mode 2 management research

Introduction

Between 1995 and 1998, the British Academy ofManagement (BAM) embarked upon a wide-spread debate about the increasing distance ofresearch from its user base and a concomitantdecline in its ability to influence policy develop-ment and practice.

Much of the discussion focused on ontologicalissues. What makes management research uniqueand different from associated disciplines such aseconomics, sociology, social psychology, psychologyor anthropology? What makes for good qualitymanagement research? How might one recognizethe boundaries of management research? Howmight we best recover the support and involve-ment of our user base? These themes have sincebeen taken up by the US Academy (see Huff,2000) and have been the focus of further debatein this journal (see Hodgkinson, 2001).

The initial discussions were led by the BAMResearch Policy Committee and subsequentlydeveloped in Council, culminating in a paper byDavid Tranfield and Kenneth Starkey. The paper

aimed to stimulate further debate on the natureof management research and highlight the potentialof so-called mode-2 practice as an avenue for fruit-ful development (Tranfield and Starkey, 1998).

This paper offers a response to Tranfield andStarkey’s discussion piece (1998) and contributesto the ongoing debate in two ways. First, webelieve that the social sciences have an estab-lished tradition of conducting research in mode 2,albeit expressed in different terms. Indeed, someforms of management research might be said tobe at the leading edge of a broader movementtowards increased levels of mode 2 knowledgeproduction. Whilst there is undoubtedly a wealthof material on approaches such as action research,cooperative inquiry and clinical method, a map-ping of these practices onto the methodologicalterritory of mode 2, as laid out by Gibbons et al.(1994), remains noticeably absent from theliterature.

The other gap in the literature is the lack of empirical accounts specifically related to thedetailed features of mode 2 proposed by Gibbonset al. (1994). The original work (predominantly

British Journal of Management, Vol. 13, 189–207 (2002)

© 2002 British Academy of Management

Mode 2 Management Research

D. MacLean, R. MacIntosh and S. Grant*Department of Business and Management, University of Glasgow, West Quadrangle-Gilbert Scott Building,

Glasgow G12 8QQ and *Scottish Health Advisory Science, 8–10 Hillside Crescent, Edinburgh EH7 5EA, UKCorresponding author email: [email protected]

The terminology of mode 2 knowledge production has become increasingly prominentin discussions on the nature and purpose of management research. This paper attemptsto move our understanding of this terminology forward, by providing a detailed ex-position of the five features of mode 2 knowledge production in the context of manage-ment research and by offering an empirical account of a research project conducted inmode 2. The paper relates the established problem-solving management researchtraditions of action research, cooperative inquiry, grounded theory and clinical methodto the conceptual territory of mode 2. It then considers a specific form of knowledgeproduction where all five features of mode 2 appear simultaneously. The paper demon-strates how the terminology of mode 2 might provide a useful basis for dialogue betweenmanagement researchers from different methodological traditions. Moreover, the paperconcludes that the specific form of mode 2 where all five features are present (calledhere 5mode2) does differ, in both its conduct and the character of its output, from anyof the established approaches considered in this paper.

01_Mac 26/11/02 1:05 pm 89

rooted in developments in the natural sciencesand technological research) offers little to justifythe authors’ choice of descriptive features (asopposed to some other set of features) in theparticular context of the social sciences. Indeed, ifthe features of mode 2 presented by Gibbons et al.are thought of as a set of ingredients, the variousrecipes which might be created from differentcombinations of these ingredients are not dis-cussed – although there is passing reference to aform of research where all the features of mode 2‘appear together’, signalling possible distinctive-ness of this particular form (Gibbons et al., 1994,p. 8).

This paper therefore has two aims. First, itattempts to relate established problem-solvingmanagement research traditions to the conceptualterritory of mode 2. The second aim of the paperis to consider the specific form of knowledgeproduction where all five features of mode 2appear simultaneously. By reviewing an empiricalexample of this form of knowledge production,the paper aims to give readers a flavour of oneparticular form of mode 2 management research,and thus assist them in relating their ownresearch practices to the concept of mode 2. Thisin turn should increase the scope for partici-pation in the debate on the development ofpractice-oriented research, whilst offering in-sights into the practice of mode 2 managementresearch on the other.

Management research – the ongoingdebate

Management research is in the throes of a healthydebate on both its ontological status and theappropriateness of various research processes.This debate has been prompted partly by arecurring concern that much of the managementresearch appearing in top-rated academic journalsis of little relevance to most practitioners (seeSchein, 1987; Gopinath and Hoffman, 1995;Starkey and Madan, 2001). More recently, thedebate on research process has begun to focus onthe differences between two distinct researchapproaches, which have been labelled mode 1 andmode 2 knowledge production (Gibbons et al.,1994).

Pettigrew (1995, 2001) draws attention to theneed for research to clear ‘double hurdles’ such as

simultaneously delivering practitioner relevanceand scholarly excellence. He discusses the fea-tures of mode 2 knowledge production in relationto these double hurdles. Tranfield and Starkey(1998) use Becher’s taxonomy to describe man-agement research as ‘soft, applied, divergent and rural’. They conclude that mode 2 reflects theontological status of management research morefaithfully than does mode 1. Their discussion of management research as mode 2 knowledgeproduction culminates in the authors advancing aseries of propositions and speculations rangingfrom themed centres of excellence, to calls fordialogue with (rather than adherence to) existingdominant (American) standards, and the need forresearchers to avoid becoming trapped in eitherpractice or academia in which they risk the perilsof ‘epistemic drift’ or ‘academic fundamentalism’respectively (p. 353).

In the USA, Anne Huff used her presidentialaddress to the Academy of Management (AoM)annual conference to address issues of practitionerrelevance, drawing attention to ‘mode 1.5’ as aterm which she used to denote the combination ofdisciplinary scientific rigour normally associatedwith mode 1 and the practically valued problem-solving orientation of mode 2 (Huff, 2000).Further reflection led to a call for ‘Mode 3’ know-ledge production (Huff and Huff, 2001). Thishighlights a growing interest in the nature androle of practice-oriented research in the US, asevidenced by AoM’s research methods division’sspecial theme on ‘practice-grounded research’ atthe 2001 and 2002 conferences.

As stated in the introduction, the academicdebate has thus far been conducted largely incommittee or meeting rooms. The papers citedabove attempted to open the debate to a broaderaudience but responses have been few and farbetween in the literature. We feel that contributoryfactors in this regard may be, on the one hand, alack of empirical accounts which are explicitlymode 2 in orientation and, on the other hand, a general confusion around the use of the termmode 2 and its relationship with other terms suchas action research. This paper thus attempts to deal with both these factors, starting with thelatter. To do this, we will draw on the original workof Gibbons et al., first presenting the concept ofmode 2 in relation to mode 1 and then detailingthe distinctive features of knowledge productionin mode 2.

190 D. MacLean, R. MacIntosh and S. Grant

01_Mac 26/11/02 1:05 pm Page 190

Mode 2 knowledge production

In the preface to The New Production of Know-ledge (Gibbons et al., 1994), Michael Gibbonsdraws the reader’s attention toward a new form ofknowledge production (mode 2) which, althoughoriginally an outgrowth from its traditionalcounterpart (mode 1), is becoming increasinglydistinctive:

‘our view is that while Mode 2 may not bereplacing Mode 1, Mode 2 is different from Mode1 – in nearly every respect . . . it is not beinginstitutionalised primarily within university struc-tures . . . (it) involves the close interaction ofmany actors throughout the process of knowledgeproduction . . . (it) makes use of a wider range ofcriteria in judging quality control. Overall, theprocess of knowledge production is becomingmore reflexive and affects at the deepest levelswhat shall count as “good science”.’ (Gibbons et al., 1994, p. vii)

Many of these themes are further refined as thekey features of mode 2, which are contrasted withthe features of mode 1:

‘Mode 1 problems are set and solved in a contextgoverned by the, largely academic, interests of aspecific community. By contrast Mode 2 is carriedout in the context of application. Mode 1 isdisciplinary while Mode 2 is transdisciplinary.Mode 1 is characterised by homogeneity, Mode 2by heterogeneity. Organisationally, Mode 1 isheirarchical and tends to preserve its form, whileMode 2 is more heterarchical and transient. In comparison with Mode 1, Mode 2 is sociallyaccountable and reflexive.’ (p. 3)

The authors then go on to discuss in some detailwhat they view as the key features of mode 2knowledge production. These are depicted insummary form in Figure 1 and discussed in moredetail below.

Knowledge produced in the context of application

This emphasizes the problem-solving nature ofmode 2, in that knowledge production is organ-ized around a particular application as opposed to the codes of practice of a particular discipline.It is perhaps this feature of mode 2 which mostexplicitly and directly links it to the perceived prob-lems of user-distance and practitioner-irrelevanceof management research mentioned in the intro-duction. The imperative to be practically useful isthe dominant consideration in framing researchaims, questions and practices which, given thedynamic nature of most organizations, ‘are pro-duced under an aspect of continuous negotiation’(Gibbons et al., 1994, p. 3).

Transdisciplinarity

Transdisciplinary problem solving involves theintegration of different skills in a ‘framework of action’ (Gibbons et al., 1994, p. 4) in whichinterwoven empirical elements and theoreticalconsensus arise and accumulate throughout theproject as practical solutions and theory buildingwhich cannot be reduced to disciplinary parts.These are then diffused through the subsequentactions of those who participate in the project.This gives rise to an unpredictable dynamic whichis difficult to direct, characterized as ‘problemsolving capability on the move’ (p. 5).

Heterogeneity and organizational diversity

Transitory, multi-faceted problems requiretransient heterogeneous teams whose memberscome and go as the situation organically unfolds.This is enabled by technologies that allow teamsto flourish between and beyond the ties of anyone institution on the one hand, and by a moresocially distributed research capability on the other.

Social accountability and reflexivity

Greater levels of communication and transparency;increased sensitivity to issues of governance andinclusion; and growing public interest in the pur-pose, nature and outcomes of research have led to much greater diversity of interests and involve-ment. These factors, along with the transdisciplin-ary team make-up, require participants continuallyto approach research from the standpoint of theother participants as part of the ongoing process

Mode 2 Management Research 191

Figure 1. The five features of mode 2 knowledge production

Knowledge produced in the context of application

Transdisciplinarity

Heterogeneity and organizational diversity

Social accountability and reflexivity

Diverse range of quality controls

01_Mac 26/11/02 1:05 pm Page 191

of negotiation. This in turn leads to a much morereflexive form of research with a deeper under-standing of the research process itself.

Diverse range of quality controls

In mode 1, the quality of knowledge produced isjudged from the standpoint of the discipline, itssenior luminaries and peer evaluation processes,whereas, in mode 2, quality controls have toreflect the concerns of a substantially broadercommunity of interest. This might include forexample, the practicality of any proposed solutionto the problem under investigation. Whilst it is conceivable that possible compromises intro-duced as a means of satisfying a broad range ofstakeholders with different interests (e.g. time-scales) might undermine the standard of suchwork, Gibbons et al. point out that, ‘it does notfollow that because a wider range of expertise isbrought to bear on a problem, that it will necessarilybe of a lower quality. It is of a more composite,multidimensional kind’ (p. 8).

Mode 2 and management research

Translated into the context of managementresearch, the criteria laid out by Gibbons et al.paint a picture of research as a high involvement,problem solving, theory building dynamic. Suchdescriptions of management research are notnew. Indeed, we have already hypothesized that agreat deal of research, using a number of method-ologies that are well established in the field ofmanagement, can be regarded as being conductedin mode 2. In this section we will examine anumber of these approaches and consider them inrelation to the features of mode 2 identified in theprevious section.

A number of the features attributed to thebroad term ‘mode 2’ might equally be applied to an array of established approaches, e.g. actionresearch, action science, participatory research,action learning, grounded theory, clinical method,cooperative inquiry, etc. An exhaustive mappingand comparison of all such approaches is beyondthe scope of any one paper.

Faced with the need to narrow our focussomewhat, we first considered Crotty’s (1998)work on the foundations of social research, whichoffers a useful schema for classifying research. Heargues that any given piece of research can bedescribed in terms of its epistemological position,

the theoretical perspective taken, the methodologyemployed and the specific methods chosen. Hiscareful analysis helps to ensure an ‘apples withapples’ comparison and, using this schema, wehave chosen to assess different methodologies in relation to the features of mode 2. In Crotty’sterms, a methodology represents a frameworkwithin which particular methods might be deployed,e.g. ethnography (a methodology in his terms),might involve participant observation or con-versation analysis (methods in his terms) (Crotty,1998, p. 4).

Having identified the category we are inter-ested in for our comparison (i.e. methodologies,in Crotty’s terminology), we are still faced withthe problem of choosing a limited number from a wide range of different methodologies. Toillustrate this problem, consider the case of actionresearch which can be traced back to Lewin’swork in the 1940s (Lewin, 1946). Action researchhas been characterized as research resulting frominvolvement with an organization over a matterof genuine concern, where there is the intentionto take action on the basis of the intervention(Eden and Huxham, 1996). However, the term isregularly used with a variety of different meanings,to the extent that the phrase itself has ‘lost some ofits original weight’ (Reason and Bradbury, 2001, p. xxiv). Indeed some argue that it may be morehelpful to consider ‘action-oriented research’ as anumbrella term (Park, 1999, p. 142) in much thesame way as that in which mode 2 might be used.

Given that entire special issues of this journalhave been devoted to delineating between variantsof action research (see Human Relations, 1993and Management Learning, 1999), we haveelected to focus here on two variants of actionresearch. The first is a relatively straightforwardapproach which still holds traditional views on theobjectivity of the research process and envisagesa clear division of labour between academic and practitioner, as characterized by Eden andHuxham (1996) and the second, an explicitlygroup-based approach called cooperative inquiry,where all those involved act as ‘co-researchersand co-subjects’ (Heron, 1996).

Our third choice is grounded theory, amethodology which its originators describe as ‘the discovery of theory from data’ (Glaser andStrauss, 1967, p. 1). The two key features ofgrounded theory approaches are theoreticalsampling (i.e. data collection is guided by the


01_Mac 26/11/02 1:05 pm Page 192

emerging theory) and constant comparison (i.e.continuously flipping back and forth betweendata coding and data analysis). Partington’s recentcontribution sets out grounded theory in moredetail and begins to build some links to the con-cepts of mode 2, highlighting its problem solvingcapabilities (Partington, 2000).

Our final choice is Schein’s clinical method(1987), which he describes using the analogy ofmedical practice, where analysis and publicationof selected cases helps to build knowledge forfellow practitioners. The research is initiated bythe practitioner and focused on an issue identifiedas important by them, and the researcher in thesesettings is called in because of his or her ‘helpingskills’ (Schein, 2001).

The contrast between mode 2 and more trad-itional mode 1 forms of knowledge production ismarked and has already been dealt with both hereand elsewhere. However, the relationship betweenthe features of mode 2 and the four method-ologies we have selected is less clear. We haveattempted to summarize these distinctions inTables 1a–1e in the appendix.

On the basis of the comparisons shown inTables 1a–1e, we would argue that research con-ducted using each of the four selected methodologiescould be regarded as producing knowledge ‘inmode 2’, reinforcing our view that managementhas a long tradition of conducting research inmode 2 and is, in many ways, further advancedthan some of its natural science counterparts inthis regard.

We will now move on to consider an empiricalaccount of a research project conducted in mode 2.In this particular project, the research was under-taken with all five features of mode 2 as simul-taneously coexisting aspects of the process. Indeed,Gibbons et al. (1994) point out that the features ofmode 2 ‘do, when they appear together, have acoherence which gives recognisable cognitive andorganisational stability to the mode of production’(p. 8). The project which follows represents ourattempt to offer an empirically driven investigationof this claim.

An example of mode 2 managementresearch

In this section of the paper we aim, as stated inthe introduction, to give readers a flavour of

one particular form of mode 2 in practice andthus assist them in relating their own researchpractices to the concept of mode 2 as defined byGibbons et al.

Moreover, by considering a project in which allfive features of mode 2 simultaneously apply, wehope to address the question implicit in theprevious section, as to whether the qualities ofmode 2 practices are primarily characterized bythe specific cocktail of the five features employed.We will do this by examining whether there is any-thing particularly distinctive about the limitingcase where all five features occur simultaneouslywithin the confines of a single project.

We begin by providing some background to themode 2 project concerned. The project spanned aperiod of over 12 months, with one- or two-daymeetings at least once per month and manyadditional conversations, telephone and emailexchanges in between. The account presentedhere is therefore selective, seeking to highlightfeatures of the project which may help to castsome light on the characteristics of the researchprocess. The account also downplays theresearch content of the project in order to focuson the implications for research process. Appro-priate pointers are given for those interested in thespecific content of the research.

The project was initiated shortly after aresearch seminar on a prior project on complexitytheory and corporate transformation. During the seminar, questions concerning the roles ofreflexivity, intuition and emotion in influencingemergent outcomes of corporate transformationprojects were raised, giving rise to relatedresearch questions and the desire to investigatethem empirically.

Subsequently, at the request of the chiefexecutive of the Scottish Health Advisory Service(SHAS), we became involved in a complexity-based transformation project which allowed us to pursue our research interests whilst aiming tohelp SHAS develop a more distinct identity andnew, more effective ways of working in existingand new areas. This desire was influenced in nosmall part by the general upheaval in the NationalHealth Service (NHS) and the recent establish-ment of the Scottish Parliament.

SHAS audits the provision of healthcare toparticular care groups in Scotland (mental health,the elderly and those with learning disabilities).The organization consists of a small core team,


01_Mac 26/11/02 1:05 pm Page 193

and a much larger reviewer network of severalhundred people. Any given audit is led by anadvisor from the SHAS core team and a group of6–8 volunteers drawn from the reviewer network.Each audit starts with an orientation day for theteam followed by a one-week visit to the par-ticular healthcare unit being assessed. The reviewteam then disperses at the end of the week butparticipates in the writing of a report which isthen fed back to the relevant health board orNHS trust. Follow-up visits by members of theSHAS core team may then be arranged if required.

Whilst the SHAS inspection and follow-upvisits were generally seen to be highly successful,a burgeoning workload, environmental turbulenceand numerous opportunities to engage in new areasof work were stretching the somewhat informaland person-dependent management processeswithin SHAS to the limit. In particular, the chiefexecutive felt that there was a clear need totransform working practices so that key decisionsand initiatives were less critically dependent onher direct involvement.

The SHAS core team felt that this required thedevelopment of a collective approach that wouldharness the distinctive strengths of individualmembers, but which could operate professionallyregardless of which individuals were involved inspecific projects. There was also a unanimous feel-ing that, in order to capitalize on opportunities to grow and extend influence in existing and newareas of work, SHAS had to develop a morecoherent external identity in tandem with its newinternal processes.

The project was thus concerned with the stra-tegic transformation of SHAS both internally interms of its management processes, and externallyin terms of its desire to leverage core capabilities,enter new areas and increase the breadth anddepth of its service delivery. The project was also intended to help sharpen the organization’sidentity and image. The interest of the chiefexecutive and the rest of the core team in tryingto apply concepts from complexity theory toorganization and management, enabled a genuineconvergence of interests between the academicresearchers and the SHAS staff – resulting in aconstructive and mutually sympathetic approachthroughout.

A central feature of the project was a series ofwhat can best be described as qualitative changesin the nature of the project itself. These changes

corresponded to changes within SHAS and tendedto be concentrated into a series of unplanned(and to some extent unexpected) episodes withincremental development and refinement occurringbetween episodes. Such episodes can be thoughtof as periods during which a system suspends itsroutine structures allowing for reflection on, andchange of, these structures (Hendry and Seidl,2002). This account will focus on these episodessince they are the most revealing aspects of theproject in terms of the research process.

Of particular note was an event which occurredin the very early stages and which influenced boththeory and practice within the project. The groupwas discussing a recurring and problematic pat-tern. Whenever a key decision or breakthroughwas about to be made, progress was suspended bythe intervention of some emergency or other. Thisemergency inevitably dragged the chief executiveaway from the rest of the core team and progresswas effectively paralysed by her absence (for thereasons of dependence mentioned earlier in thepaper), leaving the team feeling frustrated andunable to proceed.

As discussion of this issue unfolded at one ofthe first project meetings, just such an emergencyarose and there was debate as to whether or notthe meeting could be taken any further. Indeedthe survival or otherwise of the whole researchproject was raised in a somewhat panic-strickenand fraught response. The concept of emergence1

was introduced to the discussion, and the groupdecided to use the opportunity to deal with therecurrent pattern there and then. The emotionalatmosphere of the room turned from one ofdespondency and frustration to one of optimismand even celebration in which the emergency wasembraced as an opportunity to deal ‘live’ withwhat was seen as the key issue facing the organ-ization at that point in time.

The team’s response was to articulate ‘groundrules’ concerning the way members should inter-act with each other. These ground rules allowedthose present to respond, knowing that theywould have the support of those who were absent.There was a general view that this agreement onhow team members thought they should interact


1 Goldstein (1999) offers an overview of the concept of emergence and further details of our own work onemergence in the SHAS case study can be found inMacLean et al. (2001).

01_Mac 26/11/02 1:05 pm Page 194

was made meaningful by the crisis. From thatpoint onwards the ground rules were referred to asthe Cromlix principles and they governed intra-team interaction, (in the parlance of complexitytheory, these might be called order-generatingrules).2 During the project, they were frequentlyreferred to, and were invoked by participantswhen progress during meetings seemed laborious.

In this instance, a recurring and problematicpattern was identified and a response to thepattern was negotiated. Ironically, and perhapssignificantly from a psychoanalytic perspective,this was followed by a rapid dissipation of the crisis,and the threatened departure of chief executivenever came to pass. Discussions then returned tothe concept of emergence and the way in whichthe incident concerned seemed to transcend thefamiliar management dichotomy of planning andacting. There was general agreement that themeaningfulness of the incident was somehow orother tied up with its unplannedness and, more-over, that the accompanying emotional atmos-phere was somehow responsible for the team’spositive response. Developing a new pattern ofinteraction required a simultaneous appreciationof the experience, and a distancing of oneselffrom it, in order to appreciate the patterns at play.

From our point of view, the project had alreadybegun to deliver insights into the roles of re-flexivity, intuition and emotion in transformationprojects. Reflexive processes both identified thecrisis as the recurrence of a pattern, and in thiscase helped team members themselves to trans-form practice from a somewhat indisciplined,individual-centred series of exchanges to a moreharmonious group process underpinned by theCromlix principles. The change in ‘atmosphere’associated with the episode is perhaps most readilydescribed as a change in the emotional state ofthe members of the group, recognizing that theemotional dynamics of the event were both evidentand significant in shaping what transpired.

In research terms what had unfolded was in thebroad area of the project’s remit, but totallyunanticipated in terms of detailed content andtiming – driven, as it was, by the problems faced

by SHAS and the dynamic of which they werepart. Complexity theory, management and psycho-analytic perspectives were used in conversation to help interpret events, spot patterns, diagnoseresponses and ultimately agree practical stepsforward in which the distinctive theoretical originsof inputs receded into the recent past. As a prac-tical way forward for SHAS gained acceptance,conversation turned to the research questions anda more introspective examination of what hadunfolded in terms of interpersonal dynamics,unplannedness, emotion, intuition, control, etc.

The project moved into a period of more con-ventional work on scenarios, systems boundariesand possible conflicts with adjacent agencies, usingstrategy workshops and other mechanisms.During this period of the project, new practicesand service areas emerged from a series of con-versations and interactions which were ostensiblygoverned by the Cromlix principles. Throughoutthese meetings, a bi-focal stance was adopted. Onthe one hand we dealt with the micro-issuesinvolved in allowing the project to progress (NewProduct Development forms, changes to reportformats, IT/training issues etc.) as individuals and in subgroups. On the other hand, the wholegroup engaged in a collective endeavour to standback from the details of their activities and lookfor patterns, different feasible interpretations ofdeveloping situations, possible trends, etc. Thismore detached, collective view in turn fed backinto the individual micro-processes and ongoingactivities in problem-solving and research processterms.

At around the halfway stage, during areflection session, different members of the groupvolunteered that patterns of behaviour and relatinghad changed significantly since the start of theproject. According to those within the organ-ization, SHAS now had a clearer identity, workrequests from outside were addressed to SHASrather than to individual members and internalpreoccupations had moved away from inter-personal relationships and toward how the organ-ization as a whole related to its environment.There was general agreement that, at some stageover the preceding month or two, SHAS hademerged as a new organization. There was grow-ing coherence apparent in the activities of SHAS,many of which were initiating in new areas ofendeavour, though this was intermittently over-whelmed by a ‘return to the old ways’.


2 Complexity theory argues that order can be thoughtof as resulting from the repeated application of asimple set of order generating rules. Coveney andHighfield (1996) offer an overview of this issue andfurther details of our own work with order generatingrules can be found in MacIntosh and MacLean (1999).

01_Mac 26/11/02 1:05 pm Page 195

In the latter stages of the project, anothersignificant episode occurred. SHAS membersarrived for one of the project meetings somewhatpanicked and forlorn. The reason for this was aletter received from senior civil servant in theScottish Executive containing recommendationsabout how SHAS should change its service. Thecore team’s initial reading of the letter was one inwhich the message was taken as a thinly veiledthreat to SHAS’s continued operation, trying tomanoeuvre it into an unattractive and untenableposition. The letter was seen as the herald of aforthcoming crisis and the response of the SHASteam closely resembled the pre-Cromlix pattern.A reflection session focused on SHAS’s inter-pretation of the letter, the team interactionassociated with that interpretation and possiblealternative (more distanced) readings of thesituation. This culminated in general agreementthat the letter was formally asking SHAS to enactchanges which SHAS had itself developed asdesirable options during the project and whichmay well have been informally signalled to thecivil servant in question.

The letter was thus transformed from aperceived threat to a perceived endorsement. Asthe group reflected on this shift in perception,there was agreement that the Cromlix principleswere in operation during the reflection processbut had not been during the initial reading of theletter. It was also agreed that the perceived crisisassociated with the letter had largely been createdby SHAS and indeed may have served thepurpose of effecting a transition to the enactmentof the Cromlix principles. There followed a moregeneral discussion of the role of perception,intention and emotion in the creation of emergentproperties. This discussion drew on concepts fromboth complexity theory and psychodynamics. Thegroup seemed to have increased its awareness ofthe role of expectations and the quality of inter-actions in creating, in real time, the reality of theircurrent experience and future trajectory from arange of possible options.

Again the team had observed a recurrentpattern in its own behaviour, whereby crisis and aqualitative change in interrelating went hand inhand, each one in part being created by the other.In contrast with the Cromlix event, which effecteda change to internal patterns of interacting, onemight say that this particular episode bore all thehallmarks of the earlier one (threat to survival,

crisis, reflection, harmony, etc.), but was con-cerned largely with SHAS external relationships.

To summarize this brief overview of theproject, we observed a series of ‘unplanned’ criseswhich facilitated (or coincided with) a qualitativeshift in reflexive practices. These were nearlyalways manifest as a threat to the survival or con-tinuation of the organization, the project or themeeting in question. Each time such a threatappeared, new practice emerged. In terms of ouroriginal research question(s), we had developedinsights into the roles of intuition, emotion andreflexivity in organizational transformation andemergence. These insights were grounded inunplanned episodes that occurred whilst workingon a series of specific organizational problems.

Both the transformation of SHAS and the con-duct of the research occurred in a singleknowledge-producing process, the broad natureof which can best be described in metaphoricalterms as an ongoing, constructive and creativeconversation with occasional bursts of newemergent order and insight during episodes ofturbulence in an otherwise relatively smoothunfolding flow.

Insights from the practice of mode 2management research

Having presented an account of a mode 2 project,we can now relate this example to the fivefeatures of mode 2 laid out earlier in the paper.We have dedicated more attention to the sectionon knowledge production in the context ofapplication, partly because we feel that it needsmost explanation, and partly because it bestillustrates the means of delivering in the academicand practical dimensions of the project whilstproviding a detailed context for discussion of theother features.

SHAS and knowledge produced in the context of application

By way of scene-setting, we should point out that this project is part of an ongoing series ofproblem-solving engagements through which wewere exploring the applicability of complexitytheory to organizations in partnership with a net-work of practitioners (MacLean and MacIntosh,2002, offers an overview of these projects as a


01_Mac 26/11/02 1:05 pm Page 196

set). Without being diverted from the main thrustof this paper, there was broad agreement withinour network that complexity theory held somepromise, but that much of what had been writtenwas opaque on key issues (interested readers arereferred to MacIntosh and MacLean, 1999, 2001;MacLean et al., 2001 for details on the content ofthis research).

Specifically, we felt most management writingson complexity theory fail to address (at leastexplicitly) the question, ‘how do the distinctiveattributes of human beings affect the applicabilityof complexity theory, as developed in the naturalsciences, to human systems?’ Bypassing decades,if not centuries, of debate in social theory andphilosophy, the majority of organizational com-plexity theorists offer radical insights into the work-ing of organizations on issues such as strategy,change and creativity without considering thedifficulties introduced by say, the action-structuredebate, the introduction of reflexive capabilitiesor the role of self-awareness. At the outset thiswas a concern shared by both the academics and the chief executive of the organizationconcerned.

As we have stated, our initial research questionwas then translated to become: what are the rolesof intuition, emotion and reflexivity in transform-ing SHAS?

In relation to SHAS’s practical concerns, thequestion of complexity theory applied to socialsystems became translated into a concern for the appearance of new patterns of internal andexternal interaction for SHAS. This is equated to the development of new micro-processes ofworking and relating within the core team, newstrategies for SHAS at the business level and astronger collective identity for the organization.

That such developments would be relevant tothe organization should be self-evident. Theresearch questions were thus ‘framed in thecontext of application’ as befits a mode 2 project,attempting to deliver benefits to the organization,whilst generating data and insights in relation tothe roles of intuition, emotion and reflexivity inorganizational change.

Having described the initial conditions of theproject, we can now attempt to summarize whatwas produced in terms of knowledge as the dynamicevolved and unfolded.

First, in relation to practical knowledge forSHAS in the conduct of its business, several key

developments occurred. Perhaps the most signifi-cant of these was a realization that the organizationwas seeking to express its internal managementprocesses in ways which mirrored the processeswhich underpin its audits. This has resulted inexplicit plans and actions to develop managementprocesses which are distinctive to SHAS and itshistory, as opposed to yielding to external pressureto conform to types and styles which are prevalentin adjacent government organizations. This hasalso been instrumental in building SHAS’s identity,since the inspection process is something whichall core team members practise and with whichthey all identify but which, until recently, they hadviewed as separate from the management pro-cesses of SHAS as an organization. In a real sensethis issue was the culmination of the project, sinceit gave rise to simultaneous developments ininternal and external interactions whilst focusingand strengthening SHAS’s identity. The Cromlixprinciples remain central to SHAS’s view of howit aims to organize itself and conduct its activities.

In addition, SHAS developed a clear view ofpossible moves into new areas of activity, a newcommunications strategy and a new means ofprioritizing initiatives and allocating resources, all of which emerged during project interactionsand through engagement in dialogue addressingpractical issues.

As regards the academic dimension of the work,we feel that our understanding of transformationprocesses within organizations developed consider-ably as a consequence of the project describedhere. From a broad question at the outset, the pro-ject was characterized by a progressive focusingaround key issues and insights generated by thekey episodes described in the empirical account.

A full account of the theory building dimensionof the project in relation to emergence and stra-tegic change is presented elsewhere (see MacLeanet al., 2001) but, in summary, we appear to have asituation where new patterns or strategies emergeepisodically when individuals reflexively interactto transform their own emotionally stimulatingunplanned experiences into collective interpret-ations and coordinated interactions. Our currentposition in relation to the research content of theproject is that a particular process of inquirytriggered by unforeseen events both recognizesand gives rise to the occurrence of patterns,which, in stark contrast to views espoused in themajority of management writings on complexity


01_Mac 26/11/02 1:05 pm Page 197

theory, we believe can be shaped and managed bycoordinated reflexive practices which fully acknow-ledge the emotional and intuitive dimensions ofthe participants in the process.

In relation to the problem-solving nature of the work as described by Gibbons et al. (1994)and summarized earlier in this paper, it is perhapsworth noting that an ‘aspect of continuous nego-tiation’ was indeed a feature of the project. In thiscase, negotiation took two main forms; on the onehand there was an ongoing process of adjustingagreed practical actions and timescales to accom-modate changing priorities and unforeseen develop-ments in the internal and external context of theproject. These negotiations were not so mucharound departures from a predetermined plan asthey were about crafting an agreement on how toapproach an unfolding reality which is largelyunknowable in advance. On the other hand, nego-tiation focused on reaching (or not) collectiveinterpretations of events and apparent patterns inbehaviour – which in turn fed back into the othercategory of negotiation.

In this particular project, the negotiations wereconducted in the spirit of the Cromlix principles,and were inextricably bound up with the reflexivenature of the work.

SHAS & transdisciplinarity

In terms of transdisciplinarity, the research wasmainly informed from three angles – complexitytheory, management (strategy) and psychology(psychoanalysis). At the outset, managementtheory was primarily concerned with issues ofwhat SHAS did and might do. Complexity theory,perhaps the dominant input in this case, typicallyaddressed how changes came about and the psy-choanalytic theory most often addressed issues ofwhy. It is worth noting that in this case (and pos-sibly in most management research), the trans-diciplinary nature of the project stems partlyfrom the transdisciplinary nature of managementitself (see Tranfield and Starkey, 1998), and partlyfrom the transdisciplinary nature of complexitytheory. However, our sense of transdisciplinarity inrelation to this project is that it is primarily aconsequence of the problem-solving nature of the project and that in Tranfield and Starkey’spiece there is an unarticulated (and debatable)assumption that all management research is, orshould be, of this type.

The kind of problem addressed in this examplerequires action which is informed from a varietyof angles and, in the process of synthesizing newknowledge, delivers innovations which cannotthen be reduced to the original inputs. It istherefore the action which delivers transdis-ciplinary outcomes from initially interdisciplinaryprocesses, where the action is tightly bound upwith the evolving problem. As the projectevolves, transdisciplinary outcomes reflexivelyfeed back in to the project, in the ‘process ofongoing negotiation’ and the project develops amore genuinely transdisciplinary feel over time.

This was exemplified in this project by ourexperiences of infusing complexity theory withpsychoanalytics or vice versa, resulting in apossible development which would be neitherwholly complexity theory nor wholly psycho-analytics, a transdisciplinary framework emergingin context from interdisciplinary beginnings.3 Wewould therefore argue that transdisciplinaritydevelops from interdisciplinary, or intertheoreticalperspectives in a way that is largely shaped by thenature of the problem and the key players incrafting solutions in the twists and turns of con-text. This is perhaps what is meant by ‘problemsolving on the move’ (Gibbons et al., 1994, p. 5).

SHAS: heterogeneous and diverse organization of research capability

That the research capability in the SHAS projectwas socially distributed derived partly from thedivision of labour as regards specialisms and skillswithin the group and partly from a more dis-tributed appreciation of strategy and manage-ment. As regards the latter for example, somemembers of the SHAS staff had received trainingin management and research; one had an MBA,another a PhD etc. In relation to different skillsand specialisms which might help with problemsolving, the group (including the academics) hada range of capabilities to draw on ranging from IT, through team-building to PR and businessplanning.


3 There are some difficulties associated with the notionof transdisciplinarity in this context, e.g. Tranfield andStarkey (1998) argue that management itself is trans-disciplinary, and similar claims are also made aboutcomplexity theory (see Waldrop, 1992).

01_Mac 26/11/02 1:05 pm Page 198

These different skills were used to varyingdegrees at different stages of the project’s evolu-tion during the process of ongoing negotiationdiscussed in the previous section. Processes ofanalysis, interpretation and decision-making werelargely concentrated into the reflection sessionswhere the majority of team members werepresent, although during such sessions individualsalso presented opinions, reports etc. based onwork which had taken place in SHAS, on anindividual or collective basis, since the previoussession.

Meeting locations moved around from SHAS’spremises to the university and a variety of otherprivate venues. The atmosphere in the researchteam was focused but relatively informal andcertainly amicable, with research meetingsbridged by relaxed social occasions and teammeals on many occasions. Towards the end of the project, the team had cemented to such anextent that a frequent topic of conversation andoccasional anxiety on all sides was ‘life after theproject’.

This being said, the internal constitution andbehaviour of the team was relatively fluid. Onlytwo meetings had all members present; one mem-ber departed about a third of the way through andan additional academic member joined for thesecond half of the project. Who was there waspartly determined by, and partly a determinant of,what happened at the meeting.

It cannot be emphasized enough that everyonehad a contribution to make to the researchprocess which, in this case, is simply a differentfacet of the problem-solving process. Whilst at the outset there were clear distinctions in terms of primary goals and contributions, these hadblurred considerably by the end, with the projectalmost tending toward homogeneity internally,the heterogeneity being preserved by influencesand activities of individuals when working outsidethe project.

Of particular note is how this blend of hetero-geneity and homogeneity evolved over time. Inthe first half of the project, the team might best bedescribed as a heterogeneous array of skills andspecialisms integrated by a ‘framework of action’(Gibbons et al., 1994, p. 4) which sought primarilyto find ways of solving ‘the problem’. However,by the second half of the project, the distinctionsand boundaries between people and their respect-ive areas of expertise seemed to give way to a

process whereby, through the transfer of skillsand understanding between members, internaldifferentiation diminished to the extent that theteam was composed of generalist co-researchers,willing to contribute to whatever needed doing,be it theory building or management processdevelopment, or whatever.

This internal homogenization seemed to gohand in hand with the development of genuinetransdisciplinarity mentioned earlier, and isevidenced by such features as the contribution ofSHAS’s chief executive as a co-author of thispaper, or the key role played by one of the aca-demics in the subgroup which drove the adoptionof a new product development process. The team-development dimension of this project and thenotion of co-research (with the participation ofteam members as equals in one another’s originaldomains) is a key consequence of both diversityat the outset and the onset of harmonization asthe project unfolded and reflexive processes tookeffect.

SHAS and socially accountable and reflexiveprocesses

Theory building and consensus on problem-solving activities were largely concentrated intothe one- or two-day meetings described earlier.Each project meeting tended to contain a reviewof developments to date alongside reflections onpossible interpretations and future possibilities.These sessions were concerned with viewing ac-counts and conversationally searching for insightsand potential actions from the various theoreticalperspectives, professional viewpoints and common-sense approaches at play. As such, the processtended towards the subjective and creative end of the spectrum, rather than the distanced andobjective description often associated with someforms of action research.

The various lines of accountability involved inthe project influenced, either directly or indirectly,the individuals at each meeting. For example,SHAS reports directly to a senior minister of theScottish Executive, which has committed itself to improving the provision of healthcare inScotland. As a result, the performance of bodiessuch as SHAS is under constant review. More-over, healthcare is a matter of public interest andSHAS’s activities are regularly scrutinized byboth the press and a variety of pressure groups.


01_Mac 26/11/02 1:05 pm Page 199

Team members also have accountabilties andstandards in mind with regard to their individualprofessions (e.g. medicine, nursing, psychologyand management) as do academics with regard toresearch standards, theory building, publication,etc. There are also a variety of day-to-day consid-erations such as interpersonal dynamics, projecttimescales, budgets etc. and, critically, delivery ofpractically useful results.

With all of these factors at play, it should beclear that the research team could not havefunctioned if it were not possible for the membersto adopt other perspectives, so that project activ-ities and outcomes were sensitive to the interests,aspirations and powers of those involved.Reflexivity was the engine of collective creativityfor the project. For this particular project team,collective reflexivity was underpinned by theCromlix principles and the success or otherwise ofthe project may be traced to the effectiveness of them in helping align individual reflexiveprocesses.

The reflexive dimensions of the project werehowever not solely concerned with consideringthe perspectives of others. Indeed, they extendedto, and became primarily concerned with, the pro-cesses of collective knowledge production both interms of problem solving and theory building.Here, the project displayed an interesting dynamic.Whilst both dimensions were present throughout,in the first half of the project collective reflectionsessions were more concerned with practicalproblem solving related to the development of SHAS than with the theoretical details of the research question. In the second half of theproject, the balance tilted the other way. This maybe a natural evolution in all such projects or par-ticular to this case only, but is worthy of furtherinvestigation.

SHAS and a diverse range of quality controls

We have already cited Pettigrew’s notion of the‘double-hurdles’ facing management research,and in particular the need for scholarly quality tobe combined with practitioner relevance.

Dealing first with the issue of practitionerrelevance and value, the real evidence of value forSHAS must lie in the behaviour of the organ-ization itself. In essence this highlights the linkbetween change processes and organizational out-comes or performance, a dimension which Pettigrew

et al. (2001) identify as under-researched andpoorly understood.

While there is not always a straightforward linkbetween a particular initiative and overall perform-ance, the negotiation of acceptable diagnosticsand measures should be an ongoing facet of anyproject and, in this case, was addressed at virtuallyevery meeting. As described earlier, the practicalconcerns of the SHAS management team focusedon the development of new patterns of internaland external interaction. This equated to thedevelopment of new micro-processes of workingand relating within the core team, new strategiesfor SHAS at the business level and a strongercollective identity for the organization. Our reporton the development of SHAS in terms of bothidentity and working practice are in part endorsedby the contribution of the organization’s chiefexecutive to this paper and the fact that theacademic institution was paid for our involvementas agreed at the outset – subject to the delivery ofvalue to the organization.

In a sense we are addressing Pettigrew et al.’s(2001) concern for a better understanding of thelink between process and performance by focus-ing away from notions of performance in a moreabstracted, benchmark format toward perform-ance as a negotiated, context-specific agreementon desired outcomes as expressed by the keystakeholders in the project (in this case the SHAScore team and the academic researchers). Wewould argue that such a focus is more consistentwith the broad principles of mode 2; just as know-ledge is produced in the context of application inthe general sense, so knowledge relating to per-formance is inextricably bound up in a particularcontext.

The exchange of researchers’ time for directpayment is something of a soft target for thosewho, for whatever reason, prefer to see contractor applied research as a necessarily less valuable,less pure form of engagement. The propensity for scientific objectivity to fall victim to themachinations of commercialism has been pointedout elsewhere (see Eden and Huxham, 1996).Undoubtedly, sloppy research and thinly dis-guised consultancy could be conducted under theauspices of mode 2. However, we would arguethat as management scholars, the willingness ofpractitioners to pay directly for our research isone sensible diagnostic of the value which theyattach to the process. Receiving payment usually


01_Mac 26/11/02 1:05 pm Page 200

signals that the detailed terms of a problem-solving contract have been met. The fact that manypractitioners would be unwilling to pay directlyfor involvement in a research project, or in somecases even to participate in research which hasbeen heavily subsidized by another (often govern-ment) funding agency, would seem to substantiateclaims that management research holds little ofvalue or direct relevance for them, in the short tomedium term at least.

However, the availability of direct funding andthe satisfaction of practitioners in no way guaranteesthe academic quality of the research conducted.In this regard, the question of quality centres onacademic output in general and, in particular,whether a system which is predominantly mode 1in orientation can do justice to the outputs ofmode 2 projects.

In our reading of Gibbons et al. (1994), theauthors appear to adopt the view that in mode 2processes, the knowledge produced is transientand context-specific to the extent that no general-izable theory can be developed and that theory istransferred by individuals moving between con-texts and projects. If this is indeed the case, itrepresents a major obstacle to mode 2 researchers.In our experience however, this is not the case(though we should acknowledge that we areprobably drawing on more project experiencesthan the single case reported here, in stating thisview).

In the specific example of research intoorganizational change processes, mode 2 projectsrepresent an opportunity to experience the issueswhich are being explored. This leads to a mirror-ing of content and process which, in turn, rendersgeneralizable, process-level, theory buildingpossible. If this argument is accepted, it may wellpoint towards a typology of project for whichmode 2 is a suitable method.

We would argue that mode 2 is particularlysuited to addressing issues of change and know-ledge management, which may in turn go someway towards explaining current interests in theapproach. We feel that there should be no reasonin management why mode 2 papers are any lessacceptable to peer-reviewed journals (althoughperhaps only to those where a positivist bias doesnot exist). The need for new quality controls is thus debatable; the need for a broader range ofcontrols in order to incorporate the interests ofpractice is not.

Discussion

Having reviewed the way(s) in which the featuresof mode 2 manifested themselves in the SHASproject, we can begin to address the other aim ofthe paper, i.e. the question of what, if anything, isdistinctive about the version of mode 2 describedin the SHAS project in comparison with its morefamiliar counterparts as detailed in the mappingpresented earlier?

First, we believe that in discussing the individualfeatures presented in the previous section itbecomes apparent that they are actually overlap-ping facets of a single dynamic. Transdisciplinarity,context-specificity, research organization, reflex-ivity and quality control are mutually inter-dependent, increasing the significance of all fivefeatures being present in a particular project. Inthe light of this statement, the optional status ofthe five features in, say, action research (as arguedearlier) takes on new importance. It could beargued that the complex pattern set up by the inter-action between these features will be qualitativelydifferent depending on the number of them (2, 3,4 or 5) which are simultaneously present.

Second, the knowledge production process inthe project developed in a way that was beyondthe direct control of any individual or subgroup-ing. In that sense, what actually transpired wasnot designed in detail, since the existence of adetailed design may have unnecessarily con-strained the project as it unfolded. In essence thedetails and features of each project emerged outof a heterogeneous network of interacting con-cerns, skills and perspectives. There was, however,clear agreement on the processes of relating andinteracting between team members, particularlyconcerning ways of dealing with the uncertaintywhich the project would encounter and indeedcreate.

Third, whilst it is not immediately apparentfrom the example given here, there is a cumu-lativity of issues and outcomes at work acrossprojects – as we state in the introduction to theproject, the research questions in this case weregenerated through work on a prior project (andhave given rise to related questions in subsequentprojects as described in MacLean and MacIntosh,2002).

Although it could be that this is just a coin-cidental feature of our own work, we believe thatit is precisely this cumulativity which provides


01_Mac 26/11/02 1:05 pm Page 201

both the academic and practical benefits whichderive from mode 2 research at the level of sub-stantive content (as opposed to process). Thiscumulativity is most easily and identifiablylocated in the realm of theory development whichraises a question about the issue of general-izability.

In contrast with Gibbons et al.’s view of theoryas context specific and transient in mode 2, Edenand Huxham (1996) point out that, with actionresearch, generalizable outcomes are generatedat the theoretical level and they invoke this as ameasure of the quality of the research beingconducted. We would not dispute this claim insome cases (for example on change managementprojects as argued in the previous section) butwould suggest that the real issue is not that ofgeneralizability but that of transfer.

As each project proceeds, theory is experi-mentally adapted and built. This ‘interim theory’can of course be diffused through the process ofpublication, but is also held as a set of un-answered questions and embryonic concepts inthe repertoires of those involved in the work. It isnatural that these questions and models expressthemselves in subsequent, and apparently, unrelatedprojects.

Because of the interaction between theory andpractice in mode 2 research, this cumulativityapplies also to practice, although in a weakerform. We would suggest that it is this process oftransfer that leads to the application of generaltheory, which is in fact interim theory, in newcontexts and that this is most readily effected bythe researcher through practice.

We would further argue that the publicationaims of mode 2 research should not simply be‘generalizable’ theory (since there are, after all,no generalized contexts), but rather a presentationthat illustrates the process of both theoreticalabstraction and meaningful contextualization, i.e.the general theory which might flow from mode 2work is process-level theory, such as the theory oftransfer that we begin to explore above.

Conclusions

In this paper we have sought to investigate therelationship between mode 2 as described byGibbons et al. (1994) and a range of establishedmanagement research approaches. We have also

attempted to identify, what, if anything, is distinct-ive about the particular form of mode 2 researchwhere all five features are simultaneously presentin a single project. However, before drawing ourconclusions, it is worth noting that in a critique ofLevi-Strauss’s structuralism, Patrick Baert pointsout that epistemological difficulties may arise‘when the object of research is also the mediumthrough which that research becomes possible’(Baert, 1998, p. 28). This is precisely the situationwe find ourselves in here.

We are engaged in a knowledge productionprocess, the purpose of which is to examine thenature of knowledge production processes. Wetherefore proceed to draw conclusions withsome caution.

Returning to the first aim of the paper, we haveattempted to relate mode 2 as a concept to actionresearch, clinical method, grounded theory andcooperative inquiry since each of these approachesmay make a legitimate claim to producing know-ledge ‘in mode 2’. The process of relating them tothe framework developed by Gibbons et al. hashighlighted possibilities for the development ofeach approach, as well as providing a useful basisfor dialogue between researchers from each schoolof thought. We will aim to provide an example ofinput to such dialogue below as we address thesecond aim of our paper.

We have also argued that the specific form of mode 2 where all five features are present (wewill call it 5mode2) does differ, in both its conductand the character of its output, from any of theother approaches we have considered in this paper.

5mode2 is similar to action research (or at leastin so far as it is depicted in Table 1b) but actionresearch does not necessarily require diversity interms of either the disciplines, participants ororganizational location(s) involved. Action researchof the type depicted in Table 1b also tends toemphasize a relatively objective and scientificallydistanced stance (see Eden and Huxham’s twelvecontentions for action research including refer-ences to orderliness, triangulation, validity andgeneralizability, 1996, p. 84), which differs fromthe more subjective and interpersonally creativenature of the outputs from the 5mode2 projectpresented here.

5mode2 is closer still to cooperative inquiry, inthat we have stated that the SHAS project pro-gressively migrated towards a position where theparticipants became equal co-researchers. However,


01_Mac 26/11/02 1:05 pm Page 202

many of the cooperative inquiry projects reportedin the literature take place within a single organiza-tional setting, a single disciplinary setting or both(Heron and Reason, 2001). The SHAS project did encompass different disciplinary perspec-tives (complexity theory, management and psycho-dynamics) but even this example could havefeatured more organizational diversity by forexample including representatives from govern-ment, healthcare providers or pressure groups inthe research team.

We have also considered the contexts withinwhich a 5mode2 approach might be best suited.We have argued that in 5mode2 projects, the sep-aration of a dynamic into a problem-solving and aresearch dimension is somewhat artificial. Thisbeing the case, it may be that 5mode2 researchprocesses are best suited (and perhaps onlysuited) to the investigation of situations in whichaction leads to change; indeed, we might arguethat mode 2 is perhaps the only consistent way oflooking at change, i.e. ‘from the inside’ of adynamic which can only be accessed throughexperience. Furthermore, we have argued that, inthese specific conditions, the practitioner might(and perhaps should) be willing to pay directly forthe research.

Our experiences in the SHAS project have ledus to question the nature of knowledge producedin 5mode2 projects. Here, we are reminded ofSusman and Evered’s arguments surrounding ashift away from positivist grounds (1978) and thesomewhat depressing response that nineteen yearslater, Eden and Huxham (1996) report action-oriented approaches still experiencing difficultyin finding acceptance on the grounds that they are‘not science’ (p. 78). We believe that 5mode2 hasmuch to offer both the academic and practitionercommunities, but that a serious debate about theepistemological nature of management knowledge,with a more open embrace of constructionist per-spectives than has been the case to date, shouldincrease in priority.

Finally, we would like to finish the paper byhighlighting an obvious area for further work. Wepointed out in the introduction to the paper thatwe would simply accept the five features of mode2 presented by Gibbons et al. Undoubtedly for us,the combination of these features in what we havecalled 5mode2 has produced an enjoyable, reward-ing and valued (by the practitioners concerned atleast) research process. However, given that we

have argued that it is the precise cocktail of thesefive features which give 5mode2 the characteristicswe have described here, it would seem obvious toquestion the choice of these five features.

In the first instance, these questions mightrelate to the appropriateness of Gibbons et al.’schosen features in the context of managementresearch as opposed to the natural sciences andtechnological research where they were developed.A secondary set of questions might be associatedwith the features themselves in terms of bothnumber and character, e.g. why not 3mode2 oreven 7mode2?

Having argued for the value of using 5mode2,we feel that there is also a need for a moreconsidered theoretical examination of the choiceof features. Interest in mode 2 provides a wel-come opportunity for management research toembrace practice-oriented approaches and bridgewhat Starkey and Madan describe as ‘the rele-vance gap’ (2001). The features of mode 2 alsoprovide a framework within which meaningfuldialogue might take place between users of thevarious approaches discussed in this paper (alongwith users of many other approaches which wechose not to consider). Hopefully in time thatdebate may turn its attention to Gibbons et al.’s(1994) presentation of mode 2 itself, and thusgenerate a version better suited to dealing withthe specifics of management research.

Acknowledgement

The authors thank David Tranfield for his helpand encouragement in the development of thispaper.

References

Baert, P. (1998). Social Theory in the 20th Century. New YorkUniversity Press, New York.

Coveney, P. and R. Highfield (1996). Frontiers of complexity.Faber & Faber Ltd, London.

Crotty, M. (1998). The Foundations of Social Research:meaning and perspective in the research process. SagePublications, London.

Denscombe, M. (1998). The Good Research Guide for SmallScale Social Research Projects. Open University Press,Buckingham.

Eden, C. and C. Huxham (1996). ‘Action research for man-agement research,’ British Journal of Management, 7(1), pp. 75–86.


01_Mac 26/11/02 1:05 pm Page 203


Gibbons, M., C. Limoges, H. Nowotony, S. Schwartzman, P. Scott and M. Trow (1994). The New Production of Know-ledge: the dynamics of science and research in contemporarysocieties. Sage Publications, London.

Glaser, B. G. and A. L. Strauss (1967). The Discovery ofGrounded Theory: strategies for qualitative research. Aldine,New York.

Goldstein, J. (1999). ‘Emergence as a Construct: History andIssues,’ Emergence, 1(1), pp. 49–72.

Gopinath, C. and R. C. Hoffman (1995). ‘The Relevance ofStrategy Research: practitioner and academic viewpoints,’Journal of Management Studies, 32(5), pp. 575–594.

Hendry, J. and D. Seidl (2002). ‘The Structure and Significanceof Strategic Episodes: social systems theory and the routinepractices of strategic change,’ paper presented at EURAMII, Stockholm, May 2002.

Heron, J. (1996). Co-operative Inquiry: research into thehuman condition. Sage Publications, London.

Heron, J. and P. Reason (2001). ‘The Practice of Co-operativeInquiry research “with” rather than “on” people’. In: P. Reason and H. Bradbury (eds), Handbook of ActionResearch: participative inquiry and practice. Sage Publica-tions, London.

Hodgkinson, G. P. (ed.) (2001). ‘Facing the Future: The Natureand Purpose of Management Research Re-Assessed’, BritishJournal of Management, 12, Special Issue, pp. S1–S80.

Huff, A. S. (2000). ‘Changes in Organisational KnowledgeProduction: 1999 Presidential Address,’ Academy ofManagement Review, 25(2), pp. 288–293.

Huff, A. S. and J. O. Huff (2001). ‘Re-focusing the BusinessSchool Agenda,’ British Journal of Management, 12, SpecialIssue, December 2001, pp. S49–S54.

Lewin, K. (1946). ‘Action Research and Minority Problems,’Journal of Social Issues, 2, pp. 34–46.

MacIntosh, R. and D. MacLean (1999). ‘Conditioned Emer-gence: a dissipative structures approach to transformation,’Strategic Management Journal, 20(4), pp. 297–316.

MacIntosh, R. and D. MacLean (2001). ‘Conditioned Emer-gence: researching change and changing research,’ Inter-national Journal of Operations and Production Management,21(10), pp. 1343–1357.

MacLean, D., M. Mayer, R. MacIntosh and S. Grant (2001).Reflexive Strategizing, Proceeding of the Workshop on MicroStrategy and Strategizing, European Institute for AdvancedStudies in Management, Brussels, 1–3 February 2001.

MacLean, D. and R. MacIntosh (2002). ‘One Process, TwoAudiences: on the challenges of management research,’European Management Journal, forthcoming.

Park, P. (1999). ‘People, Knowledge and Change in Partici-patory Research,’ Management Learning, special issue: theaction dimension of management: diverse approaches toresearch, teaching and development, 30(2), pp. 141–158.

Partington, D. (2000). ‘Building Grounded Theories of Man-agement Action,’ British Journal of Management, 11(2), pp. 91–102.

Pettigrew, A. (1995). Distinguished Scholar Address to theOrganisation and Management Theory Division of the USAcademy of Management, Vancouver, August 1995.

Pettigrew, A. (2001). ‘Management Research After Modernism’,British Journal of Management, 12, Special Issue, December2001, pp. S61–S70.

Pettigrew, A. M., R. W. Woodman and K. S. Cameron (2001).‘Studying Organizational Change and Development:challenges for future research,’ Academy of ManagementJournal, 44(4), pp. 697–714.

Reason, P. (1999). ‘Integrating Action and Reflection ThroughCo-operative Inquiry’, Management Learning, special issue:the action dimension of management: diverse approaches toresearch, teaching and development, 30(2), pp. 207–226.

Reason, P. and H. Bradbury (eds) (2001). Handbook of ActionResearch: participative inquiry and practice. Sage Publi-cations, London.

Schein, E. (1987). The Clinical Perspective in Fieldwork. SagePublications, London.

Schein, E. (2001). ‘The Clinical Method’. In: P. Reason andH. Bradbury (eds) (2001). Handbook of Action Research:participative inquiry and practice. Sage Publications, London.

Starkey, K. and P. Madan (2001). ‘Bridging the RelevanceGap: aligning stakeholders in the future of managementresearch,’ British Journal of Management, 12, Special Issue,December 2001, pp. S3–S26.

Susman, G. I. and R. D. Evered (1978). ‘An Assessment of the Scientific Merits of Action Research,’ AdministrativeScience Quarterly, 23(4), pp. 582–603.

Tranfield, D. and K. Starkey (1998). ‘The Nature, SocialOrganisation and Promotion of Management Research:towards policy,’ British Journal of Management, 9(4),pp. 341–353.

Waldrop, W. M. (1992). Complexity: the emerging science at theedge of order and chaos. Touchstone, New York.

01_Mac 26/11/02 1:05 pm Page 204


Appendix

Table 1a. Mode 1 and the five features of mode 2

Mode 1

Knowledge produced in the Like mode 2, mode 1 characterizes a range of methodologies which might be described as context of application traditional or positivist science. However, across the range of approaches which might be

regarded as mode 1, the knowledge is not produced in the context of application. Rather it isframed, driven and produced in a predominantly theoretical context.

Transdisciplinarity Gibbons et al. point out that mode 1 is disciplinary as opposed to transdisciplinary (1994, p. 3).

Heterogeneity and Whereas mode 2 focuses on multi-faceted problems, which are typically addressed by organizational diversity heterogeneous teams populated from a range of organizations, this is not the case with mode 1.

In mode 1 problems are usually tackled by homogeneous teams from a single (academic)organization.

Social accountability Given the prevalence of both theoretical and organizational homogeneity in mode 1, it is less and reflexivity likely to produce true social accountability. Also its preference for positivist approaches reduces

the likelihood of reflexivity being a key feature of the knowledge production process, or at leastthe likelihood of this being an acknowledged feature of the research.

Diverse range of Mode 1 is typified by a comparatively uniform approach to quality control. Outputs are quality controls generally evaluated from the standpoint of the particular discipline concerned and the evaluation

itself is usually driven by senior academic peers.

Table 1b. Action research and the five features of mode 2

Action research (Eden and Huxham, 1996)

Knowledge produced in the Although we have chosen to work with a particular definition of action research, almost all of context of application the variants which we are not explicitly considering would also sympathize with the need for

knowledge to be produced in the context of application. Indeed this seems obvious given the‘pragmatic focus’ of action research and the fact that it necessarily involves a matter of ‘genuineconcern’ to the organization (Eden and Huxham, 1996).

Transdisciplinarity The transdisciplinarity (or otherwise) of action research is not typically a key feature ofdiscussions of the methodology. Undoubtedly a transdisciplinary approach is possible but it is nota prerequisite condition for most forms of action research.

Heterogeneity and Mode 2 highlights the use of heterogeneous teams and action research typically involves teams organizational diversity of practitioners and at least one academic. In this sense the team is heterogeneous and spans at

least one organizational boundary. Again, it is possible for an action research intervention toincorporate far more diversity (either in terms of team membership or organizational location)but this is not necessary.

Social accountability Given the level of interaction with the user community, action research might reasonably claim and reflexivity increased levels of social accountability over more traditional (mode 1) approaches. It is also

true that action research is reflexive in that it should generate ‘emergent theory’ and that a ‘highdegree of orderliness is required in reflecting about, an holding on to, the emerging researchcontent of each episode of involvement in the organisation’ (Eden and Huxham, 1996).

Diverse range of Action research has a long history of use in the social sciences. It operates within, and can quality controls survive, the peer-based evaluation process linked with mode 1 (as witnessed by journal articles,

books, etc). However, Eden and Huxham point out that the ‘the history and context for theintervention must be taken as critical to the interpretation of the likely range of validity andapplicability of results.’ (1996).

01_Mac 26/11/02 1:05 pm Page 205


Table 1c. Cooperative inquiry and the five features of mode 2

Cooperative inquiry

Knowledge produced in the ‘Co-operative inquiry is . . . a form of action research: it is concerned with revisioning ourcontext of application understanding of the world, as well as transforming practice within it’ (Reason, 1999, p. 208).

From this description it is obvious in cooperative inquiry, as in more traditional forms of actionresearch, knowledge is by definition produced in the context of application.

Transdisciplinarity Cooperative inquiry involves a group of people coming together to explore ideas. Thiscollaborative dimension to the research process may increase the likelihood of crossingdisciplinary boundaries but many cooperative inquiry projects are successfully performed withinthe confines of a single discipline.

Heterogeneity and This approach is described as being ‘research with people rather than research on people’ organizational diversity (Reason, 1999, p. 208). However, as with action research the heterogeneity or homogeneity of

the research team is a design choice made by the participants in the process. The same choiceapplies to the organizational diversity of the inquiry group, with some projects reported as beingconducted within a single organization and others involving more disparate membership.

Social accountability The research process associated with cooperative inquiry is explicitly participative, with and reflexivity everyone having a say in the questions to be addressed, the concepts used and the conclusions

reached (Reason, 1999). In this sense cooperative inquiry provides greater accountability thanmany research methods (arguably including action research) where these issues are the domainof the academic expert. The process also features ‘intentional interplay between reflection andsensemaking on the one hand, and experience and action on the other’ (Heron and Reason,2001, p. 179).

Diverse range of Proponents of cooperative inquiry argue that the outcome of good research is ‘not just books quality controls and academic papers . . . but also the creative action of people to address matters that are

important to them’ (Reason 1999, p. 208).

Table 1d. Grounded theory and the five features of mode 2

Grounded theory

Knowledge produced in the The originators of grounded theory developed the approach to ensure closeness to the context context of application of application, in their view research conducted in this way is practically oriented in that it is

‘grounded in the data’ (Glaser and Strauss, 1967). The approach is also characterized by the‘constant checking of the theory against the findings DURING the process of research’(Denscombe, 1998, p. 215).

Transdisciplinarity Given that grounded theory attempts to build theory from the data available it has beenobserved that ‘grounded theories are not necessarily reliant on established theoreticalperspectives’ (Partington, 2000, p. 93). In this sense, grounded theory ignores rather thantranscends existing disciplinary boundaries.

Heterogeneity and As with action research and cooperative inquiry, grounded theory is compatible with the organizational diversity principles of heterogeneous research teams located across different organizational settings.

It is not, however, a requirement, nor perhaps even a common feature, of cited applications of the approach.

Social accountability Grounded theory projects are typically led by the academic community and might therefore beand reflexivity viewed as less socially accountable than some methodologies. However, theories developed in

this way are the result of ‘continually testing and modifying the coding scheme’ (Partington,2000, p. 99) which requires the researcher to adopt a role which is both creative and reflexive.

Diverse range of Grounded theory focuses on the discovery of theory rather than the verification of theory quality controls (Glaser and Strauss, 1967). This in itself introduces difficulties in terms of quality control but

interestingly a review of exemplary grounded theory studies described their output aspredominantly ‘of the mode 1 type’ (Partington, 2000, p. 95).

01_Mac 26/11/02 1:05 pm Page 206


Table 1e. Clinical method and the five features of mode 2

Clinical method

Knowledge produced in the In research conducted using the clinical method, the focus is on an issue identified by the context of application organization (Schein, 1987). This sounds very similar to the definition of action research offered

by Eden and Huxham, but Schein argues that clinical method studies are initiated by thepractitioner, whereas action research is initiated by the academic (Schein, 2001).

Transdisciplinarity As with the other approaches, any given clinical method project could adopt a transdisciplinarystance, but this is not a central feature of this style of research. Indeed it might be argued thatsince such projects often feature the interaction of a single ‘clinician’ with an organization,transdisciplinarity is unlikely.

Heterogeneity and Each clinical intervention necessarily involves a qualified clinician who is external to the organizational diversity organization concerned but again this is a comparatively limited form of heterogeneity and

organizational diversity in the research process.

Social accountability Given the acknowledged similarities between this approach and consultancy, its social and reflexivity accountability as a research methodology might be called into question. What is not in question

however, is the highly reflexive nature of the process, for the clinician at least (Schein, 2001).

Diverse range of Since this approach is initiated by the practitioner, quality is often assessed in terms of successquality controls (or failure) to resolve the issue(s) concerned. Academics using the clinical method may well face

difficulties in satisfying these criteria whilst simultaneously attempting to meet the requirementsof more traditional, peer-based, mode 1 quality standards.

01_Mac 26/11/02 1:05 pm Page 207

mode 2 management research

Documents