4.

6 Triangulation in Qualitative Research

Uwe Flick

1 Triangulation as a validation strategy 178

2 Criticisms of triangulation 179
3 Forms of application 179
4 Systematic triangulation of perspectives 181
5 Practical problems of triangulation 181
6 Perspectives: triangulation between convergence and divergence 182

In social research the term ‘triangulation’ is has little in common with the way it is used in
used to refer to the observation of the research surveying. The debate about non-reactive mea-
issue from (at least) two different points. surement procedures (Webb et al. 1966) and the
This is most often realized by means of apply- ‘multi-trait multi-method matrix’ approach of
ing different methodological approaches. As a Campbell and Fiske (1959) constitute the starting
strategy for the validation (see 4.7) of the proce- point for the general methodological discussion
dures and results of empirical social research tri- of the concept. Greater attention within qualita-
angulation has been given special attention, tive research has been given – even in the present
particularly in the more recent publications day – to the suggestions of Denzin (1978), who
on qualitative methods (cf. Marotzki 1995a; initially understood triangulation as a validation
Schründer-Lenzen 1997). Triangulation is cur- strategy and distinguished the following four
rently also being used in the debate about the different forms.
relationship between qualitative and quantita-
tive research (Jick 1983; see 4.5). In this chapter, • Triangulation of data combines data drawn
however, we are primarily concerned with trian- from different sources and at different times,
gulation within qualitative research, which has in different places or from different people.
been the subject of serious discussion in recent • Investigator triangulation is characterized by
literature (e.g. Flick 1998c; Seale 1999a,b; the use of different observers or interviewers,
Steinke 1999). to balance out the subjective influences of
individuals.
• Triangulation of theories means ‘approaching
1 TRIANGULATION AS data with multiple perspectives and hypo-
A VALIDATION STRATEGY theses in mind … . Various theoretical points
of view could be placed side by side to assess
The idea of triangulation was imported from their utility and power’ (Denzin 1978: 297).
land surveying into the methodological litera- • Denzin’s central concept is methodological
ture of the social sciences – admittedly in a rather triangulation ‘within-method’ (for example,
metaphorical sense. Blaikie (1991) explains, for the use of different subscales within a ques-
example, that its original use in the social sciences tionnaire) and ‘between-method’.
 TRIANGULATION IN QUALITATIVE RESEARCH 179

The goal of this last strategy is described by 3 FORMS OF APPLICATION

Denzin as follows: ‘To summarize, methodo-
logical triangulation involves a complex process The four forms of triangulation suggested by
of playing each method off against the other so Denzin may be used – even bearing in mind the
as to maximize the validity of field efforts’ criticisms we have listed – as starting points for
(1978: 304). the realization of this strategy.

Triangulation of data
2 CRITICISMS OF TRIANGULATION
In addition to verbal data – interviews (see 5.2)
In a number of contexts there have been critical and group discussions (see 5.4) – visual data are
discussions of triangulation as a strategy for currently receiving considerable attention in
validation in the sense which we have outlined: qualitative research. Apart from the emphasis on
too little attention is paid to the fact that every (not only participant) observation (see 5.5),
different method constitutes the issue it seeks video-recordings and photos (Becker 1986a; see
to investigate in a specific way (e.g. Bloor 1997: 5.6) are being used with increasing frequency,
39). If this aspect is neglected, triangulation and also the analysis of cinema films (Denzin
is faced with the accusation of ‘extreme 1989c; see 5.7). As a result of this, new perspec-
eclecticism’ (Fielding and Fielding 1986: 33). tives in the triangulation of data are emerging:
Silvermann (1985: 21) feels that ‘This casts apart from their use in interviews (cf. Flick 2002,
great doubt on the argument that multiple chs 8–9; Fuhs 1997), visual data may be triangu-
research methods should be employed in a vari- lated with verbal data as an independent source
ety of settings in order to gain a “total” picture of information (Harper, in 5.6, gives an example
of some phenomenon … . Putting the picture of the linking of photos and interviews).
together is more problematic than such propo- Completely new types of data, such as electronic
nents of triangulation would imply. What goes data (see 5.8), are opening up further possibilities
on in one setting is not a simple corrective to of triangulation with traditional types of data.
what happens elsewhere – each must be under-
stood in its own terms.’ Fielding and Fielding
Investigator triangulation
(1986: 33) sum up their criticism of Denzin’s
ideas in the following terms: ‘We should
Current implementations may be found in the
combine theories and methods carefully and
proposals that interpretations of collected data
purposefully with the intention of adding
should only be carried out in groups, so as to
breadth or depth to our analysis but not for the
expand, correct or check the subjective views
purpose of pursuing “objective” truth.’ Blaikie
of interpreters. In the context of objective
(1991) complains that the combination of
hermeneutics (Oevermann et al. 1979; see 5.16),
different methods pays too little attention to
this has long been required. Different ideas
the respective theoretical backgrounds of the
about research workshops (either in the sense of
individual methods.
Strauss 1987 or as they are used in biographical
In his more recent work (e.g. Denzin 1989c:
research and objective hermeneutics, see 6.2)
246; Denzin and Lincoln 1994a: 5), Denzin has
are also indebted to this idea.
taken up these criticisms and now understands
triangulation as strategy leading to a deeper
understanding of the issue under investigation, Within-method triangulation
and thereby as a step on the road to greater
knowledge, and less towards validity and objec- This principle may be clarified using the exam-
tivity of interpretation. Triangulation is now ple of episodic interviews (Flick 1996, 2000b):
seen less as a validation strategy within qualita- there some research issue (for example, technical
tive research and more as a strategy for justify- change in everyday life) is explored by means of
ing and underpinning knowledge by gaining invitations to narrate, focusing on experiences
additional knowledge (Denzin and Lincoln in concrete situations. These are combined with
1994a: 5; cf. Flick 1992a,b). questions that focus more on definitions and
180 A COMPANION TO QUALITATIVE RESEARCH

general answers. In this, questions are asked, for investigators are part of the research situation,
example, about the concept of a computer, and non-reactive procedures (analysis of avail-
which the interview partner has developed over able materials such as documents, photos,
a long period of time (‘What do you associate diaries and the like, see 5.15), that is to say,
today with the term “computer”? What types of data that were not set up for the investigation.
equipment does it include?’). Before this the In this process, the boundaries of both methodo-
interview partner is asked to talk about the situ- logical approaches are transcended. Moreover
ation in which he or she was confronted with a the triangulation of different approaches makes
computer for the first time (‘Could you describe it possible to capture different aspects of the
for me the situation in which you first got an research issue – such as concrete examples of
idea of what a computer is?’ or ‘Under what cir- professional activity and knowledge of one’s
cumstances did you first come into contact with own modes of action and routines.
a computer? Could you tell me about that situa- In a study of trust in counselling relationships
tion?’) or situations in which the computer has (Flick 1989), subjective theories of consultants
a special influence today in everyday life. In this about confidence were collected in semi-
way, an attempt is made in such an interview standardized interviews and triangulated with
systematically to unite the methodological conversation analyses of consultation talks
approaches of the semi-structured interview and which the interviewees had had with their clients
the narrative, using their respective strengths. in their everyday professional life. While the first
On the one hand, this is intended to open up approach shed light on more general experiences
complementary perspectives on the research and ideas on the part of the consultants about
issue through the interviewees’ mode of experi- preconditions and essential prerequisites for the
ence: as for the particular process-perspective creation of confidence, the second approach
that becomes clear in (situational) narratives made it possible to show how these ideas could
(‘When I first encountered a computer … ’ ), the successfully be translated into concrete action, or
abstract description of a state (‘a computer for how and why this failed to happen.
me is … ’ ) works in a complementary way. On Methodological triangulation is of particular
the other hand, it is intended to clarify the dif- current interest in ethnography. In Lüders’s
ferent facets of the subjective approach to the opinion (1995: 321), ‘ethnography is turning
research issue. For example, a female French into a research strategy which embraces every
information technologist, at an abstract level of conceivable and ethically tenable option for
more general concepts, regularly talked of the collecting data’. Here the methodological
gender-specific obstacles that generally make it approaches necessary to realize such options
more difficult for women to handle computers are triangulated with each other, even when
or technology. In the particular situations that the term is not always mentioned explicitly. The
she recounted, on the other hand, what became end-result is less a reciprocal validation of the
clear was a consistent success story of over- discoveries made using the individual methods
coming difficult equipment and situations (cf. but an extension of the possibilities of discovery
Flick 1996). about the aspect of life under investigation.
Since different methods, such as observation or
interviewing, tend to be combined in a rather
Between-method triangulation ad hoc way in a situation of extended participation
(see 5.5), it is also possible to speak of implicit
Its is the combination of different methods, triangulation in ethnography (Flick 1998c).
however, that is most strongly associated with Explicit triangulation occurs when ethno-
the keyword triangulation, and in this different graphic methods of extended participation and
emphases are given: on the one hand this refers field observation are deliberately combined with
to the linking of qualitative and quantitative the use of (career-biographical or episodic) inter-
methods (cf. Engler 1997; Flick 2002, ch. 21; see views with individual actors at individually
4.5) in different research designs. On the other agreed times. For example, in an ongoing pro-
hand, Marotzki (1995b) proposes the combina- ject (cf. Gebauer and Flick 1998), regular ethno-
tion of reactive procedures (for example, narra- graphic observations in fields where such new
tive interviews, see 5.2, 5.11), in which the sports as inline-skating are practised are being
 TRIANGULATION IN QUALITATIVE RESEARCH 181

triangulated with episodic interviews conducted 4 SYSTEMATIC TRIANGULATION

separately with individual athletes. The first OF PERSPECTIVES
approach makes it possible to analyse the modes
of action and communication, whilst the The proposal of ‘systematic triangulation of
second clarifies the meaning of the sport and perspectives’ (Flick 1992a,b) leads in a similar
the ‘scene’ for the participants. direction. Here different research perspectives
If the concept of triangulation is taken seri- within qualitative research are combined with
ously, it is characteristic of all of these variants one another in a targeted way, to complement
that they see the procedures they combine as their strong points and to illustrate their respective
being of equal value and that they do not begin limitations. This approach can be related to the
by regarding one procedure as central and the four types of application discussed above, but
others as preliminary or illustrative. will be outlined here as an example of the inter-
relating of different methods, using the example
already cited where consultants’ subjective
Triangulation of theories theories of trust in relation to clients are recon-
structed with interviews and communicative
In combining different methods it must be validation (using the ideas of Scheele and
borne in mind that each of them was developed Groeben 1988 and Kvale 1995a; see 4.7), and tri-
against a different theoretical background. In angulated with conversation analyses and coun-
concrete situations of triangulation the partially selling conversations. Here a number of different
incompatible epistemological assumptions research perspectives are applied: the first
about the research issue, or about (qualitative) approach focuses on subjective views (of the
research in these different theoretical back- consultant), whereas the second approach targets
grounds, are carried over by the methods. descriptions of everyday routines.
This problem may be clarified with reference to In this way it was possible to realize two of the
one of the examples mentioned above. The research perspectives of qualitative research that
reconstruction of subjective theories proceeds were distinguished by Lüders and Reichertz
from an explicitly subject-oriented understanding (1986). Using a different set of terminology
of knowledge and action (summed up by the (Bergmann 1985), in the first approach a recon-
keyword of the reflexive subject, Groeben 1990). structive procedure is applied and, in the second
Conversation analysis, on the other hand, rests approach, combined with an interpretative
on a more situation-oriented view of action procedure (for examples see Flick 1992b). This
(summed up by the keyword of the conversational approach explicitly combines triangulation of
machine) that largely dictates to the individual methods and data with a triangulation of theo-
participant how he or she can or should react to retical perspectives.
particular utterances of their interlocutor (see
5.17). This becomes a problem if such differences
are not taken into account in the way the research 5 PRACTICAL PROBLEMS
issue is understood. As a solution, a number of OF TRIANGULATION
alternatives have been discussed: Blaikie (1991:
129), for instance, suggests only combining meth- Case triangulation
ods within a single research approach, and points
to the example of Cicourel (1975) who combined The most consistent variant is to apply the
different methods with one another (‘indefinite triangulated methods to the same cases: coun-
triangulation’) within an ethnomethodological seling conversations by the consultants who are
approach. As an alternative to this, Fielding and being interviewed are collected and analysed,
Fielding (1986) require that these theoretical per- and the persons being observed in a particular
spectives be included in the analysis of the data field are (all) interviewed. This procedure makes
obtained, of the convergence and divergences possible a case-related analysis of both types
which the methods produce. Finally, Denzin of data and also makes it possible to compare
(1989c) feels it is important to look at data from and interrelate, in the context of a single-case,
different theoretical angles, in order to uncover the different perspectives opened up by the
new facets of the theories in the data. methodological approaches. In addition, these
182 A COMPANION TO QUALITATIVE RESEARCH

Data sets

Method I Triangulation Method II

Single-case

Figure 4.6.1 Starting points for triangulation of methods

comparisons and interrelationships can also be both analyses and puts them in relation to each
undertaken at a higher level: systems that other. As a practical problem the question arises
emerge from a comparison of one type of data here of how comparability of the samples, where
(for example, sequential patterns in counselling the different methods have been applied, can be
conversations) can be set against patterns guaranteed. In addition it must be clarified
from the comparison of other types of data whether the different methods can be applied at
(emphases and blind spots that may be deter- the same time or whether, because of project
mined in respect of all subjective theories or planning and resources, the empirical steps have
specifically for particular professional groups). to be conducted sequentially – first the observa-
Sampling decisions (see 4.4) are only taken tional data are collected and analysed and then
once, because the same selection of cases is used the interviews are conducted and analysed. In
for both types of data. this case possible influences of the different
The disadvantages are, first, that the load for times on content should not be forgotten.
an individual participant in an investigation is
often unreasonably large: to be ready for an
interview and in addition to provide a coun- 6 PERSPECTIVES: TRIANGULATION
selling conversation is, if measured against the BETWEEN CONVERGENCE AND
normal requirement of taking part in a study, a DIVERGENCE
comparatively heavy burden. Secondly, the
danger of dropout rises markedly. Everyone who The aim of the triangulation of different
refuses to provide either an interview or a coun- approaches and perspectives at both levels (cf.
selling conversation is ‘lost’ to the entire inves- Figure 4.6.1) should be less a matter of obtaining
tigation that seeks to triangulate on the basis of convergence in the sense of confirmation of
the particular case. what has already been discovered. The triangu-
lation of methods and perspectives is particu-
Triangulation of data sets larly useful for theory-development, when it can
elucidate divergent perspectives, when – to take up
Finally, in observations on open spaces (such as the above example again – the action of the
sport ‘scenes’) there is a problem that so many consultant is different from what his or her sub-
people have to be observed that not all of them jective theory about confidence would lead us
can be interviewed. For that reason, case trian- to expect.
gulation is not possible, and so it should be Then we have a new perspective that requires
implemented at the level of data sets. theoretical explanations. From this kind of
The individual methods are initially applied understanding of triangulation we may make
independently of each other, which produces a connections to the idea of ‘theoretical sampling’
set of observations and a series of interviews. and the theoretical saturation of Glaser and
Both are analysed to assess what they have in Strauss (1967). In accordance with this, Glaser
common and where they differ. Triangulation and Strauss (1967: 68) maintain that ‘a theory
then relates in practical terms to the results of generated from just one kind of data never fits,
 TRIANGULATION IN QUALITATIVE RESEARCH 183

or works as well, as a theory generated from FURTHER READING

diverse slices of data on the same category’. In
the process of theoretical sampling (see 4.4),
further methods are also consistently used if the Denzin, N. K. (1978) The Research Act, 2nd edn.
level of knowledge can thereby be increased. If Chicago: Aldine. (3rd edn. Englewood Cliffs,
the inclusion of new data no longer delivers new NJ: Prentice Hall, 1989.)
knowledge then theoretical saturation has been
reached. Where the use of further methods can Flick, U. (1992) ‘Triangulation Revisited – Strategy
‘only’ confirm knowledge that we already have, of or Alternative to Validation of Qualitative
in the sense of validating it, then triangulation Data’, Journal for the Theory of Social
comes up against the border of theoretical satu- Behavior, 22: 175–197.
ration. Accordingly, triangulation should be
understood as a means of extending our knowl- Seale, C. (1999) ‘Quality in Qualitative Research’,
edge of the research issue. Qualitative Inquiry, 5: 465–478.
We therefore have three modes of application
for triangulation: as a validation strategy, as an
approach to the generalization of discoveries,
and as a route to additional knowledge.