Methodological Reflections–Extended Version

Methodological Reflections, Extended Version

Methodological Reflections, Extended Version Carl-Henric Nilsson, 941230. "Methodology is a difficult subject. There is no easy way out. This is understood by every conscientious scientist" (Bjerke, 1981, p. 18). The objective of this chapter is twofold. Firstly, to describe the mental glasses through which I have viewed reality and my research. This is important for the ability of the reader to understand and evaluate what I have been doing and why. Secondly, I will present some methodological and paradigmatic propositions. This has necessitated a deeper and richer description in this chapter than otherwise deemed necessary. The methodological reflections are thus intended to stand on their own merits, as well as be a contextual framing of the appended articles. My aim in this chapter has been to start with methodology as the focal point and only address issues that in a conceptual sense are at most one step removed from methodology and furthermore addressed these issues from the perspective of the research process. The objective of research in general is to increase the accumulated body of knowledge. The research process for one research project concerns the creation of a fit between its three nodes: * the paradigm, which concerns ontology and epistemological stance, that is

the basic assumptions of reality and knowledge on behalf of the researcher, * the methodology, which is the methods and techniques available as

transcendence enhancing devices, * the problem, or the research question, which is the object of the conceptual

transition.

Figure 6. Three nodes of research. Research is a broad concept, not only embracing the research problem, but also bridging the basic assumptions of reality as well as the methodology, and further

PARADIGMontology

epistemology

METHODOLOGY

methods

techniques PROBLEM

research

question


the relation between the three. Just as important as the creation of an inter-project fit between the three nodes of a research process, can the creation of an intra-project misfit, or research tension, between the three nodes be. This research tension can be utilised as a driving force for refinement of the next research project. 5.1.1 Methodology and the Problem A core question for research is whether methodology should determine the problem or the problem determine the methodology. When reading the literature, the inter-scholar conformity of methodological approach is noticeable. Most authors appear to be caught within a narrow methodological frame of reference. Few, if any, alternatives to their standard strategies of research are utilised. If a scholar is publishing articles based on mathematical modelling he or she will most often always have done this and can probably be expected to continue to do so. Rarely will any of these authors publish for instance a case study. On the other hand traditional field researchers rarely publish any theories based on mathematical modelling. Of cause exceptions exist, but that seems to be what they are, exceptions. One line of argument for this methodological focus is that each area of research is fit for one type of method. Thus between a paradigm and a problem, there is just one suitable methodology. An other standard argument for this procedure is the same as the argument by Skinner (1974) for focused factory: focus on a limited, concise manageable set of article types, concepts, techniques and journals.1 This is clearly the most efficient method for producing square boxes, but is the scientific community primarily interested in square articles by square authors? Is this the research strategy that is most apt for fostering pioneering work and novel groundbreaking results? Probably not. An other, and regrettably more plausible, explanation for the monistic approach is that researchers choose problems to fit their favourite methods or apply these methods irrespective of the character of the problems at hand. A better approach would be that; firstly that the problem are allowed to determine the methodology, and secondly, that the principle of cross-fertilisation and changes of perspectives by utilising a multi methodological approach may be advantageous for the individual researcher as well as for the collective creation of knowledge. New perspectives on a problem, may it be changes of the theoretical foundation, research method, new research constellations or some other point of view, is better suited to surpass the standardised incremental changes of an research area with significant, creative changes. These ideas will be further pursued, firstly in relation to paradigm, and secondly in relation to methodology. The two remaining nodes of research: methodology and paradigm can be used as two dimensions for extending a sphere of research possibilities.2 Some 1 For Skinner, it was a limited, concise manageable set of products, technologies, volumes and markets. 2 It is often not possible to define a study as one and only one paradigm or one and only one methodology, however it can be possible to identify the paradigm or methodology that comes


intersections are more plausible than others. During the research for the thesis I have used four methodological approaches; conceptual studies, case-studies, survey and meta-analysis. It is commonly held that the paradigm of a researcher is invariable. However, I argue that I have utilised three paradigms during my research: positivistic, system theoretical and interpretive.

Figure 7. The limited sphere of research possibilities, utilised in the thesis. The methodological dimension is described in some traditional aspects of relevance for the thesis. The paradigmatic dimension is described in more detail, ending in a paradigmatic proposition, related to my experience during the research process. Also a methodological proposition is presented in the discussion on methodology and the thesis. 5.1.2 Methodology This discussion of methodology is by no means comprehensive, but is delimited to the aspects that are relevant for the thesis. Furthermore is there a strong relation between methodology and paradigm, which makes it hard to describe methodology in an unobjectionable way irrespective of paradigmatic affiliation of the reader, and still give the methodology a content. This description rests on models and concepts that can be more apt for one paradigm than the others.

closest to the what was used in a study in order to provide a comprehensive picture of the methodological and paradigmatic aspects of the thesis.

PARADIGM

METH

ODOL

OGY case

study

survey

meta-analysis

positivistic system interpretive

reductionism------holismgeneral--------particular

objective------------------------subjective

nomothetic

ideo-

graphic

empirical

conceptual

theory

generation

theory

verification

deductive

inductive

conceptualstudy


Research can be related to two worlds, the empirical world and the theoretical world. Theory can be generated from the empiri or by development of existing theory. Theory can also be verified by relating the theory to the empirical information.

Figure 8. A model of theory development by generation and verification. Induction is development of theory based on empirical observations. According to empiricists this is the only valid way of developing theory, since all knowledge is built from experience. Deduction is development of theory from theory which is advocated by rationalists who argue that logical thinking is the most important source of knowledge. Verification is the connection back to the empirical world in which the theory's relevance can be tested. Some hypotheses, for instance those concerning the future can not be tested, therefore the demand for verification can not be satisfied and can be replaced by confirmation. The opposite method for testing is proposed by Popper (1935), arguing that most hypotheses can not be tested, but they can be falsified. This principle is connected to his idea of the growth of knowledge as the central theme for research, and as an extension he proposed that instead of choosing the most probable theory, the most improbable theory should be chosen on account of being the most interesting theory. This theory could then be tested, and if necessity, rejected. Deductive generation of theory is purely theoretical, without connection to any empirical information. This divide research into two types empirical, that is connected to the empirical world either by inducing theory from practice, or by verification of theory. On the other hand there is conceptual research which is strictly theoretical in the sense that it starts in theory and ends in theory without being affected by empirical information. Conceptual research can however at a later stage be tested for empirical validity. Research can be further divided into a idiographic and a nomothetic approach.3 The idiographic approach support the idea that one can only understand social phenomenon by getting close to the investigated subject by exploring for instance its detailed background and life history. The analysis is focused on the subjective

3 "... both nomothetic and ideographic methodologies can be employed in a deductive and inductive sense" (Burrell and Morgan, 1979, p. 9).

THEORY

EMPIRIInduction

Deduction

Verification


understanding and explanations, which are generated by getting inside situations, instead of examine as an external observer. It is further emphasised that the subject should be allowed to unfold its nature and characteristics during the research process (Burrell and Morgan, 1979). The nomothetic approach, on the other hand, focus on systematic protocol and technique. It is most often preoccupied with testing hypothesis with the "canons of scientific rigour" (ibid., p. 6) and often apply quantitative methods and standardised research instruments. The distinguishing point is whether a study should be performed through intensive, multi-aspect in-depth studies of few cases (idiographic) for descriptive studies of individuality or an extensive study across a large sample of few aspects (nomothetic) in search for laws (e.g. von Wright, 1971). 5.1.3 Methodology and Paradigm The goal of science is to increase the accumulated body of knowledge, to that most people would agree. Most scholars would also agree to that there are things, of which we do not have absolute knowledge. However the definition of science, the definition of knowledge and the interpretation of what we do know can differ substantially. Each researcher implicit or explicit make certain basic assumptions about the constituents of reality. The decisive point in these assumptions of the basic construction of reality is that they can not be empirically or logically tested, but only reflected upon intellectually (Bjerke, 1981). The sets of assumptions, stipulations or normative propositions can collectively be called a research paradigm (Kuhn, 1962). A research paradigm contains philosophical preconceptions that guide the researcher to research objects, applicable methodology for the research, and also the interpretation of the results. It defines the boundaries of research that is sanctioned by members of this area of research. These paradigmatic assumtions in essence concern the basic construction of reality, the view of research and its aim, what constitutes scientific ideals and ethical and aesthetic elements of scientific progress (Bjerke, 1981; Törnebohm, ????). Kuhn's propositions have not passed the scientific community uncontradicted (e.g. Suppe, 1974). Kuhn has, after the initial publication in 1962, subsequently made several specifications of the concept of paradigm. In this presentation I relate particularly to three conceptions of paradigm. Firstly the basic meta-theoretical and philosophical proposition concerning the concept of paradigm as an implicit or explicit view of reality that bind together scientific communities. To this view I subscribe. Secondly, that paradigms shift in phases of revolution in which one paradigm supersede the other. This conception I regard to be too restricted and monistic. Thirdly the position that one researcher can not comprehend more than one paradigm, which I will discuss more after a brief discussion on the second notion. Kuhn is referring to natural sciences, also when discussing paradigm shifts. However researchers from other fields, such as theory of science (Törnebohm, 1975, 1982) and business management (Arbnor and Bjerke, 1994) maintain that paradigms may develop evolutionary and may furthermore coexist. I conceive


paradigms as coexistent and that they transcend through evolution as well as revolution. It can hereby be implicitly assumed that I define natural sciences not to be my field of research. However, since I do not regard the isolationistic perspective to science to be the most fruitful, I argue that a forced choice of paradigmatic affiliation would not be a true reflection of my research. With educational grounding (borrowing from the useful terminology of Glaser and Strauss, 1967) in both natural and social science traditions, the possibility of shifting perspective and identifying the entities of paradigmatic content with and also as something4 (Israel, 1979) can be utilised as methodological and scientific propellants in the research process. In practical terms: keeping a foot in both camps can be advantageous, since it allows for an inside as well as outside perspective on the content of research as well as the research process. Three paradigms have shaped my view on science: the positivistic, system theoretical and interpretive paradigms. To these I will relate my research, in the assurance of the existence of other paradigms as well as differences of opinions on these three paradigms. Some researchers even doubt if paradigms other than their own qualify as science. This, however, tells more about themselves than of researchers from other fields. According to Kuhn paradigms are incommensurable, it is not possible to frame scientifically usable criteria for truth and quality of research that retain inter-paradigmatic validity. It is thus not only the results of the research that is affected by the research paradigm. The evaluation, or even the capability to evaluate the results of research at all, is dependent upon the paradigm of the consumer of the research. According to a radical interpretation of Kuhn, researchers will see what they are educated to see. Researchers from different paradigms literally live in different worlds of experience, seriously obstructing a dialogue. Such a discussion can be compared to a political discussion, where non of the parties will diverge from their basic assumptions, what so ever. The assumtions underlying research paradigms can be analysed according to a hierarchical typology. Overarching any research are the constitutive ontological assumtions of reality. Concretisation of these fundamental assumptions in terms of favoured metaphors (Morgan, 1980) or basic schools of thought, through which the assumptions become meaningful, represent a particular epistemological stance, or a particular view of the possibilities of knowledge (Oliga, 1988) and the creation of knowledge. Epistemology concerns the understanding of reality, and how we communicate this to other people. There is hence a close relationship between ontological assumption and epistemological stance. I refer to this as a paradigm5, since the way of achieving knowledge is affected by ones ontological 4 Identifying something with something means to identify based on similarities, thus arranging entities into identical taxa, and identify these entities as something, focus on the differentiating aspects of the entity that makes it unlike other entities. 5 My view of the term differ somewhat from that of the referred scholars. For instance Arbnor and Bjerke, argue that paradigms exist at two levels: the upper which connects basic assumtions to methodology, and a work paradigm, that connects the methodology to the research area, both of which I embrace. Morgan (1980, p. 606-607) has a more intuitive definition: "an implicit or


assumptions, an epistemological stance will imply a certain ontology. This view is supported by Björkegren (1989, p.117).The epistemological stance will also affect the premises of the research process. Burrell and Morgan (1979) make an extensive inquiry into the role of paradigms as views of social reality and argue that social theory could be usefully analysed in terms of four broad paradigms, which are reflected in different dimensions of meta-theoretical assumptions (Morgan, 1980). One of these dimensions is the subjective - objective dimension, in which approaches to social science is dichotomised at four levels (Burrell and Morgan, 1979, p. 3): * ontology as nominalism - realism, * epistemology as anti-positivism - positivism, * human nature as voluntarism - determinism, * methodology as idiographic - nomothetic. Dichotomisations can be advantageous since they define the opposites and thereby the endpoints of a dimension. However a dichotomisation embrace only the first two parts of a dialectic6 relation, the thesis and the antithesis. The synthesis is omitted. Much of the scientific discussion can be viewed as following the generic structure of dialectics. If phenomenon within themselves contain their own opposition and thereby intrinsically carry the prerequisites for its own transcendence, this can be taken as an explanation to that all phenomenon, whether scientific or not, are part of an continuous developmental process, that by proof of induction7 can not have an end. A comforting thought for a researcher and an explanation to the fact that much research seem to create more questions than answers. This is often illustrated in the final chapter of doctoral theses under the heading, future research. Burrell and Morgan (1979) propose that the four paradigms of sociological research that they develop are co-existing paradigms, thus take a contradictory stance or propose a anti-thesis to Kuhn's thesis of paradigms as superseding each other. Burrell and Morgan further maintain that their four paradigms are mutually exclusive, and indirectly generalise this proposition to be valid for all explicit view of reality". De Mey (1982, p. 36) suggest that a paradigm is a very elaborate cognitive structure that is supposed to specify the prerequisite knowledge necessary for processing the information in a given area of science. Kuhn (1970) finally, who introduced the concept, used paradigm to embrace at least twenty-two different meanings in the original text (Masterman, 1964). 6 Dialectics is a concept central to philosophy, and was the classical antiquity's method for reaching the truth, according to which a problem is to be discussed from several contradictory perspectives. The concepts has gone through a substantial evolution, also in modern times by for instance Hagel and Kant. According to several interpretations, dialectics is constructed around the trinity: thesis - anti-thesis - synthesis. Each thesis intrinsic contains its own anti-thesis and from the two, the synthesis is formed. The synthesis is constructed from constituents of both the thesis and the anti-thesis and further contain some novelties, whether content or structure, beyond the pure combination of the thesis and the anti-thesis. 7 A proof by induction that P(n) is true for all positive integers n proceeds in two steps: 1: Prove that P(1) is true.

2: Prove that !n( ) P n( )" P n + 1( )( )


paradigms by stating that the four paradigms are "alternatives, in the sense that one can operate in different paradigms sequentially over time, but mutually exclusive, in the sense that one cannot operate in more than one paradigm in any given point in time, since in accepting the assumptions of one, we defy the assumptions of all the others" (Burrell and Morgan, 1979, p. 25). The paradigms in their pure form are contradictory since they are based on at least one set of contradictory meta-theoretical assumptions (ibid.). Hereby Burrell's and Morgan's synthesis on paradigms can be viewed as a proposition for the co-existence of paradigms, but monistic utilisable by an individual researcher at a given point in time. Arbnor and Bjerke (1977, 1994) propose a trinity of paradigms of social science (more specifically they are focusing business management), thus in some respect take the opposing stance to Burrell's and Morgan's dichotomisational perspective on paradigms. Arbnor and Bjerke argue that three harmony groups, exist. A harmony group is a unit with a fit between problem, solution techniques and basic assumptions. Made operational to the researcher this means that the scientist should chose research techniques that are in harmony with the scientist's basic assumptions of reality and with the problem that seem to be at hand (Bjerke, 1981). The three paradigms are: the analytic or positivistic, the systems theoretical and the actors or interpretative paradigm (For a thorough discussion on these paradigms see: Arbnor and Bjerke, 1977, 1994; Bjerke, 1981; Larsson, 1989). The analytic paradigm is the oldest approach and coincides with research in natural sciences. One basic assumption is that there exist an objective reality that is independent of individual human beings, such as the researcher. The objective of research is to explore this reality. The building of scientific knowledge thus follow formal logic procedures, and is independent of subjective impressions on behalf of the researcher. Scientific knowledge further is of additive character, which means that the whole can be explained by the sum of the parts. The aim of research is to reconstruct the reality and establish causal connections between entities, often to explain and to guide or forecast. The next paradigm is the systems theoretical, which emerged in the 1950's as a reaction to the additive picture of reality of the analytic paradigm. The systems approach also believes in an objective reality but opposed to the analytical paradigm they believe in synergistic effects. This means that the relation between entities comes into focus since they can add positive or negative effects to the system. Multifinality8 and equifinality are acknowledged as well as systems within systems. A system can be studied as opened or closed. "... a system together with its environment makes up the universe of all things in a given context" (Hall and Fagen, 1963, p. 83). In a closed system, the relation to the context, the environment is omitted.

8 Multifinality means that a cause can have several effects, while equifinality means that alternative causes can have identical effects.


Systems can further be modelled from a structural perspective, focusing on the system components and relations at a given point in time in order to determine the system's state, or from a processual perspective in order to study the change between states. The latter can in system theoretical terms be viewed as an aggregated systems analysis where the systems state at different times are the components and time is the relational determinant. The scientific language becomes an important transitional device in which concepts can carry substantial content and context. The aim of research is to describe and to determine finality connections, that is the connection between the propelling forces of the system and their positive or negative effects. One of the more philosophical aims of the general systems theory movements were to promote inter-disciplinarity by improving the communication between specialists (von Bertalanffy, 1968). The idea was to develop principles, crossing individual sciences in order to promote a unity of sciences. This scientific utopia, however, yet remains to crystallise. The third paradigm is the interpretive paradigm which use a phenomenological approach to science. This is a the most recent of the three paradigms and emerged in the late 1960s. Reality is assumed to be a social construction (e.g. Berger and Luckmann, 1967), which is dependent on the actors. Reality consists of several different images, picturing the meaning and content different actors put in their actions and surrounding environment. Knowledge is a subjective property and reality does not exist in an objective sense. Systems do not exist as objective entities as in the system paradigmatic sense but exist only in the mind of the scientist. The reality of several individuals can have common components that constitute an objectified reality9 for this group of individuals, whether scientists, organisations or the society. This objectivity is not universal, but created by man and can therefor be questioned and changed. If changed, the altered objectified reality will in turn influence the agent of change dialectically. (e.g. Arbnor and Bjerke, 1994) Language are in practice used at two levels: "At the lower, he[or she, the scientist] tries to give an extensive as well as multiple description of the various meanings given to the acts of individual actors. At the scientific level, however, it is important for the scientist to denote the meaning of subjective logic at the level of the individual actor as well as at the level of their dialectical relations to one another" (Bjerke, 1981, p. 16). Intentionality is the link between object and subject. Intentionality is the structure that provides content to experience. Figuratively speaking intentionality is meta-intentions. The concept of intentionality is central to the interpretive paradigm and is inconsistent with the analytical as well as systems paradigm. Man is imaged as an active creator of surroundings objects, since the objects are argued to be shaped by the way they are apprehend by the observer. The aim of research is to catch the subjective logic in man, which is made operational through the aim of understanding and

9 Reality is however not objective in the sense of the other two paradigms.


describing dialectic relations of the multiple apprehension that exist in the social reality. The scientific arena can from this perspective be seen as a field of dialectic processes where the meanings and content of the objectified reality will affect each researches subjective image of reality. These images in turn manifest themselves in the researchers scientific contributions, changing the objectified reality of the scientific area. Science and research is hence a collective process demanding input from each individual researcher and providing a means for output of the individual research results, thus adding to the accumulated body of knowledge. Arbnor and Bjerke (1994) conclude that paradigms are not always used in there pure form and pose the rhetorical question: Does this mean that the approaches without further notice can be combined? The proposition is vigorously denied. "How can it be possible to combine the belief of an objective, from us independent, reality containing causal connections with the belief of a social constructed reality containing dialectic relations?" (Arbnor and Bjerke, 1994, p. 460). However, it is argued, that it is possible to make one of the approaches into the basic approach by accepting the basic assumptions of this approach. After that it is possible to use parts of the other paradigms within the chosen one. If results or methods from one approach is used in an other approach they will be reshaped by the context of the latter. But the basic assumption remain: paradigms do not mix. Contrasting the dichotomisational view of Burrell and Morgan (1979) in an other way than Arbnor's and Bjerke's (1994) "triadisation" of the dichotomy, Berger and Luckmann presents a forceful argumentation for complementarity. Starting in two of the most famous and influential "marching orders" for sociology: Durkheim's (1895) objective positivism versus Weber's (1922/1976) subjective German idealism, the complementarity of the objective - subjective dimension is advocated. "Durkheim tells us: 'The first and most fundamental rule is: Consider social facts as things.' And Weber observes: 'Both for sociology in the present sense, and for history the objective of cognition is the subjective meaning-complex of action.' These two statements are not contradictory. Society does indeed possess objective facticity. And society is indeed built up by activity that expresses subjective meaning. And, incidentally, Durkheim knew the latter, just as Weber knew the former. It is precisely the dual character of society in terms of objective facticity and subjective meaning that makes its 'reality sui generis'"10 (Berger and Luckmann, 1967, p. 18). Since social reality always starts in meaningful human action, it continues to carry meaning, since the original may be reconstructed, even if it is opaque to the individual at a given point in time (ibid. p. 197). Continuing in the spirit of Berger and Luckmann, Larsson (1990) scrutinise Burrell's and Morgan's four level dichotomisation of the objective - subjective dimension. The main ontological debate between nominalists and realists on

10 Reality of its kind.


whether there exist identical or merely resembling entities is punctuated by Aaron (1967) by showing that both positions have partially valid points but they are also both incorrect in assuming that the opposite position is completely wrong. The epistemological polemics described in Burrell and Morgan crude positivistic - anti-positivistic dimension is argued to neglect the issue of holism versus reductionism by positioning systems theory on the objective side with the positivists in spite of system theory's advocates' strong opposition to reductionism, mechanism and behaviourism (e.g. von Bertalanffy, 1968). If instead holism is positioned with the anti-positivists, then Muir's (1982) suggestion that holism and reductionism are compatible holds true. Therewith, also at the epistemological level, complementarity is present, even though commonly not acknowledged. The dichotomisation between voluntarism and determinism of the human nature debate is also argued to be superficial on account of several contributions concerning the possession of free will, however under external restrictions (e.g. Barnard, 1938) and propositions of actor and systems oriented approaches (Crozier and Friedberg, 1980; Baumgartner, Burns and DeVille, 1979). Finally in the methodological dimension, Larsson (1990), focus his efforts in advocating the traditional dichotomisation between idiographic and nomothetic approaches as a misinterpretation of what is really a trade of problem. The major contradiction is; whether to chose an intensive study of few objects but many aspects in depth, which is done in a traditional case-study, or to perform an extensive study of few aspects of a large sample of objects. It is argued that the major contradiction is a practical and economical restriction rather than a scientific contradiction, since it is very resource demanding to perform a study that is both extensive and intensive. There is further a lack of methodologies for performing such research. However in his second thesis, Larsson (1989) introduce and refine such a method called the case-survey method, in which intensive case-studied are treated with the extensive research arsenal of quantitative methods to reach conclusions concerning several aspects across several cases, thus utilizing the potential synergistic advantages of both idiographic and nomothetic methodologies. From this discussion we conclude that there are objects, upon which the analytical paradigm is focused, there are relations, which the system paradigm focus, and there are subjects which the interpretive approach focus. None of these paradigm deny the existence of the other two, while some orthodox scholars maintain that their paradigm is the only one that is worthy of being called science.11 Most scholars would thus agree to that objects, relations and subjects are present in most situations. With the obvious risk of being marked as heretic, I propose that:

11 However, since they are probably a dying breed and a contradictory stance on that matter would be to force an open door, the orthodox are left to their fate.


Paradigmatic flexibility is possible and can be advantageous, especially for the individual researcher, but also for the research community in bridging some of the seemingly insurmountable antagonism between advocates of different paradigms.

At the present state of affairs, each paradigmatic cluster of researchers can at best be indulgent towards the others, in the firm assurance of their own excellence, and at worst be wasting a lot of efforts and good men in defaming the work of representatives of other paradigms within their own field of research (and astonishingly enough also in fields of research other than their own!). The majority of the methodologists appear to be resistant to the idea of the simultaneous pursue of several paradigms? Are they afraid of being "stuck in the paradigmatic middle"?12 I argue for a paradigmatic flexibility as a research strategy to bridge over the threats and reaping the potentials of a multi paradigmatic approach. I furthermore propose that this can be advantageous on the level of the individual researcher. There are scholars who propose advantages from the simultaneous pursuit of different paradigms from a collective scientific community's point of view. "The challenge presented to orthodox organisation theory by these different paradigms [Burrell's and Morgan's (1979) four paradigms, authors note] is to rethink the very nature of the subject to which it is addressed. Different paradigms embody world views which favor metaphors that constitute the nature of organizations in fundamentally different ways, and which call for a complete rethinking as to what organization theory should be about" (Morgan, 1980, p. 620). According to Morgan the idea of simultaneous pursue of paradigms will have revolutionary implications to the research process as well as to the results of research. "The essence of dialectical analysis lies in the idea that we should accept all research strategies as having something to offer but attempt to use their competing insight within the context of a single analysis" (Morgan, 1983, p. 380). Also Crozier and Friedberg, (1980) and Baumgartner, Burns and DeVille (1979) argue for complementarity, especially between the systems theoretical and interpretive paradigms, thus advocating a multi paradigmatic approach. This however should not be misinterpreted as identical to Feyerabend's (1975) proposition of theoretical and methodological anarchism in science. His proposition is based on the idea that there is something good in every approach, irrespective of how absurd. There exist a vast, and apparently seldom utilised, field of paradigmatic possibilities between the traditional monistic point of view and Feyerabend's "anything goes". There is thus additional advantages to be reached by allowing an individual researcher the mental freedom of pursuing several paradigms, by not only allow the problem and the methodology to be variables in the pursuit of knowledge, but also the paradigm. This description of the three nodes of research as containing three variables instead of two run the risk of being viewed as naive, for instance

12 Drawing on the analogy of Porter's (1980) discussion of the risks of simultaneous pursue of generic strategies that I in a "Popperian" (Popper, 1959) spirit falsify in article VI of the thesis.


by Arbnor and Bjerke (1994). The reasoning in this chapter can be interpreted as narrated in order to lull the reader into a false sense of security concerning the superbness of the idea of one researchers simultaneous pursuit of several paradigms. I will therefore throw in a mental reservation, in observing that shifting the problem might be hard, and shifting methodology might be even harder, but shifting paradigm is undoubtedly the hardest. A cautious suggestion is to start by trying new problems, then try new methodologies and after that give the change of paradigm a fair chance. The proposition of paradigmatic flexibility is fascinating due to its inherent uncertainties. Uncertainties can be a double edged sword, containing threats as well as possibilities. In adopting uncertainty as one of the fundamental concepts of reality, will impose several serious implications for the conscientious researcher. For one thing it removes the shield that has traditionally protected scientists from the dilemma of choice (Morgan, 1983), especially the positivists. Evidence of uncertainty is plentiful, also from the fields of natural sciences such as Heisenberg's (1958) indeterminancy principle, suggesting that on a sub-atomic level it is not possible to attain knowledge of a particles position and momentum simultaneously. Both can be known approximately, but increasing the certainty of one will make the other more uncertain. Therefore what is determined, what becomes the result of the research, is dependent on the researcher's choice of perspective. Similarly, the duality of light, behaving as a particle or as a wave, depending on the observers point of view, is elaborated on by Bohr (1958). This led him to propose the principle of complementarity, both aspects of light are necessary to understand the nature of light. It is further useless to ask which the two aspects is the real nature of light. Light, he continued does, from this perspective, not possess inherent attributes, but are attributes of the interaction between the observer and the observed. Thus also in natural sciences there is an interaction, a special relation, between the object and the subject. Again the result of research is dependent on the researchers choices and assumptions. In social sciences these principles appear almost self-evident and intuitively appealing. Bearing Bohr's thoughts in mind, the suggestion of paradigmatic complementarity might appear more attractive. Especially when relating Bohr's notion of interaction between subject and object with the respective focus of Arbnor's and Bjerke's (1994) three paradigms: objects, relations and subjects. In Bohr's (1958) interpretation, not as exclusive paradigms, but combined in the description of the principle of complementarity. 5.1.3 Methodology and the thesis There appear to exist a widely spread methodological prejudice at the extension of the paradigmatic rigidity. According to methodologists (e.g. Arbnor and Bjerke, 1994; Morgan 1983) several researchers have specialised in using one and the same methodology. This conformity can extend over different levels of research from the individual level, to the research group and even for an entire research area. However few research areas seem to be best researched by approaching the problem with tunnel vision. By specialising in using one single method the researcher can be expected to master this method. However, if loosing the key in


the street at night, looking directly under the streetlight will facilitate the search, but you will probably never find the key... Therefore the purpose should determine the method, not the other way around. This principle has guided my research. It has become increasingly clear that research methods can not, a priori, be regarded as a neutral or atheorethical tools. (Brewer and Hunter, 1989). This is true for natural sciences, such as quantum mechanics, as well as for social sciences. For instance Walton (1966) has shown that the differences in theories of community power between political scientists and sociologists might well be accounted for by each groups isolationistic pursuit of research with different methodologies. There are probably systematic links between paradigm and favoured research methods. Ritzer (1980) has proposed such linkage for the field of sociology. At the extension this will imply that problems that are either theoretically or methodologically incompatible will be ignored. This can seriously hamper the progression of the research if important aspects of the field are not researched. One way of coping with this problem is to implement a multi methodological approach. The multi methodological approach rests on the idea that methodology should not be treated as a choice between mutually exclusive methods. Individual methods might be defective by favouring certain results, but the deficiencies of different methods are not identical. The fundamental strategy of a multi methodological approach is thus to "attack a research problem with an arsenal of methods that have nonoverlapping weaknesses in addition to their complementary strengths" (Brewer and Hunter, 1989, p. 17). Brewer and Hunter (1989, pp. 48-54) propose a number of advantages of the multi methodological approach for theoretically oriented research. Firstly, theories do not respect conventional methodological boundaries, nor should they be required to do so. Problems do not either respect conventional boarders between research areas, which has been identified by the advocates of inter-disciplinary research. However inter-disciplinarity carry a broader potential than merely sharing the problem. The possibility of sharing metaphors and drawing on analogies between different fields of research has proved a powerful tool in advancing research. For instance models from biology or physics can, and have frequently been be used in organisational theory. Furthermore, theoretical problems of importance are exposed to repeated investigations. Theoretical inquiry of a phenomenon often involve methodological replication, with an added twist at each consecutive study. However, the very fact that a method has been used on a problem repeatedly may be a good reason to apply a novel approach in the next study. As a matter of fact, a given field of research might underachieve if all of its research is being conducted within a narrow methodological frame (Dunn et al, 1993). A field of research goes through a life cycle (drawing on a biological analogy), in which different methods are more appropriate at different stages (Brewer and


Hunter, 1989). It is usually suggested that, as a discipline matures, the problems get more sharply defined leading to a natural progression from explorative studies towards more verificational studies. However, as noted by the authors, inquiry is not as unidirectional as this model implies, neither in disciplines nor in individual projects. Corroboration is an other possible advantage of the multi methodological approach. It can be needed when the ideally appropriate method is not feasible, for example, on the account of different constraints, such as of monetary resources or time restrains. Corroboration can be applied on a methodological level, in a field of research. By embracing a balance of methodological types, then "a higher form of intellectual honesty and content richness can be said to occur ... It provides more scrutiny so that dogma is continually identified and brought into question" (Dunn et al, 1993, p. 123). Finally, employing different methods helps to guard against, and correct for, inherent methodological biases. A field of research may, after some time, become self reproductive in the sense that its members have, by using a tunnel vision approach, advanced the field in a specific direction within a minor group of paradigmatic and methodological regimented researchers. By referring to prior works of other members in the group each new article is justified, on the account of prior articles in the field making research within the area a closed loop. However from an external point of view the group may be drifting away into meaninglessness, often disconnected from empirical realities. In some instances this might prove valuable, but probably not as a rule. An approach in such a field of research from a fresh methodological perspective might prove a powerful tool in reconnecting the drifters13 to a more common reality. A good example of an research area fulfilling some of the criteria of a drifter is operations management. Meredith et al (1989), found that the majority of research in operations management utilised methods that could be classified into one particular methodological paradigm and that the field is "... overwhelmingly artificial in nature" (ibid., p. 317). "... operations has been viewed with a deductive, simplistic slant and is most commonly manifested in techniques such as analytical modelling" (Dunn et al, 1993, p. 123). "... a researcher may first enter into an investigation with pre-conceived logical assumptions about relationships among perceived phenomena ... This type of methodology has the potential to be somewhat removed from the real world because theory exist a priori and its measurement is operationalized through mathematic logic and structure. ... If real world data does not support the premises of the analytical model, then the data, and not the model or its theory, are often rejected" (Dunn

13 The term "drifter" draws on an analogy from the computer game "Life" from the 1960's, which applies a few distinct rules to the inhabitants of its world, that consists of an infinite two dimensional space of squares or inhabitants that are either light, on the screen, or dark (alive or dead). After each generation the state of a square, alive or dead, is dependent on how many neighbouring squares were alive in the prior generation. After a number of generations certain constellations of eternal life appear. One is the drifter that revolves around its own axis on its individual way to infinity.


et al, 1993, p. 126). This may explain why testing against real data seldom occurs in this field of research. Mentzer and Kahn (1993) made a review of all articles in Journal of Business Logistics, 1980 to 1993 and found that "only 4 percent of JBL published literature has involved hypothesis testing procedures. Of these, few have provided information on validity and reliability" (Mentzer and Kahn, 1993, p. 150). For the field of operations management, novel methodological approaches might prove to be a vitamin injection, revitalising and reconnecting the field to an externally recognised reality. Multi methodological approaches in a research area can alleviate several of the several aspects of the drifter syndrome and thus has a multitude of advantages over the traditional tunnel vision approach. But also potential disadvantages exist. Multi methodological approaches can lead to "incoherence, confusion and fruitless controversy" (Brewer and Hunter, 1989, p. 23). Therefore some researchers (Blalock, 1978, p. 22) call for a greater consensus "in terminology and research operation". I propose that a consensus on terminology is advantageous but quite different from, actually not related at all to, consensus on research operations. I further propose that the advantages of applying a multi methodological approach to a field of research probably greatly outweighs potential disadvantages, especially over the long run. Integration, by applying a multi methodological approach, can manifest itself at different levels. At the upper level, the area of research, as discussed above. However also within the research group. Interaction between colleagues - even colleagues of quite different methodological persuasions - can be a powerful integrative force" (Brewer and Hunter, 1989, p. 24). Finally, at the individual level, integration is possible. This is perhaps the least used possibility of integration, while simultaneously providing the greatest opportunities. Some researchers even seem to be afraid of the idea. "We are not advocating that each researcher take on all methods in their research portfolio. ... What we advocate is that each researcher appreciate and encourage the diversity of various methods dedicated to the growth and understanding of logistics [the research area] as a whole" (Dunn et al, 1993, p. 124). I boldly (or foolishly depending on perspective) propose that this is exactly what should be advocated, that each researcher extends his or her research portfolio beyond the tunnel vision approach and integrate different research methods at the individual level. This could pave the way for the much coveted understanding of other methodological approaches than each researcher's own pet method. Having tried different approaches on an individual basis could probably facilitate the understanding within a research group, as well as within the research area as a whole. The methodological proposition is thus:

Implementing a multi methodological research approach in a field of research may be most advantageous by using a bottom-up strategy, starting with broadening the individual researcher's methodological portfolio.

A major benefit of a multi methodological approach may be that individual researchers begin the task of synthesise their individual findings. "Only much


more rarely is integration undertaken in the design of individual pieces of research. For the most part, we each seem to do and to publish our work in the faith and hope that somebody else will create a synthesis. ... The writing of theoretical and methodological synthezisers are among the most widely read and frequently cited works in the social science literature" (Brewer and Hunter, 1989, p. 24). Methodological complementarity has already achieved wide recognition as a means of furthering the knowledge generating capabilities of research, although the possibilities at the individual level have not yet been fully appreciated. The guiding principle of the methodology applied in the thesis is to use the most appropriate method given the problem at hand. In concluding this chapter I will discuss some of the methodological problems of the articles of the thesis from an inter-article perspective. The methodology of each article is discussed in each appended article. The methodological and paradigmatic dimension of the taxonomies of Figure 9 have already been discussed. Each article is positioned in the most appropriate place in the matrix. It is not possible to identify an article with a certain paradigm and exclude any connection to the others in an unobjectionable way. It is also an oversimplification to identify each articles as belonging solely to one methodology. However the graphic presentation can be useful as a starting point for the discussion.14

14 There appear to be a narrow-minded conception concerning the output of a doctoral program. The thesis is regarded as the output while the work performed by the doctoral student to produce it is the input. It is however quite possible to reverse the logic, the thesis can be regarded as the vehicle for the authors transformation into a certified researcher, though this perspective is usually parsimoniously treated in dissertations. With this latter view, the thesis may not be the primary result of the doctoral study. Instead the process which the author has experienced becomes the primary result. The dualistic relation between the researcher and the thesis as being simultaneously input and output of the research process is mirrored in the relation between knowledge and learning. The researcher learns by creating knowledge and creates knowledge by learning. Both knowledge and learning should be acknowledged as important results of doctoral program. I will briefly contemplate both aspects of a doctoral program.


Figure 9. Main paradigmatic affiliation and methodological approach for the articles of the dissertation. The thesis reports findings from four different fields of research (Figure 5), however the difference between the areas are not greater than they can also be arranged under one area of research, industrial management. Based on the former footnote, the aim of articles in a thesis can be of dual nature. The aim can be to improve the field of research and/or to improve the researcher's research capabilities. Viewed from this perspective the multi methodological approach has been used with the primary aim of improving the field of research in the first five articles, which have manufacturing flexibility as the lowest common denominator. The remaining three articles of the thesis, each address a different fields of research, related to the main theme: strategic manufacturing flexibility. These articles also expand the methodological dimension of the thesis. The lowest common denominator for these three articles is the researcher. An important aim of the multi methodological approach for these articles can thus be viewed as improving the researcher's capabilities. How should research be conducted? "In conceptualization, the assumptions upon which the problem definition is based are defined and if these assumptions are incorrect, then the problem that is solved may indeed be the wrong problem"(Lyles, 1981, p. 62). Therefore great caution is needed in making the assumptions. If the assumptions are to specific, the solution arrived at, the solution might be disconnected from, and thus irrelevant to, the empirical reality. On the other hand, by using to general assumptions, in order to reach general conclusions that apply to a vast number of situation, the collecting and

PARADIGMME

THOD

OLOG

Y casestudy

survey

meta-analysis

positivistic system interpretive

reductionism------holismgeneral--------particular

objective------------------------subjective

nom

othe

ticid

eo-

grap

hic

empi

rical

conc

eptu

al

theo

ryge

nera

tion

theo

ryve

rific

atio

n

dedu

ctiv

ein

duct

ive

VIV

I II

VI

VIII

III

VII

quantitative

qualitative

conceptualstudy


subsequent analysis of data may demand unreasonably much resources. Deciding upon assumptions is thus a matter of economising. The area of manufacturing flexibility was entered with some preconceptions and other assumtions were formed during the research. It is difficult to decide the exact point in time when different assumptions broke the surface of consciousness. However from an ex post perspective, it can be concluded that the research has been guided by the following basic assumptions concerning manufacturing flexibility: I Flexibility is a concept of several dimensions. II The literature has addressed the multi dimensionality primarily by

theoretically deduced taxonomies of flexibility. III In order to utilise the potentials of flexibility in manufacturing, a common

understanding of the concept among managers, can be advantageous. IV A common understanding of the means and ends of flexibility is also

advantageous. V Traditional capital budgeting techniques do not assess flexibility of

investments to its full potential, thus underrating flexible investments. VI Managers prefer simple methods of investment assessments to

theoretically correct ones. VII A theoretically correct model can be advantageous, to provide a means of

assessing the effects of the simplifications. From these preconceptions, the area of manufacturing strategy has been penetrated. Based on assumptions I, II, III, and IV a number of studies have been performed, reported in several articles, of which three are included in the thesis. These articles all built on several of the first four assumptions. Each of the studies use mainly one of the theory development principles of figure 8: induction, deduction or verification. Article I and II are aimed at providing a structural framework to relate flexibility to the manufacturing strategy of a company. The framework is build on an organisational hierarchy and a generic system of flexibility dimensions. The results of these two articles are primarily theoretically deduced, however the possibility of an empirical connection is proposed in article 2, in which a, hypothetical but realistic, example of the use of the framework is worked through. A number of mini-case studies were performed prior to the development of the flexibility structure thus implicitly providing at least a mental empirical grounding of the findings. Some of the mini-case studies are reported in Nilsson and Nordahl (1992) and account for the inductive theory development on managers' perception of flexibility in manufacturing. The next logical step is a phase aimed at verifying the induced and deduced results of these other studies, for instance, by means of a survey. Article 3 reports the preliminary set-up for this verification.


Based on assumptions V and VI a method for evaluating flexibility of the machinery in an investment was developed in article 4. The method is more aimed at usability than theoretical correctness. Base on assumption VII, a more stringent analysis and a theoretically more correct version based on continuous time is developed in article 5 . This less user-friendly version of the method is compared to the simplified version of article 4, and for all practical purposes the differences are negligible. However the continuous time version can be advantageous for further theoretical deductive research by analytical modelling. The articles 6, 7 and 8 do not follow a stringent contractive order as the first five articles. On the contrary, they are expansive. Each of the three last articles opens up a new field of research (at least to the author). Not withholding the independent merits of each articles, a major benefit of these three articles, from a thesis perspective is to expose the author to two more methodologies; case-study and meta-analysis thus improving my research capabilities by expanding the methodological portfolio. One case study is performed, the case of Scania and the heavy truck industry, which is used two ways, with different theoretical backing. Firstly, article 6, in the field of manufacturing strategy, the case of Scania is used to falsify Porter's (1980) notion of stuck in the middle and presenting some methodological propositions for studying strategy from a grounded approach, and secondly, article 8, to make first inquiries into the field of strategic manufacturing flexibility. Article 7 is used to probe the third node of entry into strategic manufacturing flexibility, that is strategic flexibility (see Figure 5). This field of research is not connected to manufacturing, but is flexibility at a higher level. For instance strategic alliances is one aspect of strategic flexibility brought forth in the literature. This literature is therefore scrutinised in article 7 based on the over optimistic picture of strategic alliances conveyed by the literature, contrasting the more modest success rates of strategic alliances reported in broad surveys. The connection between the last three articles is the concept of strategic alliance. Scania is used as an example to indicate that strategic alliances is a double edged sword, carrying threats as well as possibilities. In the case of Scania it is argued that the threats outweighs the possible advantages.

Methodological Reflections–Extended Version

Documents