Industrial Marketing Management 39 (2010) 49–55

Contents lists available at ScienceDirect

Industrial Marketing Management

Sense and sensibility: Two approaches for using existing theory in theory-building
qualitative research
Poul Houman Andersen ⁎, Hanne Kragh 1
Aarhus School of Business, Aarhus University, DoGE, Department of Management, Haslegaardsvej 12, 8210 Aarhus V, Denmark

a r t i c l e i n f o a b s t r a c t

Article history: The paper addresses the role and use of theory in theory-building case research. How to use existing theory
Received 7 July 2007 in theory-building efforts is a crucial, yet somewhat overlooked challenge within qualitative research. On
Received in revised form 13 June 2008 that background, the purpose of the paper is to explore, compare and discuss two distinct approaches for
Accepted 13 February 2009
using existing theory and developing theoretical awareness in theory-building case studies. Upon discussing
Available online 23 September 2009
the role and use of theory in case study research in general, we identify two approaches for drawing on pre-
Keywords:
existing theory, labeled the in vivo approach and the ex ante approach. The in vivo approach takes departure
Theory-building in a single theoretical framework, anchored in a single paradigm, and the aim of theory-building is to
Research methodology gradually deepen, refine and complement this framework. The ex ante approach suggests that researchers
Philosophy of science should look for paradox in the form of theoretical tensions and use these to develop theory, meaning that
researchers should develop an elaborate theoretical understanding of meta-theoretical stances and use these
to generate different readings of the case material, subsequently prioritizing and integrating them into a
coherent pattern.
© 2009 Elsevier Inc. All rights reserved.

Researchers seeking to develop new theoretical insights derived tions and expectations (Burrell & Morgan, 1979). Few people would
from case research face an important dilemma: how to use existing disagree that our access to and framing of social reality into ‘cases’ is
theoretical frameworks in case study research and still remain flexible mediated by prior knowledge. However, despite that most would agree
in their theorizing efforts. However, for qualitative researchers few on this, there is a widely held misconception that the hall marks of
guidelines exist on how and when to most appropriately include qualitative research is that it is exploratory, inductive and grounded.
theoretical perspectives in their investigations (Locke, Golden-Biddle, Researchers fray from involving pre-existing theoretical frameworks
& Feldman, 2004). Not least in the field of business-to-business mainly for two reasons: using pre-existing theories is seen as
marketing, where research on many interesting phenomena, such as incompatible with conducting explorative research, and involving pre-
for instance the coordination of marketing activities with those of existing theory is by some likened with the proposition testing canons of
other activities in- and outside the organizational perimeter must be positivist research. However, unspecified theoretical expectations or a
researched in the context where they unfold, case studies hold a lack of theoretical knowledge may lead researchers to replicate pre-
prominent role in theory development (Bonoma, 1985). However, existing findings adding little to existing theoretical knowledge, or to
there has been a call for more rigorous procedures for qualitative produce massive amounts of data without any clarity with respect to
research in terms of clear procedures for detailing the research how this data can lead to novel insights.
process in a consistent way (Bonoma, 1985; Johnston, Leach, & Liu, Contrasted to quantitative research which relies on measurement
1999; Perry, 2001). and bracketing according to close-ended categorizations, the
Pre-existing theory provides a crucial challenge to theory-building researcher's openness to sensory impressions and subjective inter-
qualitative research. Research literature warns us against being too pretations of these are central in qualitative research (Strauss &
theoretically predetermined when conducting inductively oriented Corbin, 1990, pp. 41–42). We believe that the sense of qualitative
qualitative research, as this may prematurely lock our analytical focus research strategies can be balanced by the sensibility of pre-existing
and blind us from imaginative theorizing and from revealing new theoretical frameworks, as part of the developing inter-subjectivity
insights and theoretical breakthroughs (Glaser & Strauss, 1967; and validity in qualitative research. Reflecting on pre-existing theory
Maxwell, 1996; Mills, 1959; Weick, 1989). On the other hand, can be understood as part of the process where researchers engage in
researchers do not enter a field without some theory-driven specifica- a discourse with the scientific community. The purpose of this paper is
to explore, compare and discuss two distinct approaches for using
existing theory and developing theoretical awareness in theory-
⁎ Corresponding author. Tel.: +45 89 48 66 30.
E-mail addresses: poa@asb.dk (P. H. Andersen), hak@asb.dk (H. Kragh). building case studies. Each approach represents a different way of
1
Tel.: +45 89 48 68 63. balancing the quest for new theoretical insight with the quest for

0019-8501/$ – see front matter © 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.indmarman.2009.02.008
50 P.H. Andersen, H. Kragh / Industrial Marketing Management 39 (2010) 49–55

benefiting from and maintaining a dialogue with existing theoretical the phenomenon (Daft, 1983). The research process is iterative in
frameworks. We label the two approaches in vivo and ex ante scope, suggesting that the researcher continuously moves back and
approaches for drawing on pre-existing theory. The paper is forth between field investigation and theoretical reflection (Dubois &
structured as follows. First, we discuss the role and use of theory in Gadde, 2002; Eisenhardt, 1989; Orton, 1997). In qualitative research,
qualitative research aiming at theory-building as opposed to theory- researchers are the instruments of observation and interpretation.
testing. Then, we define and discuss the two approaches for use of Moreover, as part of their ongoing sense-making efforts, researchers
theory along with their underlying dimensions and differences. In the start to evoke interpretations immediately after they enter into
following discussion, we compare the two approaches, addressing the qualitative investigations. As pointed out by Dubois and Gadde
challenges involved with each one. We conclude by discussing the (2002), the main difficulty of case studies concerns the interrelations
contribution of each approach to business-to-business marketing of these various elements in the research process.
researchers. Theory-building progresses through either interpolation or ex-
trapolation (Weick, 1989). Interpolation refers to a process, where
1. The role of theory in theory-building qualitative research concepts and partial explanations are gradually deepened building
upon broad outlines of theory which may be refined and comple-
A frequent misconception of qualitative research is that a mented as research progresses. Extrapolation refers to a process
researcher is required to enter a field of study without any theoretical where concepts and theories are developed and modified by
ballast. This perspective is often ascribed to Grounded Theory, subjecting them to alternative theoretical perspectives with different
however erroneously (Suddaby, 2006). No researcher is a tabula explanatory abilities. Whereas interpolation entails a gradual adjust-
rasa upon which reality is imprinted. We enter the field as subjects, ment of developed insights resulting from interpretation processes,
preconditioned from previous formal training or from experience, and extrapolation is characterized by deliberate theoretical leaps among
our mental ballast will interact in framing reality for us. Theoretical schools of thought which may hold axiomatically different concep-
perspectives are fragments of an autobiography developed by an tualizations of core concepts and therefore may propose radically
individual researcher (Bedeian, 2004). Rather than refraining from diverging explanations and even different framings of what is to be
theoretical predispositions, qualitative researchers should embrace focused upon analytically in research inquiry.
and understand how they interact in their sense-making efforts Exactly because social contexts are unstable over time and may
during theory-building. unfold in unpredicted ways, pre-existing scientific theory may
In order to discuss the role of theory in qualitative research, we provide an ample source of inspiration to theorizing efforts. Moreover,
need to define what we mean by the concepts of theory and theory- because theories may not be applicable as grand concepts explaining
building qualitative research. Theory is an abstract notion, which is all incidents of phenomena, there is also room for simultaneously
defined in more or less precise terms. It is used arbitrarily throughout using more theories to encourage imaginative thinking, even if such
the literature and is often confused with other concepts such as mo- theories have conflicting views on the relevance of particular concepts
dels and propositions (Sutton & Staw, 1995). Theory may be for understanding the issues studied as well as on how such concepts
conceptualized as “an ordered set of assertions about a generic are interrelated. The role of theory in case research is to support the
behavior or structure assumed to hold throughout a significantly researchers' ability to focus, and to help sort and structure data in an
broad range of specific instances” (Sutherland, 1975:9). This idea of informative manner, in the situation of data overload characteristic
generalizability of theoretical insights, however, has been challenged, for qualitative research (Miles, 1979). However, focusing too strongly
particularly in social science research. Theories aiming to reach a high on pre-existing theoretical concepts may also temporarily blind
degree of generalizability do this at the expense of other aims of researchers in their quest for establishing new insights and may
research such as accuracy and simplicity (Weick, 1995). The definition therefore hinder theory development and lead only to induction and
of theory also relates to ontological conceptions of the nature of validation of previous theorizing (Weick, 1989). Therefore, in theory-
knowledge. Social scientists have addressed this in terms of building qualitative research, pre-existing theory should be seen as a
theoretical range and argued that social researchers should refrain means for imaginative theorizing, a resourced form of musing,
from the building of grand theory in the image of natural science. allowing for the free flowing interplay of observation and multi-
Instead, theory-building should take a bottom-up approach and focus relational reflection (Locke et al., 2004).
on engaging on theorizing and minor work hypotheses closer to the
social reality which they address (Merton, 1957). 2. Two approaches for involving pre-existing theory in theory-
For our purposes, theory-building is defined as the process through building research
which researchers seek to make sense of the observable world by
conceptualizing, categorizing and ordering relationships among In the literature on qualitative research strategies, very few
observed elements. This definition builds on Astley (1985) and contributions detail the processes of involving theory in theory-
Weick (1989), both of whom also focus on theory-building as a sense- building (Van Maanen, Sørensen, & Mitchell, 2007; Weick, 1989). For
making process, where the interplay of observation and multi- instance, Eisenhardt's frequently cited paper on the use of case studies
relational reflection through interpretation and authoring plays a for theory-building purposes (Eisenhardt, 1989), except for a brief
central role for the generation of new theory. We are explicitly suggestion to defer literature studies until the later stages of the
concerned with theory-building from case studies using mainly research process, does not discuss the role of theory in the processes
qualitative data. Theory-building has been contrasted to theory- of theory-building. From the perspective of a research practitioner
testing case research. Theory-testing using case studies concerns the embarking on qualitative research, we found that the literature
confirmation or refutation of propositions or predictions derived from concerned with the use of existing literature in theory-building
a theory (Wilson & Woodside, 1999). Using cases for theory-testing generally differs in terms of how as well as when the use of pre-
has been suggested in situations where events are unique and existing theory is suggested to take place. The issue of how concerns
therefore do not allow for alternative approaches such as sampling the researcher's preparation and treatment of inspirational theoretical
(Yin, 1994) or in situations where an insider's perspective is necessary resources in field studies. Particularly, it links to a fundamental debate
in order to understand the intricacies of the phenomenon under study among different ontological conceptions of the nature of knowledge
(Johnston et al., 1999). Theoretical designs using qualitative studies to and their epistemological consequences. Whereas one strand of
build new theory contrast with this approach (Eisenhardt, 1989), philosophical thinking on the nature of science sees knowledge as
because they emphasize the meaning rather than the measurement of justified true beliefs upon which subsequent theory-building must
 P.H. Andersen, H. Kragh / Industrial Marketing Management 39 (2010) 49–55 51

The issue of when is concerned with the particular stage in the

research process at which researchers should make use of existing
theory. Although access to empirical phenomena is always mediated
by theory, there is some variation with respect to when theoretical
perspectives are actively put to the fore in interpretation activities.
Eisenhardt, for instance, argues that consideration and discussion of
existing theory should be postponed until the later stages of case
analysis, “because preordained theoretical perspectives or proposi-
tions may bias and limit the findings” (Eisenhardt, 1989:536). Others
argue against such delay stating that it will only lead to unstructured
research rediscovering what is already known (e.g. Suddaby, 2006).
The literature that we reviewed indicated a clear connection between
the two dimensions how and when, resulting in the emergence of two
distinct approaches for the involvement of theory in theory-building
qualitative research. The two approaches are illustrated in Fig. 1.
Fig. 1. Identification of the in vivo and the ex ante approaches for using theory. In the following, we develop an ‘ideal type’ description of the
perspectives presented in Fig. 1, viewing each in a purified form. We
rest, others, notably researchers inspired by postmodernist thinkers, acknowledge that this exercise may lead us to gloss over the subtleties
are more skeptical with regard to the nature of knowledge as truth found within each perspective, just as we may from time to time
and see the role of researchers as providing resistance to letting a exaggerate the intentions of the authors whose work we ascribe to a
specific definition of reality dominate over another (Addis & Podestà, certain approach. Still, we find the description useful for understand-
2003; Pritchard, 2006). ing the divergence between the two approaches and the implications
Theories may be used more or less rigorously in interpretation. of adhering to either one. Building on the core questions when and
Blumer (1954:7) points out that although researchers may seek to how theory is involved in theory-building case research, we label the
develop definitive concepts in order to make clear categorizations in two approaches the in vivo approach and the ex ante approach. Their
terms of attributes and benchmarks, theories may also serve as main characteristics are summarized in Table 1.
sensitizing instruments, which “….lacks specifications of attributes The in vivo approach seeks to tackle the challenge of balancing
and benchmarks…instead it gives the user a general sense of reference theoretical focus with theoretical detachment through a process
and guidance in approaching empirical instances…they rest on a sometimes referred to as systematic combining (Dubois & Gadde,
general sense of what is relevant”. Remaining open to alternative 2002). In this approach, researchers take as their point of departure,
interpretations of data is however quintessential to the development an overall theoretical framework, but continue to scan, select and
of new insights. As pointed out by Morgan (1997), a given theoretical discard theoretical perspectives as this framework is confronted with
perspective (or metaphor is his terms) may be persuasive to an extent empirical data.
that it blinds for alternative interpretations leading to novel insights. Throughout the research process, this interpolation helps
Using alternative metaphors in interpreting data helps the researcher researchers extend and/or combine received theory with empirical
in grasping new or alternative meanings of data. It is in the emerging findings and other theoretical perspectives in order to build new
compatibilities and similarities, but also in the tensions between theory. As opposed to traditional inductive research processes in
theoretical frameworks that novel insights may occur, which lead to which reasoning is grounded in empirical observations only, the
construction of new theory, while at the same time keeping the processes underlying interpolation are characterized by abduction,
dialogue with existing theoretical frameworks. Therefore, as we see it, which refers to the process of generating plausible theoretical
the ability to apply multiple theoretical perspectives is an essential propositions from empirical observations (Locke et al., 2004).
aspect of gaining theoretical sensitivity in theory-building qualitative The ex ante approach suggests that researchers should look for
research, because it may prevent researchers from becoming blinded paradox in the form of theoretical tensions or oppositions and use
by one particular theoretical perspective. these to develop theory (Poole & Van de Ven, 1989). Used in a

Table 1
The in vivo and ex ante approaches for using existing theory in theory-building case research.

In vivo Ex ante

Ontological assumptions about the nature Associated with critical realist assumptions Associated with post modernist assumptions
of knowledge

Role of existing theory in theory-building Theories as inspirational sources used to frame research Theories as language games which the researcher must
projects and continuously understand developments in learn in order to use them individually and to exploit
the empirical data their contradictions ‘playfully’

Theory-building approach Theory-building as interpolation — complementary Theory-building as extrapolation — deconstruction and

theories are used continuously to refine and adjust a reconstruction of theories in order to challenge and
pre-existing theoretical framework possibly change boundaries separating existing theoretical
discourses

Stage and mode of theory involvement Single-perspective use of theory in the preparation and Broad, multiparadigmatic use of theory from the outset
early data collection stages, which may later be extended
and refined

Examples of authoritative literature Dubois and Gadde (2002) Morgan (1997)

Orton (1997) Lewis and Grimes (1999)
Poole and Van de Ven (1989)
52 P.H. Andersen, H. Kragh / Industrial Marketing Management 39 (2010) 49–55

qualitative research context, this means that researchers should ways of seeing. A synthesis is not possible, since in their purest form
develop an elaborate theoretical understanding of meta-theoretical they are contradictory, being based upon at least one set of opposing
stances and oppositions both prior to and during data collection and meta-theoretical assumptions….one cannot operate in more than one
analysis. The approach is inspired by post modernist conceptions of paradigm at any given point in time, since accepting the assumptions
science and focuses accordingly on theories as metaphors or language of one, we defy the assumptions of all the others” (Burrell & Morgan,
games, each of which unfolds a specific deceptive discourse (Lyotard, 1979:25). The need for additional theoretical insights is created as the
1979; Morgan, 1997). Using multiple theoretical vantage points and research process progresses and is a result of learning from empirical
treating these as language games rather than declarations of faith data. In other words, theory development is an outcome of ‘exposing’
helps the researcher in resisting and possibly also escaping theoretical existing theory to reality in order to understand theoretical concepts
imprisonment. Theory-building unfolds when individual metaphors and the forces shaping the empirical referent (i.e. a broadly defined
as well as differences between them are systematically used to phenomenon of interest (Lewis & Grimes, 1999)) at a deeper level.
generate overlapping and sometimes contradictory readings of the This difference between the two approaches reflects their un-
case material. These images are subsequently prioritized and derlying differing ontological assumptions. Unlike the postmodern
integrated into a coherent pattern. This process has been described conception of a multifaceted reality with no fixed reference points
as storytelling, because data are essentially used to construct stories, which will later be described as underlying the ex ante approach, the
which are continuously revised and elaborated (Alvesson & Kärre- in vivo approach assumes the existence of an independent and given
man, 2007; Daft, 1983), and the emerging patterns have accordingly social reality, along similar lines as articulated within critical realism
been referred to as the storyline of a case study (Morgan, 1997). (Fleetwood, 2005). The empirical world is viewed as existing
independently of our knowledge of it and instead of ‘revealing’ itself
2.1. The in vivo approach to theory-building through subjective construction, reality to a larger extent ‘controls’ or
guides the research process. The initial theoretical framing of an
The in vivo approach to theory-building using case studies treats empirical referent is important here, because interpretation processes
theory, data and analysis as equal, interacting elements of the research start from and are influenced by the researcher's frame of reference;
process. Lying somewhere between inductive and deductive processes however, the empirical referent is assumed to exist regardless of this
for knowledge creation, the approach is frequently applied by interpretation (Easton, 2002). Although reality may not be directly
qualitative case researchers, yet the interaction between its different accessible and perfect knowledge about it not attainable, it is
elements is less frequently discussed and when discussed, often not as nevertheless possible for the researcher to confront his theoretical
one coherent approach. So far, the this approach goes by different pre-understanding with the empirical world as such and to use this
names, e.g. systematic combining (Dubois & Gadde, 2002) and confrontation to uncover more aspects of reality. In other words,
iterative grounded theory (Orton, 1997). As indicated by these labels, reality is perceived as being ready for discovery by researchers. During
the research process relies on constant iterations between theory and the process of discovery, as empirical data point to relevant com-
data. Regarding theory and data as two interrelated processes plementary theories or alternative concepts that may be used to refine
occurring to a large degree simultaneously means that the researcher emerging theory and gain a deeper-level knowledge of the phenom-
makes sense of theoretical ideas by linking them to empirical evidence enon studied, the researcher goes through a process of continuously
and at the same time, transforms empirical evidence to results through adjusting and refining decisions concerning the appropriateness of
the use of theory and ideas (Ragin, 1992). This process is one of different theories. The use of alternative theories and concepts is,
interpolation and it is characterized by a gradual deepening of however, linked to the initial framework, meaning that the frame-
knowledge of key concepts. Aimed at generating plausible proposi- work stays within the boundaries of the paradigm chosen up front.
tions from data which perhaps do not make sense when viewed in the Theories serve as input to as well as outputs from data collection and
light of an initial theoretical framework, it is often understood in terms analysis. One ambition is to achieve the best possible ‘fit’ between data
of abduction (Dubois & Gadde, 2002; Locke et al., 2004; Reichertz, and theory. This fit is reflected in the case studies, which are considered
2004). The contribution of abduction lies in the way it urges the products of research operations and a means of reaching closure of the
researcher to not only anchor theory-building in empirical data, but relationships between theory and data (Ragin, 1992), preferably in a
also to simultaneously use puzzlement and questions arising from way where “…when the case is finally turned into a ‘product’, there
these data to confront original theoretical ideas and expectations with should be no confusing pieces left” (Dubois & Gadde, 2002:558).
new ideas from other theoretical perspectives. This process is aimed at Consequently, the in vivo approach is pre-disposed towards the
systematically discovering an order, here in the shape of a theoretical evolving theoretical framework as a mechanism for understanding the
framework, that fits surprising empirical facts, yet it is acknowledged empirical data, whereas in the ex ante approach focus is to a larger
that such discovery is a mental process which depends on individual extent on the evolving storyline, which is as much a product of the
creativity and intellectual effort (Reichertz, 2004). authoring efforts of the researcher as of the empirical referent. As
The approach is different from traditional inductive approaches explained by Alvesson and Kärreman (2007), data are inextricably fused
because it relies on a theoretical pre-understanding, which serves as with theory, and rather than seeing theory as an interpretive device, the
the point of departure for initial data collection and preliminary task for researchers is to re-author experiences through introducing
analysis. Hence, it shares with the ex ante approach the notion that ‘new voices’, both in order to produce new understandings and in order
pre-existing theory plays an important role in framing a research to challenge existing ones through disciplined imagination, rather than
question and building theory. At the same time, however, the in vivo produce theory closer to reality. The demands placed on the researcher
approach is markedly different from the ex ante approach with engaged in interpolation, however, are no less than those placed on the
respect to the nature and role of this theoretical framework. Here, the extrapolating researcher.
theoretical framework is typically rooted in a single ‘grand’ perspec- Starting empirical investigations in a single, albeit broad, theoret-
tive and focuses on a limited number of concepts which are then ical framework often reflects a certain degree of specialization or
gradually extended, changed or refined as the framework is preference on behalf of the researcher, which may expand the ability
confronted with the empirical world (Andersen, Borum, Kristensen, to develop new theory, but which may also lead to fixed preoccupa-
& Karnøe, 1992). Such paradigmatic anchoring is consistent with the tions blinding the researcher to radically new or different ways of
viewpoint of Burrell and Morgan (1979), who in their much cited and seeing a problem (Vaughan, 1992). Accordingly, the challenge of
acknowledged book on paradigms in social research argued for systematic combining lies in being able to evoke theoretical concep-
paradigmatic closure and explained, that paradigms “…offer different tions — even if these are found beyond the paradigm with which the
 P.H. Andersen, H. Kragh / Industrial Marketing Management 39 (2010) 49–55 53

researcher is associated — and yet to use them within the boundaries of person or group of persons, there is a strong possibility for incon-
the defining theoretical framework. This requires sensitivity towards sistency (Weick, 1989). Although challenging, this does not make the
anomalies in data along with the ability to withdraw from pre-existing process impossible, but calls for a consistent procedure for ensuring a
theoretical framings in order to look for and introduce novel theory. strong degree of openness in the initial data collection process as well
as in the process of reverting to the empirical field for following up on
2.2. The ex ante approach to theory-building both strong and weak leads.
Although there is a growing body of literature on the value and
The ex ante approach embraces the idea of a multiparadigmatic, importance of using paradoxical thinking in the area of theory
paradoxical or dialectical approach to developing truly novel insights development (Gioia & Pitre, 1990; Lewis & Grimes, 1999; Poole & Van
(Elsbach, Sutton, & Whetten, 1999; Poole & Van de Ven, 1989; Weick, de Ven, 1989), the literature on how to approach this way of thinking
1989). The general purpose here is not to build consensus among more systematically as a strategy for research is sparse in general, and
diverging theoretical perspectives, but rather to use their divergences throughout the social sciences in particular, the number of contributions
as vantage points for creating new insights and for challenging taken- in this area are minimal. One notable exception is the procedure
for-granted assumptions, as they typically unfold within a specific suggested by Lewis and Grimes (1999) for using and juxtaposing
research paradigm or, in the words of Alvesson and Kärreman, to “use theoretical perspectives. Their framework stipulates three overall steps
theory and imagination to critically open up alternative ways of in the research process, which are iteratively rather than sequentially
framing empirical material” (Alvesson & Kärreman, 2007:1267). In linked. These consist of a groundwork phase, which is mainly aimed at
line with Morgan (1997), who in more recent work has changed his preparing and sensitizing researchers to use multiple theoretical
view on paradigms as incommensurable (Burrell & Morgan, 1979), a perspectives. This phase also deals with how the use of such multiple
way of seeing is also a way of not seeing. All theories accordingly perspectives may frame the overall research question differently, thus
provide specific focus areas with persuasive but partial lenses, also helping the researcher to focus on the research question flexibly:
providing both insights and distortions. However, the post modernist understanding how diverging theoretical framings influence not only
inspired approach to paradoxical thinking does not only see theories the perspective from which the theory is understood, but also the
as lenses, but also as giving the reader/researcher authorship. Social concepts and conjectures available for framing a relevant research
reality is dynamic and paradoxical and therefore many things at once, question linking to this specific empirical referent in the first place.
which allows the researcher to see and use multiple or even The next stage is that of data analysis, which concerns planning a
schismatic theoretical perspectives simultaneously (Morgan, 1993). paradigm itinerary and conducting multiparadigmatic coding on the
Theories, accordingly, are not regarded as projections of underlying materials used. The issue here is to make selections from the potential
generalized patterns, but rather as instruments that can help the perspectives and to retain the most interesting ones for further
researcher illuminate, understand and not least problematize the inquiry. The selection process is characterized by what Weick
phenomenon studied (Alvesson & Kärreman, 2007). Moreover, reality describes as “thought trials”, where the researcher consistently
reveals itself according to the theoretical lenses applied (Morgan, poses ‘if…then’ questions on his/her understandings, and uses the
1997). Parallel to Wittgenstein (1953), Lyotard (1979) conceives cues to either dismiss particular perspectives in respect of further
science as a language game, where each scientific discourse is discussion or to engage them in further questioning and interaction
governed by a specific set of rules and its progress is dependent on with the field. Morgan (1997) describes the retention process as
them. In their theory-building efforts, researchers seek to extend or acknowledging the most convincing and interesting storylines. Note
even break these rules in order to change the game — a process that that an important part of the data analysis phase evolves in parallel
shares much with the acts of creative destruction that Schumpeter with groundwork phase, since iterating back and forth between
(1934) envisages are undertaken by entrepreneurs. Along similar theory and data, using theoretical perspectives to discover gaps in
lines, recently, Alvesson and Kärreman (2007) have emphasized the field investigation is an important part of the process.
potential contribution to be made from identifying or creating so- The third phase in the model relates to theory-building which
called breakdowns in understanding, which broadly speaking refers to concerns detecting and using tensions and paradoxes constructively in
an open-minded, deliberate search for the unanticipated or the the development of new insights. Four ways of working with paradox
unexpected; a process which is best accomplished by knowing and to build theories within the field of management and organization are
applying an extensive repertoire of theories and vocabularies. outlined by Poole and Van de Ven (1989). These are opposition, which
Conceiving the theory-building process as a language game, however, means keeping theses A and B separated and their contrasts
does not mean that anything goes. In scientific communities, appreciated and using their tensions to stimulate further theory
authority must be legitimized in order to have any effect. To be development — presumably within the paradigmatic conjectures of
rule-breakers, researchers need to have a strong understanding of the theses A or B. A second approach is to think of theses A and B as
rules of the game and must be able to play the game accordingly. If explanations situated to explain phenomena which are separated
not, their efforts will be deemed irrational or futile by other members spatially or in level of theoretical abstraction, as also suggested by
of the community. From this perspective, theoretical tensions and proponents of multilevel approaches to analysis (Pettigrew, 1990). A
paradoxes are useful for identifying fractures and lacunas between third procedure is to separate A and B in time, suggesting for instance,
particular language games, which may guide researchers in their that one set of conditions depicted by one theoretical insight, creates
efforts to successfully change the rules of the game. the conditions for different states of reality to occur — much in the
The crucial challenge is to break the deceptive powers that spirit of dialectical reasoning. As a fourth strategy, Poole and Van de
particular theoretical perspectives have over the researcher. One way Ven (1989) suggest synthesis, which involves transforming the
of doing this is to use multiple and diverging theoretical perspectives tensions between theoretical viewpoints and in effect creating new
simultaneously in conducting research (Maxwell, 1996). It is, concepts and perspectives that may align the phenomenon studied. As
however, important to use these theoretical perspectives emphati- they point out, these approaches are combinable.
cally, that is, they need to be used on their own terms and with respect
to the specific rationale and corresponding logic specific to each of 3. Discussion
them. Imagination processes where researchers seek to sensitize their
emphatic abilities and use multiple theoretical vantage points call for The in vivo and ex ante approaches to using theory in theory-
both academic discipline and reflexivity: since the processes of building case study research represent attempts to break away from
variation, selection and retention of ideas are carried out by the same the ‘linear storyline constraint’ associated with pure inductive
54 P.H. Andersen, H. Kragh / Industrial Marketing Management 39 (2010) 49–55

research, which is often perceived as progressing from methodological The ex ante approach provides the possibility of a comprehensive
considerations through data collection, analysis and theory-building understanding of the underlying conflicts among theoretical perspec-
without much iteration between the individual stages (Orton, 1997). tives and, hence, how their different explanatory power is linked to
Both approaches accordingly dismiss the idea of the ‘empty-handed’ different contexts and circumstances. As discussed above, engaging
researcher and instead advocate engaging his/her theoretical predis- these conflicting views with empirical data may be a fruitful way to
positions. The two approaches also argue for an active, extensive use of gain novel insights. Due to the meta-theoretical nature of processes of
theory from the earliest stages of research and lasting throughout the extrapolation, however, the insights will often be of an overall nature.
research process (Irvine & Deo, 2006; Lewis & Grimes, 1999; The in vivo approach, on the contrary, focuses more on the
Meinefeld, 2004; Orton, 1997). At the same time, the approaches are imperfections or blind spots of the single theory, and how, by
markedly different with respect to the nature and engagement of pre- combining this theory with complementary theories, a deeper-level
existing theory. The ex ante approach, on the one hand, advocates understanding of the empirical referent becomes possible.
the use of broad, multiparadigmatic frameworks, the underlying
tensions of which are used to challenge data and construct the most 4. Concluding remarks
interesting storyline for a case. The in vivo approach, on the other
hand, starts with a single-perspective framework which is then The key message of this contribution is that business-to-business
continuously adjusted and refined as it is confronted with the evolving researchers embarking on theory-building through case study
empirical world. research should make conscious choices with respect to how and
The underlying idea of the ex ante approach, i.e. that the when they involve pre-existing theory in their research strategies. We
researcher should strive to break free from the constraints associated have explicated some of the concerns we think researchers should
with taking a single paradigmatic stance, offers an intriguing way of take into consideration before making that choice. We believe that
thinking about theory development. At the same time, actual making these choices consciously will help the business research
application of the approach on empirical data and case construction community in general in developing more rigorous case research in
is difficult. It is not so much because very few guidelines to such an terms of better integration with existing research and/or in reframing
application exist, but more because of the elaborate theoretical work pre-existing beliefs and assumptions in the business-to-business
required up front. Not only must the researcher have or gain a marketing research community by introducing novel theoretical
substantial theoretical overview, s/he is also required to work with perspectives. The above discussion leads to a number of speculations
multiple facets of this overview in relation to the empirical referent on the usefulness of the ex ante and the in vivo approaches in respect
under investigation. In some cases, authoring of the case material in to including existing theories in theory-building. In our view, research
the light of particular theoretical framings may prove aimless and is about both interpolation and extrapolation — and both are equally
forced and, therefore, may not contribute to any significant further important. As pointed out, the perspectives differ radically from each
understanding of the case. Also, in terms of communicating results, other with regard to the role that existing theory assumes in the
researchers may find it challenging even to address the axiomatic process. Hence, it can be discussed whether the usefulness of the
understandings of one perspective, as they may interfere with the approaches (besides the researcher's core epistemological and
implicit rules of the institutional settings of a particular research ontological beliefs) also rests upon more pragmatic issues. Such
community. However, if successful, this strategy is also the one most issues include whether the theoretical terrain(s) in which the
likely to lead to insights that may seriously challenge taken-for- researcher travels is fragmented and evolving, which may point
granted assumptions within a particular field of research. towards an ex ante approach, or whether the landscape represents
The in vivo approach on the other hand would start out with a more coherent theoretical camps or schools, equipped with particular
framework anchored in a single theoretical perspective, hence institutions, such as academic journals, which may suggest than the in
providing for a more tight ‘theoretical control’, giving the researcher vivo approach would be easier to accept and adapt. This also reflects
more sense of direction in data collection as well as in the search for the key values and norms of the social community of scientists in a
complementary theory. Young researchers embarking on learning the particular research area. As we see it, both terrains are presently found
craft of research who find themselves at the lower level of the in a business-to-business marketing context and complement rather
research hierarchy may find this research approach to be the most than dominate each other.
rewarding and the least troublesome. The pre-determination of the in
vivo approach may, however, lead to a restricted view on behalf of the References
researcher. Another problem pertaining to this approach is that there
are no articulated guidelines on how the process should actually be Addis, M., & Podestà, S. (2003). Long life to marketing research: A postmodern view.
European Journal of Marketing, 39(3/4), 386−412.
managed in order to avoid such restriction. Researchers depart from a Alvesson, M., & Kärreman, D. (2007). Constructing mystery: Empirical matters in theory
chosen theoretical perspective and apply insights from other development. Academy of Management Review, 32(4), 1265−1281.
perspectives on an ongoing basis, also as a result of changes in the Andersen, I., Borum, F., Kristensten, P. H., & Karnøe, P. (1992). Om kunsten at bedrive
feltstudier — en erfaringsbaseret forskningsmetodik. Frederiksberg: Samfundslitteratur.
empirical terrain which they study. Sudden changes in the case Astley, W. G. (1985). Administrative science as socially constructed truth. Administra-
context (for instance a merger or a sudden unavailability of key tive Science Quarterly, 30(4), 497−514.
decision-makers) may provide researchers to redirect their research Bedeian, A. G. (2004). Peer review and the social construction of knowledge in
the management discipline. Academy of Management Learning & Education, 3(2),
efforts. 198−217.
Data collection in the ex ante approach poses another particular Blumer, H. (1954). What is wrong with social theory. American Sociological Review, 19
challenge, in that it requires a meta-theoretical sample (Lewis & (1), 3−10.
Bonoma, T. V. (1985). Case research in marketing: Opportunities, problems, and a
Grimes, 1999). As discussed above, collection of such data is a
process. Journal of Marketing Research, 22, 199−208.
cumbersome task and, faced with it, researchers have resorted to the Burrell, G., & Morgan, G. (1979). Sociological paradigms and organisational analysis.
use of secondary data. The more controlled theoretical framework Aldershot: Ashgate Publishing.
Daft, R. L. (1983). Learning the craft of organizational research. Academy of Management
underlying the in vivo approach, on the other hand, provides the
Review, 8(4), 539−546.
researcher with a focus that may be used to guide and direct data Dubois, A., & Gadde, L. -E. (2002). Systematic combining: An abductive approach to case
collection, and in doing so ensures a data set that more precisely covers research. Journal of Business Research, 55(7), 553−560.
the concepts of the theoretical framework. At the same time, this focus Easton, G. (2002). Marketing: A critical realist perspective. Journal of Business Research,
55(2), 103−109.
may hinder the researcher from trying to think outside the box and be Eisenhardt, K. M. (1989). Building theories from case study research. Academy of
properly aware of the existence of complementary theory. Management Review, 14(4), 532−550.
 P.H. Andersen, H. Kragh / Industrial Marketing Management 39 (2010) 49–55 55

Elsbach, K. D., Sutton, R. I., & Whetten, D. A. (1999). Introduction: Perspectives on Reichertz, J. (2004). Abduction, deduction and induction in qualitative research. In U. Flick
developing management theory, circa 1999: Moving from shrill monologues to E. von Kardorff & I. Steinke (Eds.), A companion to qualitative research London: Sage
(relatively) tame dialogues. Academy of Management Review, 24(4), 627−633. Publications.
Fleetwood, S. (2005). Ontology in organization and management studies: A critical Schumpeter, J. A. (1934). The theory of economic development: An inquiry into profits,
realist perspective. Organization, 12(2), 197−222. capital, credit, interest and the business cycle.
Gioia, D. A., & Pitre, E. (1990). Multiparadigm perspectives on theory building. Academy Strauss, A., & Corbin, J. (1990). Basics of qualitative research. Grounded theory procedures
of Management Review, 15(4), 584−602. and techniques. Newbury Park, CA: Sage Publications.
Glaser, B. G., & Strauss, A. (1967). The discovery of grounded theory: Strategies for Suddaby, R. (2006). From the editors: What grounded theory is not. Academy of
qualitative research. New York: Aldine Publishing Company. Management Journal, 49(4), 633−642.
Irvine, H., & Deo, H. (2006). The power of the lens. A comparative analysis of two views Sutherland, J. W. (1975). Systems analysis, administration and architecture: New York.
of the Fiji Development Bank. Accounting, Auditing & Accountability Journal, 19(2), Sutton, R. I., & Staw, B. M. (1995). What theory is not. Administrative Science Quarterly,
205−227. 40(3), 371−384.
Johnston, W. J., Leach, M. P., & Liu, A. H. (1999). Theory testing using case studies in Van Maanen, J., Sørensen, J. B., & Mitchell, T. R. (2007). The interplay between theory
business-to-business research. Industrial Marketing Management, 28, 201−213. and method. Academy of Management Review, 32(4), 1145−1154.
Lewis, M. W., & Grimes, A. J. (1999). Metatriangulation: Building theory from multiple Vaughan, D. (1992). Theory elaboration: The heuristics of case analysis. In C. C. Ragin &
paradigms. Academy of Management Review, 24(4), 672−690. H. S. Becker (Eds.), What is a case? Exploring the foundations of social inquiry
Locke, K., Golden-Biddle, K., & Feldman, M. S. (2004). Imaginative theorizing in Cambridge: Cambridge University Press.
organizational research. Paper presented at the Academy of Management Conference, Weick, K. E. (1989). Theory construction as disciplined imagination. Academy of
New Orleans. Management Review, 14(4), 516−531.
Lyotard, J. F. (1979). La condition postmoderne. Paris. Weick, K. E. (1995). Sensemaking in organizations. Berkeley, CA: Sage Publications.
Maxwell, J. A. (1996). Qualitative research design. An interactive approach. Thousand Wilson, E. J., & Woodside, A. G. (1999). Degrees-of-freedom analysis of case data in
Oaks, CA: Sage Publications. business marketing research. Industrial Marketing Management, 28(3), 215−229.
Meinefeld, W. (2004). Hypotheses and prior knowledge in qualitative research. In Wittgenstein, L. (1953). Philosophische Untersuchungen. In Ludwig Wittgenstein, Werkaus-
U. Flick E. von Kardorff & I. Steinke (Eds.), A companion to qualitative research gabe, Band 1, 1e Auflage. Frankfurt am Main.
(pp. 153−158). London: Sage Publications. Yin, R. K. (1994). Case study research. Design and methods. Newbury Park, CA: Sage
Merton, R. K. (1957). Social theory and social structure. Glencoe, IL: Free Press. Publications.
Miles, M. B. (1979). Qualitative data as an attractive nuisance: The problem of analysis.
Administrative Science Quarterly, 24(4), 590−601.
Poul Houman Andersen is head of the Centre for Design of Global Enterprise (DoGE)
Mills, C. W. (1959). The sociological imagination: Penguin Books.
at the Aarhus School of Business. His main research interest concerns the global
Morgan, G. (1993). Imaginization. The art of creative management. Newbury Park, CA:
organization and governance of market exchange activities between firms. He has
Sage Publications.
published in marketing as well as in management journals, such as California
Morgan, G. (1997). Images of Organization: Sage Publications.
Management Review, Strategic Organization, Industrial Marketing Management and
Orton, J. D. (1997). From inductive to iterative grounded theory: Zipping the gap
International Marketing Review.
between process theory and process data. Scandinavian Journal of Management, 13
(4), 419−438.
Perry, C. (2001). Case research in marketing. The Marketing Review, 1, 303−323. Hanne Kragh is assistant professor at the Centre for Design of Global Enterprise (DoGE)
Pettigrew, A. M. (1990). Longitudinal field research: Theory and practice. Organization at the Aarhus School of Business. Her research interests include the organization of
Science, 1(3), 267−292. international marketing activities, management of inter-organizational relationships
Poole, M. S., & Van de Ven, A. H. (1989). Using paradox to build management and and network change.
organization theories. Academy of Management Review, 14(4), 562−578.
Pritchard, D. (2006). What is this thing called knowledge? London: Routledge.
Ragin, C. C. (1992). ‘Casing’ and the process of social inquiry. In C. C. Ragin & H. S. Becker
(Eds.), What is a case? Exploring the foundations of social inquiry (pp. 217−226).
Cambridge: Cambridge University Press.