged.
Understanding the concept of validity requires
understanding beliefs about the nature of reality. One
may believe that there can be multiple ways of understanding social life and reality, even multiple realities.
This view of reality emerges from an interpretivist perspective. Hallmarks of high-quality qualitative research
include producing a rich, substantive account with
strong evidence for inferences and conclusions and
then reporting the lived experiences of those observed
and their perspectives on social reality, while recognizing that these could be multiple and complex and that
the researcher is intertwined in the portrayal of this
experience. The goal is understanding and providing
a meaningful account of the complex perspectives and
realities studied.
In contrast, research may be based on the belief
that there is one reality that can be observed, and this
reality is knowable through the process of research,
albeit sometimes imperfectly. This perspective is
typically associated with a positivist paradigm that
underlies quantitative research, but also with the realist
paradigm found in some qualitative research. Qualitative research based on this view tends to use alternative terms for validity (eg, adequacy, trustworthiness,
accuracy, credibility) and emphasizes striving for truth
through the qualitative research process, for example,
by having outside auditors or research participants validate ndings. An important dimension of good qualitative research, therefore, is plausibility and accuracy.
Verification or Reliability
Divergent perspectives were observed on the appropriateness of applying the concept of veriability or
reliability when evaluating qualitative research. As
ANNALS O F FAMILY MEDICINE

is validity, this concept is rooted in quantitative and


experimental methods and refers to the extent to which
measures and experimental treatments are standardized
and controlled to reduce error and decrease the chance
of obtaining differences.142 Two distinct approaches to
evaluating the reliability of qualitative research were
articulated. In the rst, verication was a process negotiated between researchers and readers, where researchers were responsible for reporting information (eg, data
excerpts, how the researcher dealt with tacit knowledge, information about the interpretive process) so
readers could discern for themselves the patterns identied and verify the data, its analysis and interpretation.
This interpretivist perspective contrasts with the second, realist, perspective. Rather than leaving the auditing and conrming role to the reader, steps to establish
dependability should be built into the research process
to repeat and afrm researchers observations. In some
cases, special techniques, such as member checking,
peer review, debrieng, and external audits to achieve
reliability, are recommended and posited as hallmarks
of quality in qualitative research.|| In Table 2 we provide
a brief description of these techniques.
Perspectives on the Value of Criteria
Health care researchers also discuss the usefulness of
evaluative criteria. We observed 3 perspectives on the
utility of having unied criteria for assessing qualitative
research.
One perspective recognized the importance of validity and reliability as criteria for evaluating qualitative
research.132,133 Morse et al make the case that without
validity and reliability, qualitative research risks being
* References 19, 45, 71, 74, 78, 79, 83, 87, 96, 101-106, 108, 141.
References 69, 72, 76, 77, 80-82, 89, 95, 96.
References 45, 70, 71, 73, 74, 78, 79, 83, 86, 87, 90, 91, 93, 96, 98, 100-108, 141.
References 69, 70, 72, 81, 82, 89, 95, 109, 110.
|| References 19, 45, 71, 73, 74, 76, 78, 80, 83, 84, 86, 87, 93, 96, 100-106, 108, 141.

WWW.A N N FA MME D.O R G

334

VO L. 6, N O. 4

J ULY/AUG UST 2008

C R I T E R I A F O R Q U A L I TAT I V E R E S E A R C H

seen as nonscientic and lacking rigor.88,125 Their argument is compelling and suggests reliability and validity
should not be evaluated at the end of the project, but
should be goals that shape the entire research process,
inuencing study design, data collection, and analysis
choices. A second approach is to view the criteria of
validity and reliability as inappropriate for qualitative
research, and argue for the development of alternative
criteria relevant for assessing qualitative research.*
This position is commonly based on the premise
that the theoretical and methodological beliefs informing quantitative research (from whence the criteria
of reliability and validity come) are not the same as
the methodological and theoretical beliefs informing
qualitative research and are, therefore, inappropriate.136
Cogent criteria for evaluating qualitative research
are needed. Without well-dened, agreed-upon, and
appropriate standards, qualitative research risks being
evaluated by quantitative standards, which can lead to
assimilation, preferences for qualitative research that
are most compatible with quantitative standards, and
rejection of more radical methods that do not conform to quantitative criteria.94 From this perspective
emerged a number of alternative criteria for evaluating
qualitative research.
Alternative criteria have been open to criticism.
We observed such criticism in publications challenging the recommendation that qualitative research using
such techniques as member checking, multiple coding,
external audits, and triangulation is more reliable, valid,
and of better quality.72,82,90,91,112,127,143 Authors challenging this recommendation show how techniques such
as member checking can be problematic. For example,
it does not make sense to ask study participants to
check or verify audio-recorded transcribed data. In
other situations, study participants asked to check
or verify data may not recall what they said or did.
Even when study participants recall their responses,
there are a number of factors that may account for
discrepancies between what participants recall and the
researchers data and preliminary ndings. For instance,
the purpose of data analysis is to organize individual
statements into themes that produce new, higher-order
insights. Individual contributions may not be recognizable to participants, and higher-order insights might
not make sense.82 Similar issues have been articulated
about the peer-review and auditing processes127,143 and
some uses of triangulation.130 Thus, alternative criteria
for evaluating qualitative research have been posited
and criticized on the grounds that such criteria (1) cannot be applied in a formulaic manner; (2) do not necessarily lead to higher-quality research, particularly if
* References 72, 81, 82, 85, 94, 114, 118, 129, 136.

ANNALS O F FAMILY MEDICINE

these techniques are poorly implemented; and (3) foster the false expectation among evaluators of research
that use of one or more of these techniques in a study
is a mark of higher quality.72,81,90,91,112,123,127
A third approach suggests the search for a cogent
set of evaluative criteria for qualitative research is
misguided. The eld of qualitative research is broad
and diverse, not lending itself to evaluation by one
set of criteria. Instead, researchers need to recognize
each study is unique in its theoretical positioning and
approach, and different evaluative criteria are needed.
To fully understand the scientic quality of qualitative research sometimes requires a deep understanding
of the theoretical foundation and the science of the
approach. Thus, evaluating the scientic rigor of qualitative research requires learning, understanding, and
using appropriate evaluative criteria.123,124,135,137

DISCUSSION
There are a number of limitations of this analysis to
be acknowledged. First, although we conducted a
comprehensive literature review, it is always possible
for publications to be missed, particularly with our
identication of books and book chapters, which relied
on a snowball technique. In addition, relying on publications and works cited within publications to understand the dialogue about rigor in qualitative methods is
imperfect. Although these discussions manifest in the
literature, they also arise at conferences, grant review
sessions, and hallway conversations. Ones views are
open to revision (cf, Lincolns103,144), and relationships
with editors and others shape our ideas and whom we
cite. In this analysis, we cannot begin to understand
these inuences.
Our perspectives affect this report. Both authors
received doctoral training in qualitative methods in
social science disciplines (sociology/communication and
anthropology) and have assimilated these values into
health care as reviewers, editors, and active participants
in qualitative health care studies. Our training shapes
our beliefs, so we feel most aligned with interpretivism. This grounding inuences how we see qualitative
research, as well as the perspectives and voices we examine in this analysis. We have been exposed to a wide
range of theoretical and methodological approaches for
doing qualitative research, which may make us more
inclined to notice the generic character of evaluative
criteria emerging in the health care community and take
note of the potential costs of this approach.
In addition, we use 3 common paradigmsinterpretivism, realism, and positivismin our analysis. It
is important to understand that paradigms and debates
about paradigms are political and used to argue for

WWW.A N N FA MME D.O R G

335

VO L. 6, N O. 4

J ULY/AUG UST 2008

C R I T E R I A F O R Q U A L I TAT I V E R E S E A R C H

credibility and resources in the research community.


In this process, underlying views about the nature of
knowledge and reality have been simplied, sometimes
even dichotomized (interpretivism vs positivism). We
recognize our use of these paradigms as an oversimplication and limitation of our work, but one that
is appropriate if only because these categories are so
widely used in the works we analyze.
Our analysis reveals some common ground has
been negotiated with regard to establishing criteria for
rigorous qualitative research. It is important to notice
that the criteria that have been widely acceptedcarrying out ethical research and important research, preparing a clear and coherent research report, and using
appropriate and rigorous methodsare applicable to
all research. Divergent perspectives were observed
in the eld with regard to 3 criteria: researcher bias,
validity, and verication or reliability. These criteria are more heavily inuenced by quantitative and
experimental approaches142 and, not surprisingly, have
met with resistance. To understand the implications
of these inuences, our analysis suggests the utility of
examining how these criteria are embedded in beliefs
about the nature of knowledge and reality.
Central to the interpretivist paradigm, which
historically grounds most qualitative traditions, is
the assumption that realities are multiple, uid, and
co-constructed, and knowledge is taken to be negotiated between the observer and participants. From this
framework emerge evaluative criteria valuing research
that illuminates subjective meanings and understands
and articulates multiple ways of seeing a phenomenon.
Rich substance and content, clear delineation of the
research process, evidence of immersion and selfreection, and demonstration of the researchers way of
knowing, particularly with regard to tacit knowledge,
are essential features of high-quality research.
In contrast, fundamental to a positivist paradigm, which historically grounds most quantitative
approaches, is the assumption that there is a single
objective reality and the presumption that this reality is knowable. The realist paradigm softens this
belief by suggesting knowledge of reality is always
imperfect. Within the realist framework the goal of
qualitative research is to strive for attaining truth, and
good research is credible, conrmable, dependable,
and transferable. Thus, rigorous qualitative research
requires more than prolonged engagement, persistent
observation, thick description, and negative case analysis, but it should use such techniques as triangulation,
external auditing, and member checking to promote
attainment of truth or validity through the process of
verifying ndings.
One reason for the centrality of the realist paraANNALS O F FAMILY MEDICINE

digm in health care research may be its ability to


assimilate the values, beliefs, and criteria for rigorous
research that emerge from the positivist paradigm. In
a community that values biomedical bench research,
sees the randomized controlled trial as a reference
standard, holds a belief in an objective reality, and
values research that is reliable, valid, and generalizable (typically positivist ideals), it is not surprising that
realist views with regard to qualitative research have
found favor. Unlike interpretivism, realism adopts a
philosophy of science not at odds with the commonly
held ideals of positivism. By maintaining a belief in
an objective reality and positing truth as an ideal
qualitative researchers should strive for, realists have
succeeded at positioning the qualitative research enterprise as one that can produce research which is valid,
reliable, and generalizable, and therefore, of value and
import equal to quantitative biomedical research.
Although qualitative research emerging from a
realist paradigm may have successfully assimilated
into the clinical research community (as it has in other
disciplines), it may be at a cost. Qualitative approaches
most compatible with traditional values of quantitative
research may be most likely to be accepted (published
and funded). More radical methods (eg, feminist standpoint research, critical postmodern research), which
can make innovative contributions to the eld, may
be marginalized because they do not t the evaluative
criteria that have emerged in the health care community.94,115 In addition, doing rigorous qualitative
research in the way realists prescribe involves using a
number of techniques that may foster the appearance
of validity and reliability, but can be problematic if
inappropriately applied.*
The search for a single set of criteria for good
qualitative research is grounded in the assumption
that qualitative research is a unied eld.124,135,137,145
Qualitative research is grounded in a range of theoretical frameworks and uses a variety of methodological approaches to guide data collection and analysis.
Because most manuscript and grant reviewers are not
qualitative experts, they are likely to embrace a generic
set of criteria. Reviewers and researchers need to be
aware of the 7 criteria for good qualitative research,
but also they need to be aware that applying the same
standards across all qualitative research is inappropriate. Helping reviewers understand how an unfamiliar
qualitative approach should be executed and standards
for evaluating quality are essential, because reviewers,
even qualitative experts, might not be well-versed in
the particular qualitative method being used or proposed. Panel organizers and editors need to recognize
* References 72, 81, 90, 91, 112, 123, 127, 145.

WWW.A N N FA MME D.O R G

336

VO L. 6, N O. 4

J ULY/AUG UST 2008

C R I T E R I A F O R Q U A L I TAT I V E R E S E A R C H

that a qualitative expert may have only a very narrow


range of expertise. Moreover, some researchers may be
so entrenched in the dogma of their own approach that
they are unable to value qualitative methods dissimilar
from their own. This type of ax grinding harms not
only the efforts of qualitative researchers, but the eld
more generally.
Future work needs to focus on educating health care
researchers about the criteria for evaluating qualitative
research from within the appropriate theoretical and
methodological framework. Although the ideas posited
here suggest there may be a connection between how
quality is dened and the kind of work published or
funded, this assumption is worthy of empirical examination. In addition, the eld needs to reect on the value
of qualitative health care research and consider whether
we have the space and models for adequately reporting
interpretive research in our medical journals.

15. Pope C, Mays N, eds. Qualitative Research in Health Care. 2nd ed.
London: BMJ Books; 2000.
16. Crabtree BF, Miller WL, eds. Doing Qualitative Research. Thousand
Oaks, CA: Sage Publications; 1992.
17. Crabtree BF, Miller WL, eds. Doing Qualitative Research. 2nd ed.
Thousand Oaks, CA: Sage Publication; 1999.
18. Holloway I. Basic Concepts for Qualitative Research. Oxford: Blackwell Science; 1997.
19. Malterud K. Qualitative research: standards, challenges, and
guidelines. Lancet. 2001;358(9280):483-488.
20. Asbury J. Overview of focus group research. Qual Health Res.
1995;5(4):414-420.
21. Freeman T. Best practice in focus group research: making sense
of different views. J Adv Nurs. 2006;56(5):491-497.
22. Kidd PS, Parshall MB. Getting the focus and the group: enhancing analytical rigor in focus group research. Qual Health Res.
2000;10(3):293-308.
23. Sim J. Collecting and analysing qualitative data: issues raised by
the focus group. J Adv Nurs. 1998;28(2):345-352.
24. Graneheim UH, Lundman B. Qualitative content analysis in nursing
research: concepts, procedures and measures to achieve trustworthiness. Nurse Educ Today. 2004;24(2):105-112.

To read or post commentaries in response to this article, see it


online at http://www.annfammed.org/cgi/content/full/6/4/331.

25. Kuzel AJ. Naturalistic inquiry: An appropriate model for family


medicine. Fam Med. 1986;18(6):369-374.

Key words: Qualitative research; methodology; health care research

26. Mays N, Pope C. Qualitative research: observational methods in


health care settings. BMJ. 1995;311(6998):182-184.
27. Savage J. Ethnography and health care. BMJ. 2000;321(7273):
1400-1402.

Submitted April 15, 2007; submitted revised November 7, 2007;


accepted November 14, 2007.

28. Britten N. Qualitative research: qualitative interviews in medical


research. BMJ. 1995;311(6999):251-253.

Funding support: Preparation of this report was supported by a grant


from the Robert Wood Johnson Foundation (#053512).

29. Bauman LJ, Greenberg Adair E. The use of ethnographic interviewing


to inform questionnaire construction. Health Educ Q. 1992;19(1):9-23.

Acknowledgments: We are indebted to Mary Dixon-Woods, PhD, for


her insightful comments on earlier versions of this work.

30. Griffiths P, Gossop M, Powis B, Strang J. Researching hidden poulations of drug users by privileged access interviewers: methodological and practical issues. Addiction. 1993;88(12):1617-1626.

References

31. Hoddinott P, Roisin P. Qualitative research interviewing by general


practitioners. A personal view of the opportunitites and pitfalls.
Fam Pract. 1997;14(4):307-312.

1. Cicourel A. Method and Measurement in Sociology. New York, NY:


Free Press; 1964.

32. Nunkoosing K. The problem with interviews. Qual Health Res.


2005;15(5):698-706.

2. Schutz A. Collected Papers. Vol 1. The Hague: Martinus Nijhoff; 1962.


3. Schutz A. Collected Papers. Vol 2. The Hague: Martinus Nijhoff; 1962.

33. Atkinson P. Narrative turn or blind alley? Qual Health Res.


1997;7(3):325-344.

4. Garfinkel H. Studies in Ethnomethodology. Cambridge, MA: Polity


Press; 1967.

34. Kleinman A. The Illness Narratives. New York, NY: Basic Books; 1998.
35. Mishler EG. The Discourse of Medicine: Dialectics of Medical Interviews. Norwood, NJ: Ablex; 1984.

5. Feinstein AR. Additional basic approaches in clinical research. Clin


Res. 1985;33(2 Pt 1):111-114.

36. Traynor M. Discourse analysis: theoretical and historical overview


and review of papers in the Journal of Advanced Nursing 19962004. J Adv Nurs. 2006;54(1):62-72.

6. Meinert CL. Clinical Trials: Design, Conduct and Analysis. New York,
NY: Oxford University Press; 1986.
7. Goffman E. Asylums. New York, NY: Doubleday Anchor; 1961.
8. Glaser B, Strauss A. Awareness of Dying. New York, NY: Aldine; 1965.

37. Heritage J, Maynard DW, eds. Communication in Medical Care.


Cambridge: Cambridge University Press; 2006.

9. Garfinkel H. Good Organizational Reasons for Bad Clinic Records.


Studies in Ethnomethodology. Cambridge, MA: Polity Press;
1967:186-207.

38. Crabtree B, Miller W. Researching practice settings: a case study


approach. In: Crabtree B, Miller W, eds. Doing Qualitative Research.
2nd ed. Thousand Oaks, CA: Sage Publications; 1999:293-312.

10. Mishler E. Discourse and meaning in medical interviews. Research


grant proposal submitted to the Medical Founcation, Inc.; 1979.

39. Yin RK. Enhancing the quality of case studies in health services
research. Health Serv Res. 1999;34(5 Pt 2):1209-1224.

11. Waitzkin H. Medicine, superstructure and micropolitics. : Soc Sci


Med [Med Psychol Med Sociol]. 1979;13A(6):601-609.

40. Hugentobler MK, Israel BA, Schurman SJ. An action research


approach to workplace health: Integrating methods. Health Educ Q.
1992;19(1):55-76.

12. Kleinman A. Culture, and illness: a question of models. Cult Med


Psychiatry. 1977;1(3):229-231.
13. Kleinman AM. Some issues for a comparative study of medical
healing. Int J Soc Psychiatry. 1973;19(3):159-163.

41. Meyer J. Using qualitative methods in health-related action


research. In: Pope C, Mays N, eds. Qualitative Research in Health
Care. 2nd ed. London: BMJ Books; 2000:59-74.

14. Kleinman AM. Toward a comparative study of medical systems: an


integrated approach to the study of the relationship of medicine
and culture. Sci Med Man. 1973;1(1):55-65.

42. Kuzel A. Sampling in qualitative inquiry. In: Crabtree B, Miller W,


eds. Doing Qualitative Research. 2nd ed. Thousand Oaks, CA: Sage
Publications; 1999:33-45.

ANNALS O F FAMILY MEDICINE

WWW.A N N FA MME D.O R G

337

VO L. 6, N O. 4

J ULY/AUG UST 2008

C R I T E R I A F O R Q U A L I TAT I V E R E S E A R C H

43. Morse J. Strategies for sampling. In: Morse J, ed. Qualitative Nursing Research: A Contemporary Dialogue. Newbury Park, CA: Sage
Publications; 1991:127-145.

67. Mishler E. Validation in inquiry-guided research: The role of exemplars in narrative studies. Harv Educ Rev. 1990;60(4):415-442.
68. Bailey PH. Assuring quality in narrative analysis. West J Nurs Res.
1996;18(2):186-194.

44. Sandelowski M. Sample size in qualitative research. Res Nurs


Health. 1995;18(2):179-183.

69. Angen MJ. Evaluating interpretive inquiry: reviewing the


validity debate and opening the dialogue. Qual Health Res.
2000;10(3):378-395.

45. Patton MQ. Enhancing the quality and credibility of qualitative


analysis. Health Serv Res. 1999;34(5):1189-1208.

70. Britten N, Jones R, Murphy E, Stacy R. Qualitative research


methods in general practice and primary care. Fam Pract.
1995;12(1):104-114.

46. Malterud K. The art and science of clinical knowledge: evidence


beyond measures and numbers. Lancet. 2001;358(9279):397-400.
47. Morse JM. Qualitative methods: The state of the art. Qual Health
Res. 1999;9(3):393-406.

71. Cutcliffe JR, McKenna HP. Establishing the credibility of qualitative


research findings: the plot thickens. J Adv Nurs. 1999;30(2):374-380.

48. Miller WL, Crabtree BF. Qualitative analysis: how to begin making
sense. Fam Pract Res J. 1994;14(3):289-297.

72. Eakin JM, Mykhalovskiy E. Reframing the evaluation of qualitative


health research: reflections on a review of appraisal guidelines in
the health sciences. J Eval Clin Pract. 2003;9(2):187-194.

49. Pope C, Ziebland S, Mays N. Qualitative research in health care.


Analysing qualitative data. BMJ. 2000;320(7227):114-116.

73. Elder NC, Miller WL. Reading and evaluating qualitative research
studies. J Fam Pract. 1995;41(3):279-285.

50. Morgan DL. Practical stategies for combining qualitative and


quantitative methods: applications to health research. Qual Health
Res. 1998;8(3):362-376.

74. Fitzpatrick R, Boulton M. Qualitative methods for assessing health


care. Qual Health Care. 1994;3(2):107-113.

51. Back AL, Starks H, Hsu C, Gordon JR, Bharucha A, Pearlman


RA. Clinician-patient interactions about requests for physicianassisted suicide: a patient and family view. Arch Intern Med.
2002;162(11):1257-1265.

75. Frankel RM, Devers K. Qualitative research: a consumers guide.


Educ Health (Abingdon). 2000;13(1):113-123.
76. Giacomini MK, Cook DJ. Users guides to the medical literature:
XXIII. Qualitative research in health care B. What are the results
and how do they help me care for my patients? Evidence-Based
Medicine Working Group. JAMA. 2000;284(4):478-482.

52. England M, Tripp-Reimer T. Imminent concerns of filial caregivers reporting recent experiences of crisis. Int J Aging Hum Dev.
2003;56(1):67-88.

77. Hall JM, Stevens PE. Rigor in feminist research. ANS Adv Nurs Sci.
1991;13(3):16-29.

53. Gubrium JF, Rittman MR, Williams C, Young ME, Boylstein CA.
Benchmarking as everyday functional assessment in stroke recovery. J Gerontol B Psychol Sci Soc Sci. 2003;58(4):S203-S211.

78. Hamberg K, Johansson E, Lindgren G, Westman G. Scientific


rigour in qualitative research--examples from a study of womens
health in family practice. Fam Pract. 1994;11(2):176-181.

54. Wetle T, Shield R, Teno J, Miller SC, Welch L. Family perspectives


on end-of-life care experiences in nursing homes. Gerontologist.
2005;45(5):642-650.

79. Hinds PS, Scandrett-Hibden S, McAulay LS. Further assessment of


a method to estimate reliability and validity of qualitative research
findings. J Adv Nurs. 1990;15(4):430-435.

55. Duncan MT, Morgan DL. Sharing the caring: family caregivers
views of their relationships with nursing home staff. Gerontologist.
1994;34(2):235-244.

80. Inui TS, Frankel RM. Evaluating the quality of qualitative research:
a proposal pro tem. J Gen Intern Med. 1991;6(5):485-486.

56. Irwin K, Bertrand J, Mibandumba N, et al. Knowledge, attitudes and


beliefs about HIV infection and AIDS among healthy factory workers
and their wives, Kinshasa, Zaire. Soc Sci Med. 1991;32(8):917-930.

81. Koch T. Establishing rigour in qualitative research: the decision


trail. 1993. [discussion 101-103]. J Adv Nurs. 2006;53(1):91-100.
82. Koch T, Harrington A. Reconceptualizing rigour: the case for
reflexivity. J Adv Nurs. 1998;28(4):882-890.

57. Bradley EH, McGraw SA, Curry L, et al. Expanding the Andersen
model: the role of psychosocial factors in long-term care use.
Health Serv Res. 2002;37(5):1221-1242.

83. Krefting L. Rigor in qualitative research: the assessment of trustworthiness. Am J Occup Ther. 1991;45(3):214-222.

58. Jervis LL. The pollution of incontinence and the dirty work of caregiving in a U.S. nursing home. Med Anthropol Q. 2001;15(1):84-99.

84. Kuzel AJ, Engel JD, Addison RB, Bogdewic SP. Desirable features of
qualitative research. Fam Pract Res J. 1994;14(4):369-378.

59. Robinson JD. Closing medical encounters: two physician practices


and their implications for the expression of patients unstated concerns. Soc Sci Med. 2001;53(5):639-656.

85. Engel JD, Kuzel AJ. On the idea of what constitutes good qualitative inquiry. Qual Health Res. 1992;2(4):504-510.
86. Maxwell J. Understanding validity in qualitative research. Harv
Educ Rev. 1992;62(3):279-300.

60. Ware NC, Lachicotte WS, Kirschner SR, Cortes DE, Good BJ. Clinician experiences of managed mental health care: a rereading of
the threat. Med Anthropol Q. 2000;14(1):3-27.

87. Mays N, Pope C. Qualitative research in health care. Assessing


quality in qualitative research. BMJ. 2000;320(7226):50-52.

61. Wittink MN, Barg FK, Gallo JJ. Unwritten rules of talking to doctors about depression: integrating qualitative and quantitative
methods. Ann Fam Med. 2006;4(4):302-309.

88. Morse J, Barrett M, Mayan M, Olson K, Spiers J. Verification strategies for establishing reliability and validity in qualitative research.
Intl J Qual Meth. 2002;1(2):1-19.

62. Rabago D, Barrett B, Marchand L, Maberry R, Mundt M. Qualitative aspects of nasal irrigation use by patients with chronic sinus
disease in a multimethod study. Ann Fam Med. 2006;4(4):295-301.

89. Popay J, Rogers A, Williams G. Rationale and standards for


the systematic review of qualitative literature in health services
research. Qual Health Res. 1998;8(3):341-351.

63. Ahles TA, Wasson JH, Seville JL, et al. A controlled trial of methods
for managing pain in primary care patients with or without cooccurring psychosocial problems. Ann Fam Med. 2006;4(4):341-350.

90. Sandelowski M. The problem of rigor in qualitative research. ANS


Adv Nurs Sci. 1986;8(3):27-37.

64. Reed J, Payton VR. Focus groups: issues of analysis and interpretation. J Adv Nurs. 1997;26(4):765-771.

91. Sandelowski M. Rigor or rigor mortis: the problem of rigor in


qualitative research revisited. ANS Adv Nurs Sci. 1993;16(2):1-8.

65. Webb C, Kevern J. Focus groups as a research method: a critique


of some aspects of their use in nursing research. J Adv Nurs.
2001;33(6):798-805.

92. Seale C, Silverman D. Ensuring rigour in qualitative research. Eur J


Public Health. 1997;7(4):379-384.
93. Shortell SM. The emergence of qualitative methods in health services research. Health Serv Res. 1999;34(5 Pt 2):1083-1090.

66. Waitzkin H. On studying the discourse of medical encounters. A


critique of quantitative and qualitative methods and a proposal for
reasonable compromise. Med Care. 1990;28(6):473-488.

ANNALS O F FAMILY MEDICINE

94. Yardley L. Dilemmas in qualitative health research. Psychol Health.


2000;15:215-228.

WWW.A N N FA MME D.O R G

338

VO L. 6, N O. 4

J ULY/AUG UST 2008

C R I T E R I A F O R Q U A L I TAT I V E R E S E A R C H

95. Altheide D, Johnson J. Criteria for assessing interpretive validity


in qualitative research. In: Denzin N, Lincoln Y, eds. Handbook
of Qualitative Research. Thousand Oaks, CA: Sage Publications;
1994:485-499.

119. Dixon-Woods M, Shaw RL, Agarwal S, Smith JA. The problem of appraising qualitative research. Qual Saf Health Care.
2004;13(3):223-225.

96. Creswell J. Qualitative Inquiry and Research Design: Choosing Among


Five Traditions. Thousand Oaks, CA: Sage Publications; 1988.

120. Dingwall R, Murphy E, Watson P, Greatbatch D, Parker S. Catching


goldfish: quality in qualitative research. J Health Serv Res Policy.
1998;3(3):167-172.

97. Frankel R. Standards of qualitative research. In: Crabtree B, Miller


WL, eds. Doing Qualitative Research. 2nd ed. Thousand Oaks, CA:
Sage Publications; 1999:341-.

121. Emden C, Sandelowski M. The good, the bad and the relative,
Part one: Conceptions of goodness in qualitative research. Int J
Nurs Pract. 1998;4(4):206-212.

98. Hammersley M. Reading Ethnographic Research. New York, NY:


Longman; 1990.

122. Emden C, Sandelowski M. The good, the bad and the relative,
Part Two: Goodness and the criterion problem in qualitative
research. Int J Nurs Pract. 1999;5(1):2-7.

99. Kirk J, Miller M. Reliability and Validity in Qualitative Research. Vol 1.


Newbury Park, CA: Sage Publications; 1986.
100. Kuzel A, Like R. Standards for trustworthiness in qualitative studies
in primary care. In: Norton P, Stewart M, Tudiver F, Bass M, Dunn
E, eds. Primary Care Research Traditional and Innovative Approaches.
Vol 1. Newbury Park, CA: Sage Publications; 1991:138-158.
101. Leninger M. Importance and use of ethnomethods: ethnography
and ethnonursing research. Recent Advances in Nursing.1987;17:
12-36.

123. Harding G, Gantley M. Qualitative methods: beyond the cookbook. Fam Pract. 1998;15(1):76-79.
124. Johnson M, Long T, White A. Arguments for British Pluralism in
qualitative health research. J Adv Nurs. 2001;33(2):243-249.
125. Morse JM. Myth 93: Reliability and validity are not relevant to
qualitative inquiry. Qual Health Res. 1999;9(6):717-719.
126. Morse JM. considering the peer in peer review. Qual Health Res.
2002;12(5):579-580.

102. Leninger M. Evaluation criteria and critque of qualitative research


studies. In: Morse JM, ed. Critical Issues in Qualitative Research Methods. Thousand Oaks, CA: Sage Publications; 1994:95-115.
103. Lincoln Y, Guba E. Naturalistic Inquiry. Newbury Park, CA: Sage
Publications; 1985.

127. Morse JM. The significance of standards. Qual Health Res.


2003;13(9):1187-1188.
128. Nolan M, Behi R. Validity: a concept at the heart of research. Br J
Nurs. 1995;4(9):530-533.
129. Nolan M, Behi R. Alternative approaches to establishing reliability
and validity. Br J Nurs. May 25-Jun 7 1995;4(10):587-590.

104. Marshall C, Rossman G. Designing Qualitative Research. 3rd ed.


Thousand Oaks, CA: Sage Publications; 1999.

130. Nolan M, Behi R. Triangulation: the best of all worlds? Br J Nurs.


Jul 27-Aug 9 1995;4(14):829-832.

105. Mays N, Pope C. Quality in qualitative health research. In: Mays


N, Pope C, eds. Qualitative Research in Health Care. London: BMJ
Books; 2000:89-102.

131. Peck E, Secker J. Quality criteria for qualitative research: Does context make a difference. Qual Health Res. 1999;9(4):552-558.

106. Miles M, Huberman A. Qualitative Data Analysis: An Expanded


Sourcebook. 2nd ed. Thousand Oaks, CA: Sage Publications; 1994.

132. Poses RM, Isen AM. Qualitative research in medicine and health
care: questions and controversy. J Gen Intern Med. 1998;13(1):32-38.

107. Munhall P. Nursing Research: A Qualitative Perspective. 3rd ed. Boston, MA: Jones and Bartlett Publishers; 2001.

133. Poses RM, Levitt NJ. Qualitative research in health care. Antirealism is an excuse for sloppy work. BMJ. 2000;320(7251):1729-1730.

108. Patton M. Qualitative Evaluation and Research Methods. 2nd ed.


Newbury Park, CA: Sage Publications; 1990.

134. Ratcliffe JW, Gonzalez-del-Valle A. Rigor in health-related


research: toward an expanded conceptualization. Int J Health Serv.
1988;18(3):361-392.

109. Seale C. The Quality of Qualitative Research. Thousand Oaks, CA:


Sage Publications; 1999.

135. Rolfe G. Validity, trustworthiness and rigour: quality and the idea
of qualitative research. J Adv Nurs. 2006;53(3):304-310.

110. Silverman D. Interpreting Qualitative Data Methods for Analyzing


Talk, Text and Interaction. 2nd ed. Thousand Oaks, CA: Sage Publications; 2001.

136. Sandelowski M, Barroso J. Reading qualitative studies. Intl J Qual


Meth. 2002;1(1):1-47.
137. Sparkes AC. Myth 94: Qualitative health researchers will agree
about validity. Qual Health Res. 2001;11(4):538-552.

111. Avis M. Valid arguments? A consideration of the concept of validity in establishing the credibility of research findings. J Adv Nurs.
1995;22(6):1203-1209.

138. Ward MM. Study design in qualitative research: a guide to assessing quality. J Gen Intern Med. 1993;8(2):107-109.

112. Barbour RS. The newfound credibility of qualitative research?


Tales of technical essentialism and co-option. Qual Health Res.
2003;13(7):1019-1027.

139. Whittemore R, Chase S, Mandle C. Validity in qualitative research.


Qual Health Res. 2001;11(4):522-537.

113. Barbour RS, Barbour M. Evaluating and synthesizing qualitative


research: the need to develop a distinctive approach. J Eval Clin
Pract. 2003;9(2):179-186.

140. Deetz S. Differences in approaches to organizational science:


Rethinking Burell and Morgan and their legacy. Organ Sci.
1996;7(2):191-207.

114. Beck CT. Qualitative research: the evaluation of its credibility, fittingness, and auditability. West J Nurs Res. 1993;15(2):263-266.

141. Leininger M. Ethnomethods: the philosophic and epistemic bases


to explicate transcultural nursing knowledge. J Transcult Nurs.
1990;1(2):40-51.

115. Boulton M, Fitzpatrick R, Swinburn C. Qualitative research in


health care: II. A structured review and evaluation of studies. J Eval
Clin Pract. 1996;2(3):171-179.

142. Cook T, Campbell D. Quasi-Experimentation: Design and Analysis Issues


for Field Settings. Boston, MA: Houghton Mifflin Company; 1979.

116. Bryman A. Quantity and Quality in Social Research. Vol 18. New
York, NY: Routledge; 1988.

143. Barbour RS. Checklists for improving rigour in qualitative research:


a case of the tail wagging the dog? BMJ. 2001;322(7294):1115-1117.

117. Chapple A, Rogers A. Explicit guidelines for qualitative research: a


step in the right direction, a defence of the soft option, or a form
of sociological imperialism? Fam Pract. 1998;15(6):556-561.

144. Lincoln Y. Emerging criteria for quality in qualitative and interpretive research [Keynote address]. Annual meeting, American Educational Research Association. San Francisco, CA; 1995.

118. Devers KJ. How will we know good qualitative research when we
see it? Beginning the dialogue in health services research. Health
Serv Res. 1999;34(5 Pt 2):1153-1188.

145. Dixon-Woods M, Shaw R, Agarwal S, Smith J. The problem


of appraising qualitative research. J Health Serv Res Policy.
2007;12(1):42-47.

ANNALS O F FAMILY MEDICINE

WWW.A N N FA MME D.O R G

339

VO L. 6, N O. 4

J ULY/AUG UST 2008

CobenDJ,CrabtreeBF,2008,Evaluativecriteriaforqualitativeresearchinhealth
carecontroversiesandrecommendations.Annalsoffamilymedicine6(4):331339.

1.

2.

3.

(1) ethical
(2) research
(3)
(4) methods
(5)

(6) validitycredibility

(7) verificationreliability

verification

tacitknowledge

THEORY INTO PRACTICE / Summer 2000

Getting Good Qualitative Data

John W. Creswell
Dana L. Miller

Downloaded By: [Canadian Research Knowledge Network] At: 11:25 15 February 2009

Determining Validity
in Qualitative Inquiry

inquiry is challenging on many levels. Multiple perspectives about it flood the pages of books
(e.g., Lincoln & Guba, 1985; Maxwell, 1996; Merriam, 1998; Schwandt, 1997) and articles and chapters (e.g., Altheide & Johnson, 1994; Lather, 1993;
Maxwell, 1992). In these texts, readers are treated to
a confusing array of terms for validity, including authenticity, goodness, verisimilitude, adequacy, trustworthiness, plausibility, validity, validation, and
credibility. Various authors have constructed diverse
typologies of validity (e.g., Maxwells five types,
1992; Lathers four frames, 1993; and Schwandts
four positions, 1997). It is little wonder that Donmoyer (1996), who wrote an editorial on validity
in the Educational Researcher, commented on the
diverse perspectives of validity by contrasting Miles
and Hubermans (1994) traditional conception of
validity with Lathers (1993) ironic validity (p.
21). Novice researchers, in particular, can become
increasingly perplexed in attempting to understand
the notion of validity in qualitative inquiry.
There is a general consensus, however, that
qualitative inquirers need to demonstrate that their
studies are credible. To this end, several authors identify common procedures for establishing validity in
RITING ABOUT VALIDITY IN QUALITATIVE

John W. Creswell is professor of educational psychology


at the University of Nebraska-Lincoln; Dana L. Miller
is assistant professor of research methods at Doane
College, Lincoln, Nebraska.

qualitative projects (e.g., Lincoln & Guba, 1985;


Maxwell, 1996; Merriam, 1998). Qualitative researchers routinely employ member checking, triangulation,
thick description, peer reviews, and external audits.
Researchers engage in one or more of these procedures and report results in their investigations.
As helpful as they are, these discussions about
validity procedures provide little guidance as to
why one procedure might be selected for use by
researchers over other procedures. In this article,
we suggest that the choice of validity procedures
is governed by two perspectives: the lens researchers choose to validate their studies and researchers
paradigm assumptions. We advance a two-dimensional framework that can help researchers identify appropriate validity procedures for their studies.
The use of this framework can provide a rationale for choice of a procedure beyond what the
setting and participants will bear and what colleagues and faculty advisers recommend. The
framework helps researchers select procedures
based on who assesses the credibility of a study and
their own philosophical positions toward qualitative
inquiry. We begin by discussing the two perspectives
of the framework and then identify nine validity procedures that fit the framework. We end by describing
how the lens and paradigm assumptions help guide
our choice of validity procedures.
In this discussion we define validity as how
accurately the account represents participants realities of the social phenomena and is credible to

THEORY INTO PRACTICE, Volume 39, Number 3, Summer 2000


124
Copyright
2000 College of Education, The Ohio State University
0040-5841/2000$1.50

Creswell and Miller


Determining Validity

Downloaded By: [Canadian Research Knowledge Network] At: 11:25 15 February 2009

them (Schwandt, 1997). Procedures for validity


include those strategies used by researchers to establish the credibility of their study. Throughout
this discussion, we make the assumption that validity refers not to the data but to the inferences
drawn from them (Hammersley & Atkinson, 1983).

the interpretations accurately represent them. A


third lens may be the credibility of an account by
individuals external to the study. Reviewers not
affiliated with the project may help establish validity as well as various readers for whom the account is written.

The Lens Used by the Researcher

Paradigm Assumptions

When we refer to the lens, we mean that the


inquirer uses a viewpoint for establishing validity
in a study. Qualitative inquirers bring to their studies a different lens toward validity than that brought
to traditional, quantitative studies.
In quantitative research, investigators are
most concerned about the specific inferences made
from test scores on psychometric instruments (i.e.,
the construct, criterion, and content validity of interpretations of scores) (AERA, APA, & NCME, 1982)
and the internal and external validity of experimental
and quasi-experimental designs (Campbell & Stanley, 1966). In contrast, qualitative researchers use a
lens not based on scores, instruments, or research
designs but a lens established using the views of people who conduct, participate in, or read and review a
study.
For example, one lens to determine the credibility of a study is the particular lens of the researcher. Researchers determine how long to remain
in the field, whether the data are saturated to establish good themes or categories, and how the
analysis of the data evolves into a persuasive narrative. Patton (1980) describes this process as one
where qualitative analysts return to their data over
and over again to see if the constructs, categories,
explanations, and interpretations make sense (p.
339). Altheide and Johnson (1994) refer to it as
validity-as-reflexive-accounting (p. 489) where researchers, the topic, and the sense-making process
interact.
Qualitative inquirers may use a second lens
to establish the validity of their account: the participants in the study. The qualitative paradigm
assumes that reality is socially constructed and it
is what participants perceive it to be. This lens
suggests the importance of checking how accurately
participants realities have been represented in the
final account. Those who employ this lens seek to
actively involve participants in assessing whether

The lens researchers usetheir own, study


participants, or individuals external to the project
is not the only perspective that governs the choice
of validity procedures. Researchers paradigm assumptions or worldviews (Guba & Lincoln, 1994)
also shape their selection of procedures. As suggested by Ratcliffe (1983),
Quite different notions of what constitutes validity
have enjoyed the status of dominant paradigm at different times, in different historical contexts, and under different prevailing modes of thought and
epistemology. (p. 158)

Three paradigm assumptions, labeled by Guba


and Lincoln (1994) as postpostivist, constructivist,
and critical influence researchers choice of validity procedures. These assumptions have been associated with different historical moments in the
evolution of qualitative inquiry (Denzin & Lincoln,
1994). A brief overview of these paradigm assumptions is advanced here.
The postpostivist researcher assumes that qualitative research consists of rigorous methods and systematic forms of inquiry. Identified by Denzin and
Lincoln as the modernist phase of qualitative inquiry (1994, p. 8), this philosophical perspective
emerged in social science research during the 1970s
and continues today. Individuals embracing the
postpostivist position both recognize and support
validity, look for quantitative equivalence of it,
and actively employ procedures for establishing
validity using specific protocols. Maxwell (1996),
in Qualitative Research Design: An Interactive
Approach, for example, exemplifies postpostivist
assumptions toward qualitative validity.
The constructivist or interpretive position
emerged during the period of 1970 to 1987 (Denzin
& Lincoln, 1994), and it is reflected in stances toward validity today. Constructivists believe in pluralistic, interpretive, open-ended, and contextualized
(e.g., sensitive to place and situation) perspectives

125

THEORY INTO PRACTICE / Summer 2000

Downloaded By: [Canadian Research Knowledge Network] At: 11:25 15 February 2009

Getting Good Qualitative Data

toward reality. The validity procedures reflected


in this thinking present criteria with labels distinct
from quantitative approaches, such as trustworthiness (i.e., credibility, transferability, dependability,
and confirmability), and authenticity (i.e., fairness,
enlarges personal constructions, leads to improved
understanding of constructions of others, stimulates
action, and empowers action). The classical work
by Lincoln and Guba, Naturalistic Inquiry (1985),
provides extensive discussions about these forms
of trustworthiness and authenticity.
A third paradigm assumption is the critical
perspective. This perspective emerged during the
1980s as the crisis in representation (Denzin &
Lincoln, 1994, p. 9). As a challenge and critique
of the modern state, the critical perspective holds
that researchers should uncover the hidden assumptions about how narrative accounts are constructed,
read, and interpreted. What governs our perspective about narratives is our historical situatedness
of inquiry, a situatedness based on social, political, cultural, economic, ethnic, and gender antecedents of the studied situations. The implication
for validity of this perspective is that validity is
called into question, its assumptions interrogated
and challenged, and the researchers need to be reflexive and disclose what they bring to a narrative.
Richardson (1994) uses the metaphor of a
crystal as an image for validity: Crystals are prisms
that reflect externalities and refract within themselves.
. . . What we see depends on our angle of repose (p.

522). To this end, researchers engage in validity procedures of self-disclosure and collaboration with
participants in a study. These procedures help to
minimize further the inequality that participants
often feel. For example, Carspeckens Critical Ethnography in Educational Research (1996) reports
validity procedures for tracking bias and interviews
with oneself as ways for researchers to be situated
in a study.

Validity Within Lens and Paradigms


As shown in Table 1, we use the lens and
paradigm assumptions to create a two-dimensional
framework for locating nine different types of validity procedures. The discussion now turns to these
nine procedures with a brief definition of each,
their location within a lens and paradigm perspective, and approaches for implementing each procedure. This list is not exhaustive but includes those
procedures commonly used and cited in qualitative
literature.
Triangulation
Triangulation is a validity procedure where
researchers search for convergence among multiple and different sources of information to form
themes or categories in a study. The term comes
from military navigation at sea where sailors triangulated among different distant points to determine
their ships bearing (Jick, 1979). Denzin (1978)
identified four types of triangulation: across data

Table 1
Validity Procedures Within Qualitative Lens and Paradigm Assumptions
Paradigm assumption/Lens

Postpositivist or
Systematic Paradigm

Constructivist
Paradigm

Critical Paradigm

Lens of the
Researcher

Triangulation

Disconfirming
evidence

Researcher
reflexivity

Lens of Study
Participants

Member checking

Prolonged engagement in the field

Collaboration

Lens of People External to the Study


(Reviewers,
Readers)

The audit trail

Thick, rich
description

Peer debriefing

126

Downloaded By: [Canadian Research Knowledge Network] At: 11:25 15 February 2009

Creswell and Miller


Determining Validity

sources (i.e., participants), theories, methods (i.e.,


interview, observations, documents), and among
different investigators.
As a validity procedure, triangulation is a step
taken by researchers employing only the researchers lens, and it is a systematic process of sorting
through the data to find common themes or categories by eliminating overlapping areas. A popular
practice is for qualitative inquirers to provide corroborating evidence collected through multiple
methods, such as observations, interviews, and documents to locate major and minor themes. The narrative account is valid because researchers go
through this process and rely on multiple forms of
evidence rather than a single incident or data point
in the study.

ing beliefs and biases early in the research process


to allow readers to understand their positions, and
then to bracket or suspend those researcher biases
as the study proceeds. This validity procedure uses
the lens of the researcher but is clearly positioned
within the critical paradigm where individuals reflect on the social, cultural, and historical forces
that shape their interpretation.
Researchers might use several options for incorporating this reflexivity into a narrative account.
They may create a separate section on the role of
the researcher, provide an epilogue, use interpretive commentary throughout the discussion of the
findings, or bracket themselves out by describing
personal experiences as used in phenomenological
methods (Moustakas, 1994).

Disconfirming evidence
A procedure closely related to triangulation
is the search by researchers for disconfirming or
negative evidence (Miles & Huberman, 1994). It
is the process where investigators first establish
the preliminary themes or categories in a study
and then search through the data for evidence that
is consistent with or disconfirms these themes. In
this process, researchers rely on their own lens,
and this represents a constructivist approach in that
it is less systematic than other procedures and relies on examining all of the multiple perspectives
on a theme or category.
In practice, the search for disconfirming evidence is a difficult process because researchers have
the proclivity to find confirming rather than disconfirming evidence. Further, the disconfirming
evidence should not outweigh the confirming evidence. As evidence for the validity of a narrative
account, however, this search for disconfirming
evidence provides further support of the accounts
credibility because reality, according to constructivists, is multiple and complex.

Member checking
With member checking, the validity procedure shifts from the researchers to participants in
the study. Lincoln and Guba (1985) describe member checks as the most crucial technique for establishing credibility (p. 314) in a study. It consists
of taking data and interpretations back to the participants in the study so that they can confirm the
credibility of the information and narrative account.
With the lens focused on participants, the researchers systematically check the data and the narrative
account.
Several procedures facilitate this process. A
popular strategy is to convene a focus group of
participants to review the findings. Alternatively,
researchers may have participants view the raw data
(e.g., transcriptions or observational field notes)
and comment on their accuracy. Throughout this
process, the researchers ask participants if the
themes or categories make sense, whether they are
developed with sufficient evidence, and whether
the overall account is realistic and accurate. In turn,
researchers incorporate participants comments into
the final narrative. In this way, the participants
add credibility to the qualitative study by having a
chance to react to both the data and the final narrative.

Researcher reflexivity
A third validity procedure is for researchers to
self-disclose their assumptions, beliefs, and biases.
This is the process whereby researchers report on
personal beliefs, values, and biases that may shape
their inquiry. It is particularly important for researchers to acknowledge and describe their enter-

Prolonged engagement in the field


Another validity procedure is for researchers
to stay at the research site for a prolonged period

127

THEORY INTO PRACTICE / Summer 2000

Downloaded By: [Canadian Research Knowledge Network] At: 11:25 15 February 2009

Getting Good Qualitative Data

of time. Fetterman (1989) contends that working


with people day in and day out for long periods of
time is what gives ethnographic research its validity and vitality (p. 46). During repeated observation, the researchers build trust with participants,
find gatekeepers to allow access to people and sites,
establish rapport so that participants are comfortable disclosing information, and reciprocate by giving back to people being studied. This lens is
focused on gaining a credible account by building
a tight and holistic case.
Being in the field over time solidifies evidence because researchers can check out the data
and their hunches and compare interview data with
observational data. It is not a process that is systematically established, but constructivists recognize that the longer they stay in the field, the more
the pluralistic perspectives will be heard from participants and the better the understanding of the
context of participant views. In practice, prolonged
engagement in the field has no set duration, but
ethnographers, for example, spend from 4 months
to a year at a site.
Collaboration
Credible data also come from close collaboration with participants throughout the process of
research. Collaboration means that the participants
are involved in the study as co-researchers or in
less formal arrangements. This validity lens is one
of building the participants view into the study. It
belongs to a critical paradigm perspective because
the intent of the process is to respect and support
participants in a study, not further marginalize
them.
In practice, collaboration may assume multiple forms. For example, participants may help form
the research questions, assist with data collection
and analysis, and be involved in writing the narrative account. Some qualitative researchers may
share the profits, such as book royalties or co-authorship publication rights. By actively involving
participants in their studies, qualitative inquirers
add further credibility to their narrative accounts.
The audit trail
Now the lens for establishing validity shifts
again. The credibility of a study is established by

128

turning to individuals external to the project, such


as auditorsformally brought into the studyor
readers who examine the narrative account and attest to its credibility. In establishing an audit trail,
researchers provide clear documentation of all research decisions and activities. They may provide
evidence of the audit trail throughout the account or
in the appendices. Researchers may also use an external auditor to review their study. The goal of a
formal audit is to examine both the process and product of the inquiry, and determine the trustworthiness of the findings.
Lincoln and Guba (1985) use the analogy of
a fiscal audit to describe this process. The audit is
often used in formal studies, such as in dissertations, particularly when committee members are
trained quantitatively and may be skeptical about
qualitative studies. Certain audiences appreciate the
rigor of the audit process, and the lens for establishing credibility becomes someone external to the
project. It is a systematic procedure in that the
reviewer writes an analysis after carefully studying the documentation provided by the researcher.
An audit trail is established by researchers
documenting the inquiry process through journaling and memoing, keeping a research log of all
activities, developing a data collection chronology, and recording data analysis procedures clearly.
The external auditor examines this documentation
with the following questions in mind: Are the findings grounded in the data? Are inferences logical?
Is the category structure appropriate? Can inquiry
decisions and methodological shifts be justified?
What is the degree of researcher bias? What strategies were used for increasing credibility? (Schwandt
& Halpern, 1988). Through this process of documenting a study and a review of the documentation by an external auditor, the narrative account
becomes credible.
Thick, rich description
Another procedure for establishing credibility
in a study is to describe the setting, the participants,
and the themes of a qualitative study in rich detail.
According to Denzin (1989), thick descriptions are
deep, dense, detailed accounts. . . . Thin descriptions,
by contrast, lack detail, and simply report facts (p.
83). The purpose of a thick description is that it

Downloaded By: [Canadian Research Knowledge Network] At: 11:25 15 February 2009

Creswell and Miller


Determining Validity

creates verisimilitude, statements that produce for


the readers the feeling that they have experienced,
or could experience, the events being described in
a study. Thus, credibility is established through
the lens of readers who read a narrative account
and are transported into a setting or situation.
To use this procedure for establishing credibility, researchers employ a constructivist perspective to contextualize the people or sites studied.
The process of writing using thick description is
to provide as much detail as possible. It may involve describing a small slice of interaction, experience, or action; locating individuals in specific
situations; bringing a relationship or an interaction
alive between two or more persons; or providing a
detailed rendering of how people feel (Denzin,
1989).
With this vivid detail, the researchers help
readers understand that the account is credible. Rich
description also enables readers to make decisions
about the applicability of the findings to other settings or similar contexts.
Peer debriefing
A peer review or debriefing is the review of
the data and research process by someone who is
familiar with the research or the phenomenon being explored. A peer reviewer provides support,
plays devils advocate, challenges the researchers
assumptions, pushes the researchers to the next step
methodologically, and asks hard questions about
methods and interpretations (Lincoln & Guba,
1985).
The lens for establishing credibility is someone external to the study, and a critical paradigm
is operating because of the close collaboration between the external reviewer and the qualitative researcher. This procedure is best used over time
during the process of an entire study. Peer debriefers can provide written feedback to researchers or
simply serve as a sounding board for ideas. By
seeking the assistance of peer debriefers, researchers add credibility to a study.

Positioning Ourselves
Our approach is to use several validity procedures in our studies. Certainly some strategies
are easier to use than others, particularly those in-

herent in the study design, such as triangulation of


methods, prolonged observations in the field, and
the use of thick, rich descriptions. In deciding to
use a formal audit or peer debriefer, researchers
should consider their audiences, the availability of
such individuals, and the expense of using them.
Member checking is always important as well as
keeping research logs to document the rigor of our
research processes. When faced with students or
faculty committees that seek rigor and a systematic review of procedures, the process of establishing a clear audit trail is most important.
As we review the nine validity procedures,
we acknowledge the importance of all three lenses
and that their emphasis in a study will vary depending on the project, the audience for whom we
are writing, and the people available to provide an
assessment of our project. Our primary lens, however, is always that of the participants in a study,
and we have become more reflexive in our studies,
acknowledging the inseparableness of the researcher and the process of inquiry.
As for our paradigm stances, we most closely align ourselves with the use of systematic procedures, employing rigorous standards and clearly
identified procedures (e.g., Creswell, 1998). However, we also resonate with the critical perspective
and engage in collaborative research practices that
are respectful of the individuals we study (e.g.,
Miller, Creswell, & Olander, 1998). What is most
important is that the credibility of the account be
conveyed in a qualitative study. We suggest that
the use of validity procedures requires thinking
beyond specific proceduresto acknowledge the
lens being employed in a study and the paradigm
assumptions of the researchers.

References
Altheide, D.L., & Johnson, J.M. (1994). Criteria for
assessing interpretive validity in qualitative research. In N.K. Denzin & Y.S. Lincoln (Eds.),
Handbook of qualitative research (pp. 485-499).
Thousand Oaks, CA: Sage.
American Educational Research Association, American
Psychological Association, & National Council on
Measurement in Education. (1982). Standards for educational and psychological testing. Washington,
DC: American Educational Research Association.
Campbell, D.T., & Stanley, J.C. (1966). Experimental
and quasi-experimental designs for research. In

129

THEORY INTO PRACTICE / Summer 2000

Downloaded By: [Canadian Research Knowledge Network] At: 11:25 15 February 2009

Getting Good Qualitative Data


N.L. Gage, Handbook of research on teaching (pp.
1-76). Chicago, IL: Rand-McNally.
Carspecken, P.F. (1996). Critical ethnography in educational research: A theoretical and practical
guide. New York: Routledge.
Creswell, J.W. (1998). Qualitative inquiry and research
design: Choosing among five traditions. Thousand
Oaks, CA: Sage.
Denzin, N.K. (1978). The research act: A theoretical
orientation to sociological methods (2 nd ed.). New
York: McGraw-Hill.
Denzin, N.K. (1989). Interpretive interactionism. Newbury Park, CA: Sage.
Denzin, N.K., & Lincoln, Y.S. (1994). Introduction: Entering the field of qualitative research. In N.K. Denzin & Y.S. Lincoln (Eds.), Handbook of qualitative
research (pp. 1-17). Thousand Oaks, CA: Sage.
Donmoyer, R. (1996). Educational research in an era
of paradigm proliferation: Whats a journal editor
to do? Educational Researcher, 25(2), 19-25.
Fetterman, D.M. (1989). Ethnography: Step by step
(Applied Social Research Methods Series, No. 17).
Newbury Park, CA: Sage.
Guba, E.G., & Lincoln, Y.S. (1994). Competing paradigms in qualitative research. In N.K. Denzin &
Y.S. Lincoln (Eds.), Handbook of qualitative research (pp. 105-117). Thousand Oaks, CA: Sage.
Hammersley, M., & Atkinson, P. (1983). Ethnography: Principles in practice (2nd ed.). New York:
Routledge.
Jick, T.D. (1979). Mixing qualitative and quantitative
methods: Triangulation in action. Administrative Science Quarterly, 24, 602-611.

130

Lather, P. (1993). Fertile obsession: Validity after poststructuralism. The Sociological Quarterly, 34, 673-693.
Lincoln, Y.S., & Guba, E.G. (1985). Naturalistic inquiry.
Newbury Park, CA: Sage.
Maxwell, J.A. (1992). Understanding and validity in qualitative research. Harvard Educational Review, 62,
279-300.
Maxwell, J.A. (1996). Qualitative research design: An
interactive approach. Thousand Oaks, CA: Sage.
Merriam, S.B. (1998). Qualitative research and case study
applications in education. San Francisco: Jossey-Bass.
Miles, M.B., & Huberman, A.M. (1994). Qualitative data
analysis: An expanded sourcebook (2nd ed.). Newbury Park, CA: Sage.
Miller, D.L., Creswell, J. W., & Olander, L.S. (1998).
Writing and retelling multiple ethnographic tales of
a soup kitchen for the homeless. Qualitative Inquiry,
4, 469-491.
Moustakas, C. (1994). Phenomenological research methods. Thousand Oaks, CA: Sage.
Patton, M.Q. (1980). Qualitative evaluation methods.
Newbury Park, CA: Sage.
Ratcliffe, J.W. (1983). Notions of validity in qualitative
research methodology. Knowledge: Creation, Diffusion, Utilization, 5(2), 147-167.
Richardson, L. (1994). Writing: A method of inquiry. In
N.K. Denzin & Y.S. Lincoln (Eds.), Handbook of qualitative research (pp. 516-529). Thousand Oaks, CA:
Sage.
Schwandt, T.A. (1997). Qualitative inquiry: A dictionary
of terms. Thousand Oaks, CA: Sage.
Schwandt, T.A., & Halpern, E.S. (1988). Linking auditing and metaevaluation: Enhancing quality in applied research. Newbury Park, CA: Sage.