1
Use Matters … and Matters of Use: Building Theory for Reflective Practice
Chris Huxham and Paul Hibbert
Advanced Institute for Management Research and
The University of Strathclyde Business School
The University of Strathclyde Business School Department of Management
199 Cathedral Street Glasgow, G4 0QU
T: +44 (0)141 553 6143/6113
F: +44 (0)141 552 8851 E: [email protected] / [email protected]
Acknowledgements
We would like to thank our colleagues: Siv Vangen, who has been centrally involved in the development of the theory of collaborative advantage, the themes-based approach to theory building and the interpretive clustering approach; Colin Eden, who has been central to the development of the principles of research oriented action research and one of the collaborative themes; and, Nic Beech, who has been central to the development of our understanding of the use of theory for reflective practice and some of the collaborative themes. We would like to thank the many managers and other practitioners with whom we have worked over the past 18 years. We would like to acknowledge the support for this research of the U.K. Economic and Social Research Council (ESRC) and the Engineering and Physical Sciences Research Council under ESRC grant numbers 000234450 and L130251031 and the ESRC/EPSRC Advanced Institute of Management Research grant number RES-331-25-0016.
2
Use Matters … and Matters of Use: Building Theory for Reflective Practice
Debate about validity and rigour in qualitative management research has ranged widely in
recent years (Gergen and Gergen 2000). Many validity concepts have been proposed and
terminology such as trustworthiness, transferability, credibility, dependability, plausibility,
authenticity of various sorts, interpretive validity and so on has entered the academic
discourse alongside the more traditional notions of internal, external and measurement
validity, and reliability (summarized, for example, in Lincoln and Guba (2000) and Miles and
Huberman (1994)). Such concepts highlight alternative criteria for considering the validity of
the link between raw data and theoretical or other output of the research. In this article,
however, we aim to extend the discussion of validity and rigour beyond this single link,
through holistic consideration of the wider research process. We present the discussion
through an analysis of a particular approach to organizational research, which is explicitly
concerned with developing theory to support reflective practice. This is the Interpretive
Clustering Approach (ICA) (Author 1 2003; Author 1 and Co-author 2000) to theory building
in a Research Oriented Action Research (RO-AR) (Eden and Huxham 2006) context. We
have related the specific issues that we raise to this approach, to provide an illustrative case;
the broad issues, however, relate to a wide range of organizational research methods.
A holistic notion of validity and rigour
Within the breadth of recognized approaches to organizational theory development there is
usually an implicit link between the research methods and the form in which the theoretical
output relating to the substantive content appears (see, for example, Langley (1999); Van
Maanan, Sørensen and Mitchell 2007). By substantive content, we mean the subject matter or
area of interest of the research issue under study. Under research method we include the
epistemological basis of the research, the approach to data collection and the process of
3
theory building. The form of the theory refers to the nature or style of the conceptualization of
the substantive content. Though the links are often only made implicitly in reports of
research, these are clearly inter-related: any research method is only amenable to certain
types of subject matter and will to some extent, dictate the style in which the research output
is framed. Substantive content, research method and form can thus be thought of as three
dimensions of theory development.
These three critically important dimensions have become embedded in the literature (e.g.
Bryman and Bell 2007; Marshall and Rossman 2006) and are clearly evident, if not always
explicitly discussed, in accounts of empirical work. A fourth, equally important, dimension,
however, generally remains hidden. This relates to the intended use of the theory (see Figure
1). Theory might, for example, be intended for use as a contribution to academic debate or
intended to inform an aspect of the practice of those who work within organizational
environments. Within each of these two broad areas, there are many possibilities. For
example, theory intended for academic use might be in the form of propositions for later
testing or presented as qualitative insights for appropriation in future rounds of theory
building. Theory intended to inform practice might be intended to develop practitioners’
contextual awareness or to prescribe precise courses of action to be taken in particular
circumstances.
---------------------------------------------------------
Insert figure 1 about here
---------------------------------------------------------
The fourth dimension thus refers to the envisaged manner of use of the research output by the
projected audience, whoever they may be. This goes beyond notions such as catalytic
validity, fittingness, ecological validity or the capacity for practical wisdom (Bryman 2001;
Lather 1991; Miles and Huberman 1995; Schwandt 1996) which argue the case for the
4
usability of research output in practice situations. The concern here is not only with whether
theory is usable and useful to the intended audience, but with how it is to be used. We see the
design of the manner of its use as integrally related to the other three dimensions, since the
output needs to be in a form, and with content – and thus produced through a research method
– that is consistent with that manner of use.
We suggest that rigorous organizational research requires that the four dimensions are treated
in a mutually consistent way, whatever the validity criteria used. In brief, this means that the
validity of each must be judged partly in the context of its appropriateness to the others and
that ontological presumptions need to relate to all four. We are not arguing that each of these
aspects needs to be fully understood at the start of a piece of research; rather that the
development of a holistic and mutually consistent understanding of all four should be a
central part of the research process.
In this article we consider these four dimensions of the research process from the perspective
of the hidden dimension, intended use. As will be seen in the next section, we will focus on a
particular type of use for theory, returning to the generality of the relevance of the four
dimensions in the conclusion.
Theory to support reflective practice
Although we have stressed that the intended use of theory may legitimately be to inform
academic debate, the focus of discussion here will be on theory that is intended to support
practice (AMJ 2007; Van de Ven and Johnson 2006) and which is thus broadly consistent
with the style of research sometimes labelled “mode 2” (Hodgkinson 2001). We are
particularly concerned with theory building that explicitly aims to address Pettigrew’s (1997)
“double hurdles” of academic rigour and relevance to managerial practice (Aram and
5
Salipante 2003) and thus explicitly challenges the polarization of the two concepts (Gulati
2007).
There is a wide range of types of theory that is intended to support practice including those
that focus on lifecycles, success factors, process steps, competencies and so on (e.g. Hibbert,
Huxham and Ring 2008; Kieser and Nicolai 2005; McGahan 2007; Stubbs 2000). The ICA
approach that we focus on here, however, takes a very particular view of the way in which its
theoretical output is intended to inform practice. Its perspective is based on an ontological
presumption that the organizational situations to which its theoretical output seeks to have
relevance are intrinsically complex and indeterminate - and therefore not amenable to precise
conceptual description. They are also characterized by tensions – trade-offs – between
alternative possible courses of action and so are also not amenable to highly prescriptive
advice. ICA’s stance on the use dimension therefore pitches the role of theory as giving
support for reflective practice (Schipper 1999; Schön 1987) by providing conceptual handles
to stimulate practitioners’ consideration of the situations facing them and the actions they
need to take (Author 1 and Co-author 2003).
From this perspective, the quality of research output therefore needs to be judged against its
ability to be effective in this role. Theory needs to provide concepts and frameworks that are
both sufficiently recognizable to seem relevant to users and sufficiently generic to be
translated by them for application to their own context. It must be framed in such a way as to
support the user in action-focussed reflection; this could be either formal reflection in, for
example, structured workshop processes, or informal post-hoc reflection. The framing needs
to portray organizational processes in a mode that mirrors the way users are likely to
experience them (Stubbs 2000). It also, however needs to provide insight beyond that which
6
users would naturally have themselves (Raelin 2001). Theory is not expected to provide
solutions, but it is nonetheless important that it does not mislead the user towards
inappropriate action (Ghoshal, 2005; Raelin 2002).
Building this article
In summary, therefore, this article explores issues relating to the holistic consideration of
validity and rigour implied by the four dimensions of theory development when the use
dimension is the support of reflective practice. Our purpose, in focussing on reflective
practice is at two levels. On the one hand, we see this exploration of theory building for
reflective practice as an example of the way in which viewing the research process from the
use dimension influences considerations of research practice and rigour. On the other hand,
we pose the tenets about reflective practice that we uncover as having value in their own
right, either as points of principle for researchers concerned with supporting reflective
practice or as points of comparison for those with other uses in mind.
We used our own research programme as the basis of our investigation and use examples
from it to exemplify the points we wish to make here. The details of the programme are not
central to the thrust of this article, but a minimal description is necessary to provide the
context for the examples. The 17-year programme is concerned with the process of managing
inter-organizational collaboration. The theory of collaborative advantage (Author 1 1993;
Author 1 and Co-author 2004, 2005) is based around a central tension between the opposed
concepts of collaborative advantage and collaborative inertia; that is, between the presumed
synergistic benefits of collaboration and the more usual outcome, that collaborations make
slow, difficult or even no progress (Parise and Casher 2003). Particular conceptions of the
four dimensions of rigour have been the cornerstones of the programme (see Figure 2). The
7
substantive content is the accumulating Theory of Collaborative Advantage. The method is
the ICA used within a RO-AR setting. The form of the theory is overlapping descriptive
conceptualizations highlighting issues, contradictions, tensions and dilemmas and the
intended use is as handles for reflective practice.
---------------------------------------------------------
Insert figure 2 about here
---------------------------------------------------------
To produce this article, we carefully reviewed three projects from the research programme.
Our “data” were taken from research notes, reviewers’ reports and comments associated with
six articles, together with source data from the articles themselves and the various versions of
their narratives. We use two of these projects here to illustrate the arguments; these were on
leadership in collaborative situations and learning in collaborative situations (Author 2 and
Author 1 2005; Author 1 and Author 2 2005; Author 1 and Co-author 2000; Co-author and
Author 1 2003). Our exploration here takes the form of an interpretation of the ICA as a
means of theory development. This is not intended to be a full description; readers who wish
to know more about the approach should read this in conjunction with Author 1 (2003) and
Author 1 and Co-author (2000). Our concern here is with the aspects of this approach that
demonstrate how the consideration of the intended use of theory (in this illustration, reflective
practice) shapes the development process and suggests responses to the related validity and
rigour questions. In making those connections we highlight a number of important aspects of
the approach, which have not been discussed elsewhere.
The remainder of this paper is structured in two main parts. The first explains the
methodological principles of RO-AR and ICA in sufficient detail to support discussion of
how the approach addresses issues of rigour, validity and the intended use of theory,
especially in relation to the process of moving from data in the field to generalized
8
conceptualizations. The second focuses on the aspects of that process that relate to writing
conceptualizations from a structured analysis of the data, and the reception of theory
constructed in this way, which are clearly important issues in relation to supporting the final
use of the theory.
Building Theory From and For Practice
Background: Research Oriented Action Research and the Interpretive Clustering
Approach
Research Oriented Action Research (Eden and Huxham 2006) is a research methodology that
resembles ethnography in the sense that it relies on “naturally occurring” data rather than, for
example, interview or focus group data (Galibert 2004; Golden-Biddle and Locke 1993). It
has parallels with case study research to the extent that the generalizable output is usually
derived from a small number of researched situations (Eisenhardt and Graebner 2007, Stake
1995). The feature that distinguishes it from these approaches is that the researcher, acting as,
for example, a consultant, facilitator or participant, actively seeks to intervene in the
researched situation (Ayas 2003). Research data, collected during the course of the
intervention, typically includes: notes or recordings of verbal communication and
observational data; email; and, organizational documents, including formal material
workshop process outputs.
The Interpretive Clustering Approach (ICA) is one way in which RO-AR data can be
developed into theory (Author 1 2003; Author 1 and Co-author 2000 (but note that the label
ICA, and the discussion of the procedures, is introduced for the first time in this article)).
Theory is developed through a process of successive generalization, the stages of which
9
provide a platform from which “empirical material facilitates theorization” by “provid(ing)
resources for both imagination and discipline” (Alvesson and Kärreman 2007: 1266). This
will be discussed in more detail below. The resulting “themes-based” theory – comprised, as
indicated earlier, of descriptive conceptualizations in which issues, contradictions, tensions
and dilemmas are particularly highlighted – does not involve the generation of synthetic
explanatory variables (Langley 1999). However it does aim to provide purchase on the
reasons underlying the experienced world. The intention is that the conceptualizations
support the understanding and negotiation of possibilities rather than the development of
highly prescriptive guides. The purpose is to provide conceptual handles to support the
consideration of similar organizational problems (Author 1 and Co-author 2003).
Application of the ICA for theory building
We argued earlier that theory to support reflective practice needs to be sufficiently
recognizable to seem relevant to users, sufficiently generic to be translated by them for
application to their own context, framed in a mode that reflects the way users are likely to
experience their worlds and, at the same time, provide them with novel insight. The
interpretive clustering approach performs three functions for theory development that relate
to these requirements. Firstly it is a means of building sufficient generalizability from single
instance action-derived data. Secondly, it helps to ensure the theory is faithful (Gergen and
Gergen 2000; Lincoln and Guba 2000) to the data and thus reflects the experienced world as
well as promoting recognizability. Thirdly, it is a device for aiding the process of working
creatively with raw data to find below-the-surface insights. This means that it is likely to be
able to direct attention to aspects of the experienced world that, though recognizable once
explicated, might otherwise not have been recognized.
10
The process of moving beyond individual items of data to generalized theory for reflective
practice requires a commitment to that intended use at every step. Most particularly, it is
highly discursive, requiring reflective conversation (Jacobs and Heracleous 2005; Stubbs
2000) – including debate and negotiation – on the part of the researchers. The theory
development process involves the development of generalizable interpretations from
individual elements of data that are then linked to develop three successively higher levels of
generalization, We have used the software, Decision Explorer (DE), which can store and
display data, generalizable interpretations and links between them as “maps”, to assist in the
process of moving from items of data to theoretical conceptualizations (Bryson et al 2004).
This software is helpful, but not essential; the process could be completed using paper-based
methods, but it would be cumbersome. The levels are effectively equivalent to stages in
theory-building (see Figure 3). We discuss each of these three stages in turn below.
---------------------------------------------
Insert figure 3 about here
---------------------------------------------
Beginning to generalize: from datum to interpretation
Data derived from action settings is likely to be unique. Its value is in its potential to capture
the detail of what happens in situations where people are acting “for real” (Argyris and Schön
1974) rather than relying on espoused accounts of what happened or what ought to happen
(Alvesson 2003). This is important both for the recognizability of the resulting output and for
capturing below the surface aspects of situations that may not be recalled or may be
suppressed in, for example, post hoc accounts. However, it cannot in itself generate
generalizable conclusions. This means that, as with all small sample approaches,
generalizable output is in the form of interpretative insight rather than analytical “truth”.
11
At the first level of generalization, therefore, the focus is on possible generic interpretations
of individual data elements. These are generated by asking the question: in relation to the
focus of our research, what phenomenon could this datum be an example of? For example, in
our learning in collaborative settings project, a manager delivering a community service
asked colleagues,
“how do we create an atmosphere in which is OK to ask questions?”
Our DE map shows us having worked with four quite different interpretations of this data
element as shown in Figure 4.
---------------------------------------------
Insert Figure 4 about here
---------------------------------------------
Often, detailed debate reveals many possible interpretations from a small item of data. This is
partly because of the alternative constructions of what happened that can be placed upon any
event (Boje, Luhmann and Baack 1999; Kilduff and Mehra 1997); the appropriate
characterization is always a matter of debate. As Weick (2007, p17) notes “If things seem
simple… you’re not paying attention”. More importantly, even if a single construction of
what the data item is “about” could be settled on, it may still have relevance to many different
reflective practice challenges. As the example in Figure 4 demonstrates, the possibilities can
be highly varied; these are not usually alternatives to be chosen between, but contribute
additively to the potential utility of aspects of the theoretical picture in a range of similar
situations.
Although multiple interpretations are normal for any piece of data, two considerations are
important to the derivation of appropriate ones. Firstly, context has a significant effect on the
meaning of both utterances (Bechky 2003) and events (Pettigrew 1985,1990), so data should
be interpreted in the light of the circumstances in which it has been collected. Thus, for
12
example, data collected during a workshop involving a range of collaborating organizations
would need to be interpreted in the light of the dynamics and politics of the collaboration.
Context cannot, however, ever be fully appreciated and important aspects of it are often
unknowable. Thus, from the leadership in collaboration project, the quotation,
“the leadership task is to enthuse and develop anyone who can deliver the partnership’s
aims (Wendy)” (Author 1 and Co-author 2000: p1164),
could be interpreted as a typical statement of naivety about what can be achieved for a
partnership or as highlighting an important task area, depending on how experienced and
capable Wendy is judged to be. Statements can thus be approached both appreciatively and
suspiciously, leading to multiple plausible interpretations (Kock 2004; Prasad 2002). In
supporting reflective practice, the resultant multiplicity of interpretation is unproblematic, but
care is needed to avoid misleading the user by implying that any particular interpretation is to
be preferred.
Secondly, and also related to the avoidance of misleading the user, it is also important to be
clear about whether data were provoked by the nature of the intervention, for example by
asking direct questions or focussing the participants’ attention directly on the issue being
researched. Provoked data can surface issues that would otherwise have been obscured by,
for example, lack of interest (c.f. Heracleous 2001) but must be interpreted with caution. Like
interview or focus group data, it may be an honest articulation, but this does not necessarily
mean that it transparently represents past practice or future intentions (Alvesson 2003).
Fleshing out: from interpretation back to data
In Figure 4, the four interpretation concepts did not all originate with the “creating an
atmosphere” datum (285). Those numbered 44 and 280 were derived from data considered
13
earlier in the process. Thus, while this datum might not have triggered interpretations around
trust and social ritual, it did seem relevant to these, given that we had already identified them
when considering other data. This leads to ICA’s second level of generalization, which is
concerned with building understanding of the interpretation concepts.
As more data items are considered, many interpretations become informed by several of
them, sometimes via lower level (that is, less general) interpretation concepts. Thus, for
example, as shown in Figure 5, the interpretation concept 44, “the need for trust in learning”,
is informed by data from four collaborative situations and has several “sub-interpretations”.
---------------------------------------------
Insert figure 5 about here
---------------------------------------------
Each time a data item is linked to an interpretation concept it adds weight to our belief in the
relevance of that concept to the phenomenon of interest. Importantly, however, the role of
additional data is not just to confirm the interpretation, but also to flesh out the nature of the
concept; the process of comparison adds richness (Weick, 2007). It is likely to introduce both
increased central coherence and a variety of possible actualisations to our understanding of
the concept, all of which may have potential utility in a variety of situations of reflective
practice. Thus, in the example, our data suggested that issues relating to “protecting yourself”
(121) and to judgements of “partners’ competence” (206/7) might, in addition to “creating a
learning atmosphere”, be reasons for considering the “role of trust in learning”. In this and
similar forms of theory development, therefore, confidence in the validity of the central
theory is underpinned at the same time as its intended use, in reflective practice, is also
further supported.
14
As mentioned earlier, reflective conversation – involving debate and negotiation – is
paramount in this process of moving from single data items to concepts (Author 1 2003).
Joint research is therefore the ideal pattern. For example, if co-researchers have been
involved in all or parts of an intervention that is contributing data to the analysis, there is
more opportunity for spotting relevant data, more potential for debating what was going on in
that situation and greater possibility of generating alternative interpretations of it. In the
analysis phase, the process of proposing to each other data items that might have significance
for the research topic, debating whether they do have relevance, considering what their
generic significance could be and how that links to the interpretations of other data – which is
the basis of level 1 and 2 generalizing – can all be encapsulated within a single process of
negotiation (Wallace, Ross and Davies 2003). Such debate is important both for broadening
the range of possibilities considered and for challenging each researcher’s assumptions. Notes
made during our analysis for the learning in collaboration project run:
“(There is) quite a lot of overlap in data items (we have each) chosen but many
differences in initial interpretation of these data items - but also no unresolved matters
after discussion”.
We also noted that we,
“… did not insert data when we did not have a good understanding / sense of a robust
interpretation”.
It is not impossible for a lone researcher to conduct most of the aspects of this research.
However, this would mean developing processes to challenge oneself during the stages of
analysis and finding situations in which the reflective conversation can be established with
others as the interpretive process is finalized.
15
Conceptualizing: from interpretation to narrative
The third level of generalizing is concerned with pulling together interpretations into clusters
which then form the basis of the final conceptualization. As shown in Figure 6, by the end of
our analysis we had included the “need for trust in learning” concept in clusters relating to
emotion and power in learning in collaboration and these in turn ultimately formed a small
part of our conceptualization of attitudes to learning in collaboration (Author 1 and Author 2
2005).
---------------------------------------------
Insert figure 6 about here
---------------------------------------------
At the third level of generalization the process moves from mapping to writing. Obviously a
key part of that process is deciding on the broad line of argument that holds the clusters
together. There are always many possibilities for debate and discussion. Experience shows
that the structure decided upon will be moulded and to some extent changed during the
writing process. Because the process of converting clusters into writing is so important to the
theory development process (Eden and Huxham 2006), we have devoted the next main
section to it.
Writing Theory From Clusters
Stage three is an extremely significant, and often the most difficult, part of the process. While
the move from data to interpretive cluster can be done systematically, turning clusters into
prose is a mysterious process. The content of the writing comes directly from the clusters;
what is difficult to trace, however, is the development of a suitable style of presentation,
consistent with the use dimension of rigour. There are at least three interrelated aspects to
16
this, relating to conveying creative meaning, generality and style of use, and interest – all of
which connect with and influence both the use of the text in reflective practice and the way it
conveys validity to academic audiences. We address these points in turn below.
Drawing creativity and meaning
To illustrate the point about meaning we take an example from the second of our
conceptualizations of learning, which introduced the notion of local collaborative process
learning (Author 2 and Author 1 2005). When we began to write we were working from
clusters that we had so far labelled as “learning to collaborate in a particular situation”, and
“about partners”. We had quite a lot of data – for example,
“there is a lot of uncertainty around him, connected to unknown details of his role and
the potential availability of money”
– which, indicated things people feel they need to know in order to progress a collaboration.
The challenge was to use the associated generic interpretations in a way of that would say
something that was not immediately obvious, emphasizing issues, contradictions, tensions
and dilemmas. In situations like this, getting to the point of establishing a compact, tangible
and useful concept often seems a painful process of struggling with the data. With hindsight
it seems that two principles may be important in coming out on top of the struggle. The first
is continually to force the writing to focus on the intended topic, resisting the temptation to
take a more obvious, but less useful line. Thus for example, a sentence originally written as,
“the process must involve (at least) making a combined judgement about …”
was altered to,
“the process must involve (at least) the learning required to make a combined judgement
about …” (Author 2 and Author 1 2005: p 63. Current emphasis),
thus shifting the focus from the judgement itself to the learning task – our focal interest – that
17
would be required to make it. The second involves crafting second order interpretation
concepts from combinations of the original ones as the prose is worked on. Thus, for
example, in the context of the same piece of research, the notion of,
“… grappl(ing) simultaneously with a number of interrelated learning tasks …”
(Author 2 and Author 1 2005: p 63)
seems relatively straightforward when the individual tasks are considered together but is not
anywhere directly evident in the data. However, whereas the process of developing first order
interpretations can be done relatively routinely by following the rules, second order ones are
often derived through a trial and error process of “playing” with different possible ways of
writing about the concepts in combination. If this reveals something quite new, then it
inevitably challenges the researcher’s initial understanding, which can be discomforting
(Gadamer 1998). Sometimes inspiration (Langley 1999) or a relatively straightforward topic,
can make this process less painful, but in our experience, on most occasions it requires a lot
of hard grind. On a positive note, that does suggest that it is possible to achieve worthwhile
output through determined effort and reflection.
Rhetoric for generality and use
The notion of playing with writing leads to our next set of issues. Whilst rhetoric performs
many functions in the presentation of research output (Golden-Biddle and Locke 1993;
1997), we are particularly concerned with its role in conveying generality and the particular
form of usability relevant to reflective practice. The claim to generality, and thus the validity
of the research, must be conveyed through the style of language used to present it. The
writing style needs also to reflect the intended use.
18
To make the theory tangible, users need examples of real situations to which they can relate,
in which the theory is seen to apply. The theory itself does not need to be precise in its
description of the concepts because users will never employ them in exactly the same
circumstances as they were derived from. Rather, the concepts must convey possibilities that
might occur in the user’s situation. This is not simply an argument for the way in which the
theory is presented to practitioners; it is a statement about the nature of the academic rigour
of the theorizing process, and hence about the way in which the theory must be presented to
academic audiences too. It challenges acceptance of the difference between academics’ and
practitioners’ frames of reference (Rynes, Bartunek and Daft 2001). Thus, the rhetoric must
convey “accuracy” (Weick 1979) through faithfulness (Gergen and Gergen 2000; Lincoln and
Guba 2000) to, rather than close depiction of, the data from which it was derived. In so doing,
it facilitates a “match” between the portrayal of the original situation and the social
constructions of likely users. This means being able to trace the theory back to its roots in the
original data without undue abstraction, and to convey a sense of the data in the rhetoric.
At the same time, the rhetoric must convey “generality” through a sense of the tentative
relevance of the abstracted concepts to other situations. The trick – and this is the difficult bit
– is to make the relevance stand out while ensuring that the tentativeness is evident. Thus, to
fall back on our learning project example, we do not have to be tentative about the concept of
local collaborative process learning, because we collated masses of data supporting its
“existence”. On the other hand, we do have to be cautious in suggesting that any particular
form of it is a generic feature of collaboration because we have relatively few observations
and, in addition, they are open to a range of competing possible understandings (Boje,
Luhmann and Baack 1999; Kilduff and Mehra 1997). In writing a conceptualization it would
be appropriate to be firm in the first to demonstrate relevance, and slightly tentative in the
19
second to indicate that these are indicative rather than firm concepts.
Getting started on the writing process involves rhetorical trial and error in parallel with the
“meaning-deriving” trial and error discussed above. Our review of the six articles focussed on
in the construction of this paper (as described earlier) did, however, reveal a number of
rhetorical devices in play. Not surprisingly, “may”, “might”, “ possibly”, “could”,
“sometimes”, “often”, “frequently”, “generally” and “tends to” feature regularly in relation to
lower-level concepts and convey a degree of gentleness in the assertions being made about
them. Such assertions are usually backed up with data examples which are indicative of a
wider range of possibilities, and the phrase, “our data suggests”, is a useful way of conveying
“cautious certainty”. More definite phraseology is generally reserved for the higher level
concepts (i.e. those that are informed by many other interpretation concepts), those that can
be backed up with extant literature (including that reporting our own previous research) and
“the obvious” (i.e. statements that we believe most people would not question).
In developing a line of argument, another useful stylistic tool is the rapid switching between
past and present tense, as a means of distinguishing exemplars (past tense),
“One way they got around this was by ensuring that members had time to read and digest
information in preparation for meetings”,
from general assertions (present tense),
“One way of getting around this is to ensure that members have time to read and digest
information in preparation for meetings” (from the leadership in collaboration project
(Co-author and Author 1 2003: S67)).
Quite often a sentence originally constructed as a past tense example is redrafted into generic
phraseology.
20
Building interest and engagement
Our concern with conceptualizing has so far focussed on the issue of generalizing from
instances of data about practice in a way that is both valid and useful. At face value, “valid
and useful” might be thought good enough criteria to legitimate the output (Ayas 2003;
Marsick and Gephart 2003). In practice, however, the combination of the conceptualization
and the rhetoric through which it is presented needs also to engage the audiences – both the
academic and the practitioner audiences – through sparking interest and perhaps entertaining
them (Golden-Biddle and Locke 1993; 1997). Building interest into conceptualizations,
however, raises two problems.
Firstly, there is the difficulty of conveying the detail of process issues in an interesting
manner. This became clear to us during the second part of our leadership in collaborative
situations project. The first conceptualization had indicated the importance of the activities
that “leaders” carry out in order to progress a collaboration. Our intention in the second stage,
therefore, was to uncover and report on these. The problem that we encountered was that
although such a list might be very useful, it cannot be presented as conceptual or interesting.
Such detail is effectively the realm of appendices or technical manuals. Instead we
conceptualized the activities as being of two types – those that were in the spirit of
collaboration and those that suggested what we labelled collaborative thuggery – and
structured the argument around the tension between them (Vangen and Huxham 2003). We
also introduced sub-concepts associated with each type of activity; these were embracing,
empowering, involving and mobilizing, and manipulating and politicking respectively.
Structuring the data in such a way has the double function of engaging interest and
introducing significant concepts. As one manager put it,
21
“(the terminology) collaborative thuggery is doing the rounds of the (region)”.
However, it does not help with the important original agenda of mapping out the activities
relating to those concepts, other than by use of a limited number of examples in each of the
conceptual categories.
Secondly, the use of catchy conceptual labels as rhetorical devices is a helpful way of
engaging and retaining interest in theory, but can be provocative. Whilst a manager, for
example, commented to colleagues that,
“I can remember three things from (Author’s) presentation – collaborative advantage,
collaborative inertia and collaborative thuggery”
– which suggests labels are an important aspect of the handles that facilitate reflective
practice – the choice of label is not straightforward and can itself raise issues. For example,
we deliberately chose “collaborative thuggery” (a term originating in our data) for its surprise
– and hence, we hoped, memorable – element. It clearly achieves these ends. However, it is
also sometimes distracting. For example, while the term was introduced and published in the
context of research on leadership and collaboration, a brief reference to it was subsequently
objected to by a reviewer of a paper on a different subject:
“The notion of “thuggery” is a very loaded term …”.
The spirit of this remark has sometimes been echoed by others. The paradox is that the very
aspects of such labels that make them attractive and useful can also detract from their value
for some audiences. By contrast, the label collaborative advantage appears uncontroversially
attractive. The problem with that, though, is that it is increasingly being used by others who
either do not know – or feel bound by – the meaning that we have assigned it (Doz and
Hamel 1998; Dyer 2000; Hansen and Nohria 2004; Lank 2006; Kanter 1994). Multiplicity of
meaning is inevitable as people draw concepts into their own framework of understanding
22
(Heracleous 2001), but is problematic if it is incorporated in theory which takes a different
interpretation of the four dimensions of rigour of Figure 1. In such cases a concept label may
become associated with very different types of rhetoric and rigour and it may be difficult for
third parties to follow the distinctions in the use of the theory.
Summary and Conclusion: Use Matters, Specifically and Generally
The specific case: principles of theory development for reflective practice
Our starting tenets were that theory to support reflective practice needs to be sufficiently
recognizable to seem relevant to users, sufficiently generic to be translated by them for
application to their own context, framed in a mode that reflects the way users are likely to
experience their worlds and, at the same time, provide them with novel insight. We also
argued that it should have sufficient integrity in terms of not misleading the user towards
inappropriate action. Drawing together key points from the preceding discussion of the ICA
and RO-AR, we can now summarise the ways in which the application of the tenets is
demonstrated in that illustrative example approach; this has relevance for the consideration of
similar methodological approaches that have the same concern for the use dimension of
rigour. Having done that, we will comment on the general implications for theory that is
designed to support reflective practice. Finally in the light of the specific example of
reflective practice, we will return to the four dimensions of rigour.
Addressing the tenets: the ICA example. Each of the features necessary for the development
of theory for reflective practice is built into the ICA / RO-AR approach in a number of ways,
and each of these features could potentially be incorporated in similar ways in other methods
oriented towards reflective practice. Recognizability and closeness to the experienced world
23
is in-built in two ways. The first is through an emphasis on “faithfulness” and, thus, its close
link between theoretical output and data gained in action settings. The second is the inclusion,
in the portrayal of the theory, of examples from the data from which it was derived.
Integrity and generalizability is incorporated in three ways. The first is through a focus on
“possibilities” at every stage of the approach. The derivation of interpretation concepts and
the process of fleshing them out leads to coherence. At the same time, the emphasis on
conversational development through debate, and the carefully tentative portrayal of the
essence of the clustered material in the final rhetoric, support theoretical integrity through
minimising the possibility of superficial readings of the data. The second means of
incorporation is the interpretation of “accuracy” as relating to faithfulness to the data, rather
than close depiction of it; the third is the style of rhetoric and its ability to convey
possibilities that might occur. The notions of “tentative relevance” and “cautious certainty”
trigger a style of presentation that helps the user to know how to interpret the value of the
presented concepts.
The incorporation of creativity is demonstrated in the ICA / RO-AR approach in four ways.
The first is through the principle of taking data at the point of action and aiming to let it speak
for itself. Closely related to this, the second is the processes for “pushing at possibilities”
during the stages of analysis; the derivation of multiple interpretation concepts from data and
the fleshing out of these interpretations both lead to gains in variety. The third is the attention
to conveying insightful meaning in a manner relevant to the research theme through “forcing
the rhetoric”. This notion involves resisting superficial engagement with the data, thereby
taking elements of the emerging theory beyond bland statements of the obvious. Finally, the
fourth is attention to building interest and engagement through impactful conceptualizations
24
and attractive conceptual tags that have the potential to draw in the user.
General considerations: usability and acceptability
Usability: developing theory for use in reflective practice. Our discussion above has shown,
through the example of the ICA / RO-AR approach, how the tenets of a use dimension
concerned with reflective practice can be incorporated into methodological approaches. We
suggest that the tenets – recognizability and closeness to the experienced world, generality,
creativity and integrity – are relevant to any research approach that takes the “intended use”
dimension of rigour as relating to the support of reflective practice. We suggest that many of
the ways in which the tenets are incorporated in the ICA / RO-AR approach are translatable
to other approaches and worthy of consideration, in the spirit of “possibilities”. That is, while
the stages of analysis described earlier are specific to the ICA, the notions of faithfulness,
possibilities, tentative relevance, cautious certainty, forcing the rhetoric, inclusion of
examples, attractive conceptual tags and so on all have the potential to be interpreted in the
context of other approaches.
Acceptability: integrity in writing for academic gatekeepers. By considering here the
derivation of theory to support reflective practice, we have aimed both to derive tenets and
principles for theory in the context of that kind of use and to exemplify the more general
principle that “use matters” and that it cannot be separated from the other 3 dimensions of
rigour relating to content, method and form, if academic research is to do more than hide
behind façades of relevance (Kieser and Nicolai 2005: p277-278).
Clearly, the tenets and principles posited here for a reflective practice mode of use, are quite
different from those that would apply to other modes of use. For example, the notion that
25
multiplicity of interpretation is unproblematic, which is a central part of the argument here
for the reflective practice mode, might be untenable if theory was intended to suggest more
precise courses of action. Equally, it would be untenable in a purely positivist academic mode
of use. This suggests that as a discipline, management research needs to appreciate the use
dimension – in a way that is mutually consistent with the content, method and forms
dimensions (Figure 1) – across the range of usages if it is to understand validity and method
holistically. We argued in the introduction that the development of a holistic understanding of
the four dimensions should be a central part of any research process. If broader concepts of
rigour associated with different usages can be developed generally and accepted, the
challenges of actually doing research for different modes of use – and, in particular, that with
relevance to practice, which tends to be most contested (Gulati, 2007) – will be reduced.
There are clear implications for editorial policies and refereeing processes. These are
particularly relevant in the context of calls for editors to solicit a wider range of articles
reflecting academic-practitioner knowledge transfer techniques (Rynes, Bartunek and Daft,
2001). Broader concepts of rigour are likely to mean both acting outside traditional comfort
zones (Gulati, 2007) and critiquing work that is within traditional criteria for rigour but which
does not address – or avoids (Palmer, 2006; Tushman and O’Reilly, 2007) – use criteria. In
these terms, taking a ‘closet’ approach – in which the data that inspires the theory is not
acknowledged in the research account (Sutton 1997) – is unacceptable as it suppresses the
assumptions, norms and values that underlie the research (Rynes, et al, 2001). By
introducing here the four dimensions of rigour we aim to provide a way of anchoring
dialogue about this.
26
REFERENCES Alvesson, Mats 2003 ‘Beyond neopositivists, romantics and localists: a reflexive approach to interviews in organizational research’. Academy of Management Review, 28: 13-33. Alvesson, M. & Kärreman, D. 2007. ‘Constructing mystery: Empirical matters in theory development’. Academy of Management Review, 32(4): 1265-1281. AMJ 2007. Editor's forward: Carrying Sumantra Ghoshal's torch: Creating more positive, Relevant and ecologically valid research. Academy of Management Journal, 50(4): 745-747. Aram, John D., and Paul F. Salipante Jr 2003 ‘Bridging scholarship in management: epistemological reflections’. British Journal of Management, 14: 189-205. Argyris, Chris, and Donald Schön 1974 Theories in practice. San Francisco: Jossey Bass. Ayas, Karen 2003 ‘Managing action and research for rigour and relevance: the case of Fokker Aircraft’. Human Resource Planning, 26:2 19-29. Bechky, Beth A. 2003 ‘Sharing meaning across occupational communities: the transformation of understanding on a production floor’. Organization Science, 14: 312-330. Boje, David, John T. Luhman and Donald E. Baack 1999 ‘Hegemonic stories and encounters between storytelling organizations’. Journal of Management Inquiry, 8: 340-360. Bryman, Alan 2001 Social research methods. Oxford: Oxford University Press. Bryman, Alan. & Emma Bell. 2007. Business research methods. Oxford: Oxford University Press. Bryson, John, Fran Ackermann, Colin Eden and Charles Finn 2004 Visible thinking: unlocking causal mapping for practical business results. Chichester: Wiley Doz, Yves, and Gary Hamel 1998 Alliance advantage: The art of creating value through partnering. Boston, Mass: Harvard Business School Press Dyer, Jeffrey 2000 Collaborative advantage: winning through extended enterprise supplier networks. New York: Oxford University Press.
27
Eden, Colin, and Chris Huxham 2006 ‘Researching organizations using action research’ in Handbook of organization studies (Second Edition). S. Clegg, C. Hardy, W. Nord and T. Lawrence (eds.). London: Sage, forthcoming. Eisenhardt, Kathleen M., and Melissa E. Graebner 2007. ‘Theory building from cases: opportunities and Challenges.’ Academy of Management Journal, 50: 25-32. Gadamer, Hans-Georg 1998 Truth and method (revised second edition). New York: Continuum. Galibert, Charlie 2004 ‘Some preliminary notes on actor-observer anthropology’. International Social Science Journal, 56: 455-466. Gergen, Mary M. and Kenneth J. Gergen 2000 ‘Qualitative inquiry: tensions and transformations’ in Handbook of qualitative research (second edition).N.K. Denzin and Y.S. Lincoln (eds),1047-1065. Thousand Oaks, CA: Sage. Ghoshal, Sumantra 2007. ‘Bad management theories are destroying good management practices’. Academy of Management Learning and Education, 4: 75-91. Golden-Biddle, Karen, and Karen Locke 1993 ‘Appealing work: an investigation of how ethnographic texts convince’. Organization Science, 4: 595-616. Golden-Biddle, Karen, and Karen Locke 1997 ‘Constructing opportunities for contribution: structuring intertextual coherence and “problematizing” in organizational studies’. Academy of Management Journal, 40: 1023-1062. Gulati, Ranjay 2007 ‘Tent poles, tribalism and boundary spanning’: The rigour-relevance debate in management research. Academy of Management Journal, 50(4): 775-782. Hansen, Morten.T., and Nitin Nohria 2004 ‘How to build collaborative advantage’. MIT Sloan Management Review, 46: 22-30. Heracleous, Loizos T. 2001 ‘An ethnographic study of culture in the context of organizational change’. Journal of Applied Behavioural Science, 37: 426-446. Hibbert, Paul, Chris Huxham and Peter Ring 2008 ‘Managing collaborations’ in Handbook of interorganizational relations, S. Cropper, M. Ebers, C. Huxham and P Ring (eds). Oxford: OUP (forthcoming). Hodgkinson, Gerard 2001 ‘Facing the future: the nature and purpose of management research re-assessed’. British Journal of Management, 12: S1-S80.
28
Huxham, Chris, and Siv Vangen 2003 ‘Researching organizational practice through action research’. Organizational Research Methods, 6: 383-403. Jacobs, Claus D. and Loizos T. Heracleous 2005. ‘Answers for questions to come: reflective dialogue as an enabler of strategic innovation’. Journal of Organizational Change Management, 18: 338-352. Kanter, Rosabeth M. 1994 ‘Collaborative advantage: The art of alliances’. Harvard Business Review, 72: 96-108. Kilduff, Martin, and Ajay Mehra 1997 ‘Postmodernism and organizational research’. Academy of Management Review, 22: 453-481. Kieser, Alfred, and Alexander T. Nicolai 2005 ‘Success factor research: overcoming the trade-off between rigor and relevance?’ Journal of Management Inquiry, 14: 275-279. Kock, Ned 2004 ‘The three threats of action research: a discussion of methodological antidotes in the context of an information systems study’. Decision Support Systems, 37: 265-286. Langley, Anne 1999 ‘Strategies for theorizing from process data’. Academy of Management Review, 24: 691-710. Lank, Elizabeth 2006 Collaborative advantage: how organizations win by working together. Basingstoke: Palgrave Macmillan. Lather, Patricia A. 1991 Getting smart: feminist research and pedagogy with/in the postmodern. New York: Routledge. Lincoln, Yvonna, and Egon Guba 2000 ‘Paradigmatic controversies, contradictions and emerging confluences’ in Handbook of qualitative research. Denzin, N. and Lincoln, Y. (eds), 163-188. London: Sage. McGahan, Anita M. 2007 Academic research that matters to managers: On zebras, dogs, lemmings, hammers and turnips. Academy of Management Journal, 50(4): 748-753. Marshall, Catherine and Rossman, Gretchen 2006 Designing qualitative research. (4th ed.) Thousand Islands, Ca: Sage. Marsick, Victoria, and Martha Gephart 2003 ‘Action research: building the capacity for learning and change’. Human Resource Planning, 26:2 14-18.
29
Miles, Matthew, and Michael Huberman 1994 Qualitative data analysis: an expanded sourcebook. 2nd ed. Thousand Oaks, Ca: Sage. Palmer, Donald 2006. ‘Taking stock of the criteria we use to evaluate one another's work: ASQ fifty years out’. Administrative Science Quarterly, 51: 535-559. Parise, Salvatore, and Amy Casher 2003 ‘Alliance portfolios: designing and managing your network of business-partner relationships’. Academy of Management Executive, 17: 25-39. Pettigrew, Andrew M. 1997 ‘The double hurdles for management research’ in Advancements in organizational behaviour: essays in honour of Derek Pugh. T. Clarke (ed.), 277-296. Aldershot: Ashgate Pettigrew, Andrew M. 1985 The awakening giant. Oxford: Blackwell. Pettigrew, Andrew M. 1990 'Longitudinal field research on change theory and practice'. Organization Science, 1: 267-292. Prasad, Anshuman 2002 ‘The contest over meaning: hermeneutics as an interpretive methodology for understanding texts’. Organizational Research Methods, 5: 12-33. Raelin, Joseph A. 2002. ‘Public I don’t have time to think! versus the art of reflective practice’. Reflections, 4: 66-89. Raelin, Joseph A. 2001. ‘Public reflection as the basis of learning’. Management Learning, 32: 11-30. Rynes, Sara., Bartunek, Jean. M. and Daft, Richard.L. 2001 ‘Across the great divide: Knowledge creation and transfer between practitioners and academics’. Academy of Management Journal, 44(2): 340-355. Schipper, Frits 1999 ‘Phenomenology and the reflective practitioner’. Management Learning, 30: 473-485. Schön, Donald 1987 Educating the reflective practitioner. San Francisco: Jossey-Bass. Schwandt, Thomas A. 1996 ‘Farewell to criteriology’. Qualitative Inquiry, 2: 58-72 Stake, Robert E. 2000 ‘Case studies’ in Handbook of qualitative research (second edition). N. Denzin and Y. Lincoln (eds.), 435-454. Thousand Oaks, Ca: Sage. .
30
Stubbs, Mark 2000. ‘Action, knowledge and business-environment research: a case for grounded constitutive process theories and a sense of audience’. Business, Strategy and Environment, 9: 24-35. Sutton, Robert I. 1997 ‘The virtues of closet qualitative research’. Organization Science, 8: 97-106. Tushman, Michael. & O'Reilly, Charles III 2007. ‘Research and relevance: Implications of Pasteur's quadrant for doctoral programs and faculty development’. Academy of Management Journal, 50(4): 769-774. Van de Ven, Andrew H., and Paul E. Johnson 2006 ‘Knowledge for theory and practice’. Academy of Managment Review, 31(4): 802-821. Van Maanen, John., Sørensen, Jesper.B. & Mitchell, Terence.R. 2007. ‘The interplay betweeen theory and method’. Academy of Management Review, 32(4): 1145-1154. Vangen, Siv, and Chris Huxham 2003 ‘Enacting leadership for collaborative advantage: dilemmas of ideology and pragmatism in the activities of partnership managers’. British Journal of Management, 14: S61-S76 Wallace, Brendan, Alastair Ross, and John Davies 2003 ‘Applied hermeneutics and qualitative safety data: the CIRAS project’. Human Relations, 56: 587-607. Weick, Karl E. 1979 The social psychology of organizing. Reading: Addison-Wesley. Weick, Karl E. 2007 ‘The generative properties of richness’. Academy of Management Journal, 50: 14-19.
31
Figure 1: Dimensions of Rigour in Theory Building
Figure 2: Dimensions of Rigour in the Case Study
SUBSTANTIVECONTENT
RESEARCHMETHOD
INTENDEDUSE
FORMOF THETHEORY
SUBSTANTIVECONTENT
RESEARCHMETHOD
INTENDEDUSE
FORMOF THETHEORY
CONTENTThe Theory of
Collaborative AdvantageMETHODResearchOrientedAction
Researchand the
InterpretiveClusteringApproach
USEHandles for
Reflective Practice
FORMThemes -
OverlappingConceptual-izations of
Issues,Tensions,Dilemmas
CONTENTThe Theory of
Collaborative AdvantageMETHODResearchOrientedAction
Researchand the
InterpretiveClusteringApproach
USEHandles for
Reflective Practice
FORMThemes -
OverlappingConceptual-izations of
Issues,Tensions,Dilemmas
32
Figure 3: Levels in the Process of Generalization
Figure 4: Multiple Generic Interpretations of a Data Point
44 need for trust inlearning280 role of social
rituals in(blocking) learning
285 CS7 "how do wecreate an atmosphere
in which it is ok toask questions?"
286 creating alearning atmosphere
287 learning byquestioning
42
49 206 121 43
8
283 282 281 42
96
12
Key: italic script – data; plain script – interpretation; numbers – arbitrary identifiers (which indicate the sequence in which the items were entered); arrows – linkage; dashed arrows – links to data or interpretations not shown here; code (CS7) – identifier of the source of the data
forming interpretive clusters and writing conceptualizations
fleshing out individual interpretation concepts with data points
deriving interpretation concepts from individual data points
Generalising Activity
3
2
1
Level
forming interpretive clusters and writing conceptualizations
fleshing out individual interpretation concepts with data points
deriving interpretation concepts from individual data points
Generalising Activity
3
2
1
Level
33
Figure 5: Multiple Data Points Informing a Generic Interpretation Concept
43 BGN17 "We areworking with peoplethat may or may notbe unscrupulous… we
need to hold someknowledge back" sme
director
44 need for trust inlearning
119 SBN5b LAd'slegal advice re
knowing who you aregetting into bed
with, eg partner Coor its parent
120 learning aboutother parties inorder to protectyourself (ratherthan enhancingcollaboration)
121 implications foraspects of learning
of the need toprotect yourself
123 SBN5b "if itmatters if a partnerwalks off and speaksto your competitor,
includeconfidentiality in
your legalagreement" (LAd)
124 legal agreementsas a way of
protecting yourself
142 codes of conductas a way of
protecting networkmembers
143 SBN5d thenetwork's code of
conduct forconfidentiality
194 LA019 "can thepublic sector and
the voluntary sectorhave an equal
share?… those withexperience (public
sector) had theadvantage first"
195 greater priorlearning gives
advantage to thatpartner
205 LA039 "whenasking communities
questions, whatweight should you
give to theiranswers?"
206 basis forevaluating partner
knowledge offerings
207 prejudices aboutpartners' competence
285 CS7 "how do wecreate an atmosphere
in which it is ok toask questions?"
45
48 46
42
52
126 122
115125
208
287 286 280
Figure 6: From Interpretation Concepts to Clusters
42 role of emotionin learning
44 need for trust inlearning
49 interface ofpower and learning
307 ENACTMENT OFATTITUDE
284 280 177 8 41
285 206 121 43
275 257 192 48
139 8 35 125 22
Key: bold script – cluster labels; calligraphy script – conceptualization label