Collaborative Action Research: A Dynamic Intervention Project to Create a Purposeful Virtual Community-of-Practice


This is a chronological account of a 2-year inductive collaborative action research by Cranfield University on Hewlett Packard Enterprise Services’ efforts to form a virtual community-of-practice for connecting geographically dispersed project professionals. As organizations increasingly rely on more “hands-on” learning over traditional education, researchers will have to adopt study strategies that can balance academic and business needs. The focus of the case reported here is the collaborative action research process. It was designed to be academically rigorous for the research institution, yet practically useful for the participating organization. The extended duration of the study was key, as it allowed both early active collaboration and later, close-quarter independent observation by the research institution, including the production of periodic feedback to Hewlett Packard Enterprise Service management for dynamic intervention.

Learning Outcomes

By the end of this case, students should be able to

  • Appreciate the importance of credibility perceptions when conducting in-company research
  • Appreciate the importance of balancing the goal-diversity needs of academic research and business intervention
  • Understand how to design and conduct a collaborative action research project
  • Select different methods for data collection, analysis, and feedback that are commensurate with the needs of the research project and business organization

Context and Collaboration Overview

The case study relates to the Hewlett Packard Enterprise Service (thereafter, HPES). At the time of research, HPES was responsible for a range of application development, infrastructure, business process management, and other outsourced services to industrial and public sector organizations. The subject of interest is HPES’ distributed Project Management Offices (PMOs). Their sheer number (c.200), variety, and geographical spread created a range of challenges for HPES. Although HPES PMO employees are experienced project professionals, they tended to work in small local groups throughout the global organization. Low visibility and connection meant that individuals were largely unaware of “who knows what” or “who knows how” beyond their local PMO. Senior management and many PMO managers agreed on the need to leverage the PMOs’ potential in facilitating project delivery and success. Their discussions led to the idea of a community-of-practice (CoP) as a useful forum to disseminate innovative local practices and reduce PMO personnel isolation. They also agreed that assistance would be required with the research and development of such a CoP. Cranfield University was invited by the senior director to engage as their research partner.

To illustrate the learning outcomes of credibility perceptions and goal diversity, I will explain how we decided upon the research team for the project and summarize the context and considerations for collaboration between HPES and Cranfield.

Formation of the Research Team

The research project was the responsibility of the International Centre for Programme Management (ICPM) at Cranfield. A first consideration for the ICPM group was the formation of a “credible” team for HPES. Both the research literature and ICPM’s past experience with in-company projects suggest that managers are more ready to accept research that is problem-focused, context specific, and drawing directly from their professional experiences to produce future-oriented, practical outcomes (McNiff, Lomax, & Whitehead, 2001; Robinson, 1993). Therefore, members for the HPES project must be knowledgeable about business practices and can understand the technical language of IT project management. An ICPM-wide email was sent seeking initial expressions of interest. A meeting of the interested parties followed to discuss suitability, capacity, and research strategy. The eventual research team comprised a professor of IT/IS strategy and two academic colleagues from the project and program management community (of whom I was one). A dedicated research fellow was co-opted to provide additional research and administrative support. All four of us have prior industry experience and are actively engaged with business-related research.

Pre-Research Scoping Meeting

At the pre-research scoping meeting, an HPES representative (who was to become the in-company CoP convener and point-of-contact for the Cranfield team) provided the background to the “PMO problem.” He also explained senior management’s vision of a CoP as an online forum for disseminating local best practices to a wider audience. Senior management wanted the CoP to have the clear mandate of developing a professional PMO entity guided by specific goals, explicit accountability, and clear executive oversight. We were approached for assistance because they were unclear how to design and implement the PMO CoP or, indeed, how it would be received or function in the longer term.

In turn, we explained the potentials of adapting the principles and phases in action research (AR; see McNiff et al., 2001; Reason & Bradbury, 2008) to the proposed collaborators. Meeting HPES’ practical desire to form a CoP without sacrificing academic rigor was our primary concern. Following our discussion on study duration, roles, and contributions by the two respective organizations (i.e., HPES and Cranfield), it was agreed that the research would span a period of time and Cranfield’s role (as the research institution) would be twofold: starting in active collaboration mode to research the literature on CoP characteristics and formation; then stepping back as non-participatory independent observers at CoP events. As financier and owner of the CoP initiative, HPES would appoint a dedicated convener who could facilitate meetings with PMO personnel and be Cranfield’s point-of-contact for feedback to HPES. The meeting ended with both parties agreeing that the research design must be academically robust and interim reports were to be submitted to HPES to assist practical implementation of the CoP.

Method in Action: The Collaborative Action Research Process

On reviewing the minutes of the initial meeting, we concluded that a collaborative action research (CAR) strategy would be appropriate for the joint study. AR has been around for a long time and is often used to help organizations study workplace practices and develop their capacity for improvement. Many AR projects would have followed a cyclical model similar to Kurt Lewin’s (1946)plan, act, observe, reflect, and replan model. As this was a joint study, we chose to base our CAR design on Gerald Susman and Roger Evered’s (1978) AR cycle of diagnosing, action planning, action taking, evaluating, and specify learning. Their much-cited paper in the Administrative Science Quarterly specifically discussed the scientific merits of AR. Richard L. Baskerville and A. Trevor Wood-Harper’s (1996) paper was also pertinent for this research. They argue that AR is a rigorous approach, is well suited for certain domains such as systems development, can be done in collaboration, and the effects of specific interventions can be tracked through the iterative phases. In our case, both of the participating organizations would have a clear role and contribution to make to the business and research projects, albeit collaboration outcomes must satisfy the dual agendas of practical application and study insights that are worthy of academic sharing. Hence, a distinguishing feature of our CAR design beyond the classic AR is defined partnering roles that would best serve their reasons for undertaking the research project.

Actual duration, research activities at each phase, roles, and interim reports are shown in Figure 1.

Figure 1. The CAR process, roles, and reports.

Briefly, we started as active collaborators working with HPES at the initial diagnosis and action planning phases to establish the scale of the PMO problem before reviewing the literature for characteristics of online CoPs and lessons from previous CoP implementations to develop an action plan. The collaborative inquiry nature of our research meant that our design was aligned with—but not exactly as—Susman and Evered’s (1978) iterative AR phases. We needed to reflect the different roles and responsibilities of the academic and business research partners. Action taking was split into two distinct phases (Phase 3 and Phase 4), and evaluation and learning were depicted as the final phase. Once the CoP was launched, we retreated from our active collaborator role to attend community events as non-participating, independent observers. Guided by tentative hypotheses from the literature review and using a mix of qualitative and quantitative methods, we collected and analyzed the data. In line with the AR principle of evidence-based refinement of current arrangements and monitoring impact of changes for further refinement, our interim reports would help HPES develop an implementation plan, launch the CoP, monitor, and make adjustments to the CoP and its activities. A final evaluation report to HPES marked the end of our involvement in the CoP project.

Each of the five CAR phases, including the interim reports to HPES, is explained in greater detail in the following sections to demonstrate the learning outcomes on how to design and conduct a CAR project, and how to select different methods for data collection, analysis, and feedback that are commensurate with the needs of the research project and business organization.

Phase 1: Diagnosis

Research is about finding out. Our first task at the beginning of the CAR project was to ascertain the state and extent of the PMO problem before assisting in the design and implementation of the purposeful CoP.

Our main challenge at this point was locating the dispersed PMO employees. We could not verify senior management’s estimate of 200 PMOs and did not know how to access the individuals. A conventional sampling approach was not possible owing to the lack of a universal PMO directory. Working with HPES in their role as our engaged sponsor, eight key client accounts with a linked or embedded PMO were identified (e.g., Rolls Royce, UK Ministry of Defence, UK Department of Works and Pensions). Our HPES convener approached the eight PMOs to explain the purpose of the research and to seek their agreement to participate. With HPES management input, we developed a “getting to know you” questionnaire and sent it to everyone on the list of names and email addresses provided by the convener. As the purpose of the questionnaire was merely to provide background information on PMO size, location, number of employees, and nature of work to facilitate later in-depth interviews, we did not feel a pilot was essential. In addition, given the welfare orientation of the intended intervention and the fact that our research subjects are mature experienced professionals, they were not categorized as “vulnerable” in the conventional sense. Nonetheless, we were careful to abide by the AR principles of transparent methodology, voluntary consent and right to withdraw, and data and identity confidentiality. Recipients were assured on the voluntary nature of the research and that only summary findings would be accessible by respondents and HPES management.

Without a PMO directory, access to other PMOs was achieved using the snowballing (referral) technique. We would ask, at each interview, for participants to refer us to colleagues that they knew were working in a PMO. The fact-finding questionnaire was sent to 120 referred PMO contacts (with the expressed assurance about voluntary participation and individual confidentiality) along with an invitation to attend an open workshop at Cranfield University. The purpose of the workshop was to understand the extent to which the issues raised in the key-account interviews were general and to discuss the notion of a CoP as a possible solution. Twenty-five of the 120 referred attended the workshop along with six invited academic researchers active in PMO research from a number of academic organizations in the United Kingdom, Europe, and North America.

The workshop started with a general discussion on the value of establishing an intentional CoP. Major themes from the key-account interviews (e.g., poor visibility, low learning, weak career prospects) were shared to generate more in-depth discussions on how the CoP could address the “PMO problem.” These were held in separate breakout sessions, each facilitated by a member of the CAR team. Permission to tape the sessions was sought and given by the workshop attendees. Thematic analysis from the workshop sessions yielded rich data for our first interim report showing the state “as was” and confirming the common view that a CoP could be the mechanism to overcome connection challenges. A subsequent facilitated brainstorming session for HPES management was held at Cranfield to discuss the contents of the report and to agree the strategy for intervention.

Phase 2: Action Planning

Having completed our diagnosis of the problem, the next step was the co-development of an implementation plan. The key consideration for HPES and Cranfield at this stage was how to create the virtual CoP with an appropriate governance framework. In our capacity as active collaborators, our task was critical engagement with the literature. The main reasons for doing any literature review are to find out what has already been said or studied about our topic of interest and to check our initial ideas and findings against that literature. While it is not possible to detail all aspects of the literature review, nor are we able to share the action plan for the CoP’s implementation for proprietary reasons, I will explain the rationale and basis of our literature search against the questions of, “How do we define the virtual CoP?” and “What are the aims, objectives, and conditions for its formation?”

We began our search on the concept of “community-of-practice.” We found a substantial body of literature on the subject. The classic view by learned researchers Jean Lave and Etienne Wenger (1991) was that a CoP is a situated social system for spontaneous learning. Situated interaction was problematic for our dispersed PMO subjects. However, of resonance was the general assumption that a CoP can facilitate effective problem solving, practice improvement, and self-renewal (Wenger, 1998; Wenger, McDermott, & Snyder, 2002; Wenger & Snyder, 2000). We were also informed by Wenger’s (1991) observations that CoP success and longevity depended on the willingness to participate and the physical artifacts, connection protocols and professional focus around which people can connect and establish a shared PMO identity.

Next, with HPES’ desired virtual community in mind, we turned to search the literature for references to “online CoPs.” It would appear that they are becoming more accessible with powerful technology advances (e.g., Ma & Agarwal, 2007; Silva, Goel, & Mousavidin, 2009; Vaast & Walsham, 2009). Lars Lindkvist’s (2005) paper on the varying definitions (e.g., communities of interest, communities of knowing, collaborative networks) reminded us that a common view was needed on what is a “virtual CoP.” Working on the notion that management has a responsibility to harness fragmented business (Brown & Duguid, 2001; Wenger & Snyder, 2000), we managed to construct a definition for our CoP that represented the purposeful intent of the parent organization:

… a community designed and implemented as an organisational intervention, utilising multiple synchronous and asynchronous electronic platforms to enable local, project and organisational peer-to-peer engagement and mutual learning.

The second question of “What are the aims, objectives and conditions for its formation?” relates to HPES’ expressed desire at the pre-research scoping meeting for clear accountability and executive overview. We interrogated the literature for theories on “CoP governance.” Both Wenger’s (1998) work on the mix and configuration of structural and coordination mechanisms and artifacts, and, more recently, Mark Thompson’s (2005) seeding and controlling structures are instructive for encouraging and managing community dialogue. Another relevant reference for this project was Christopher Koliba and Rebecca Gajda’s (2009) synthesis of the major goals and relationships for learning, mode and quality of knowledge transfer, degree of formalization, and strength of coupling. Our virtual CoP would not have a physical presence in the form of a building or office, so it must rely on tangible artifacts such as an explicit mission statement and terms-of-reference to define its boundaries and governance (see Lee-Kelley & Turner, 2017, for sight of the mission statement and terms of reference).

Another important consideration for the virtual CoP was the “role of technology.” Koliba and Gajda (2009) had already highlighted the need to pay attention to the modes and mechanisms for online exchange. Indeed, something that emerged from the diagnosis phase was the need for regular webcasts on issues valuable to the members, so technology would play a central role in the CoP. Examples in the plan included the access to the company’s web-based meeting facility; a LinkedIn networking forum; ability to record and store each event to provide access to members who could not attend a meeting or others who wanted to remind themselves at a later date of all or certain aspects of a prior meeting.

Our report at the end of this phase was a summary of our review of the concepts, theories, and gaps around the CAR topic. HPES would be able to “take action” guided by our literature findings.

Phase 3: Action Taking 1 (Implementation)

We had been active collaborators during Phase 1 (Diagnosis) and Phase 2 (Action Planning), being responsible for investigating the extent of the PMO problem and researching the literature to help HPES plan their intervention. Meanwhile, HPES management and their team had assumed the role of the engaged sponsor, showing interest, providing general support, answering queries, and making decision as and when required. Phase 3 required HPES to transit from their relatively passive role to become the Implementer-Owner (Figure 1). They would have to enact the structures, artifacts, and mechanisms identified during the action planning phase. The manager who walked us through the PMO problem at the pre-scoping meeting had been formally appointed as convener by senior management. He would be the source of coherence in the founding of the CoP and responsible for negotiating its purpose, mission, terms of reference, and governance. Being only part-time in this role, an early task for him was to secure additional resources by way of volunteers who could help organize gatherings, connect with members off- and online, invite and arrange topics of interest, set meeting agendas, and moderate interactions during events. As HPES stepped up their active involvement in the project, we could withdraw to our new role as non-participatory independent observers.

The first event organized by the convening group was the inaugural meeting. An important task for them ahead of that meeting was to draft the mission statement and terms-of-reference (identified in the literature as important governance artifacts) for senior management authorization and subsequent presentation at the inaugural meeting. The meeting was by invitation only to those who had responded to the “getting to know you” questionnaire, were interviewed as a PMO member of the identified key accounts, or had attended the open workshop. Its purpose was primarily as a thank-you for their participation and to indicate what the CoP could do for them at future meetings. Forty-five individuals from across Europe, the Middle East, and Africa registered as members that day. The online meeting lasted about an hour.

We were also invited. In accordance with agreed good ethical practice, the convener who was chairing the meeting introduced us and sought agreement for our continuing involvement as “fly-on-the-wall” observers at future CoP events. Voting was unanimous as the CAR project was already known to the attendees as we had been in contact with many during the diagnosis phase.

Phase 4: Action Taking 2 (CoP Events)

Phase 4 relates to the second part of action taking which chronicled the monthly events held via the company’s online meeting software suite. Priority areas which had been identified at the diagnosis phase formed the basis for the early events relating to PMO roles and classification, value proposition and relevant performance measures, up-to-date toolkits, rationalization of the extensive yet difficult to use organizational reference materials, compatible job codes and defined career paths for PMO staff, and sharing of “best practices.” These were categorized by the convener into two themes: “collateral and toolkit” and “value and classification” and given to his core group to identify specific presentations within the themes and to organize event speakers.

This second action taking phase lasted the longest period (16 months) in the CAR project. The questions of “how to monitor attendance at events?” and “what kind of evidence to show participation?” were discussed and dealt with during the planning phase. As a technology-intensive organization, HPES was able to build an automated online knowledge system for logging attendances, video capturing presentations, and recording Q&A discussions. This was an important resource for vital data that we, as observers, would find hard to capture in its entirety. We continued attending the monthly community events to observe “firsthand” the formal presentations, practice discussions, and social interactions. We could also approach the convener, his core group, or any individual CoP member for clarification.

Interim Reports

The ability to make changes is a core tenet of any AR (McNiff, 2016). An item in the minutes of the pre-scoping meeting was the agreement for us to submit periodic reports to HPES so that they could take “dynamic action.” An inductive logic was adopted throughout, using a range of qualitative and quantitative research-gathering and analysis techniques to generate our reports: the first upon completion of the diagnosis phase (Phase 1), the second followed the action planning phase (Phase 2), and the third was submitted during the first quarter of the CoP events (Phase 4) to allow the convening team to reflect on the feedback and make changes to improve CoP experience for the membership. Here, I explain how the three interim reports were generated. The fourth report, which marked the formal end to the CAR research, is described in Phase 5 (evaluating and learning).

Interim Report 1

The purpose of the first report was to ascertain the “PMO problem” and whether a CoP could be an acceptable option to address the problem. We had achieved this via a “getting to know you” questionnaire sent to all “known” PMO employees, face-to-face interviews with PMO staff from eight key HPES accounts, and a major workshop at our university. The questionnaire collected background data on where the PMOs were located, how many PMO staff in each, how they communicated locally and with other PMOs, how often, and the specific issues of dispersion including any career concerns. We used simple descriptive analyses such as frequency counting, percentages, and averages, including cross-tabulations for comparison. The face-to-face interviews lasted between 40 and 60 min each. Thematic analysis of subjective accounts using the “mind-mapping” technique produced a richer picture on dispersed working, low visibility, and connection. Facilitators of the breakout sessions at the open workshop used a matrix of key themes and categories from the questionnaire and interviews to instigate session debate and to score the items against the proposed CoP solution. Session scores were aggregated to arrive at a ranked-order set of results for subsequent discussion and strategy generation by HPES management. We were able then to produce a summary report of the major findings and conclusions from the diagnosis phase for authorization by HPES to proceed to the next CAR phase.

Interim Report 2

This report followed the action planning phase, which involved a systematic search of the extant literature. The resultant report was pivotal in guiding the CoP’s design and implementation; specifically on (a) the need for a shared vision of what the CoP was about; (b) the benefits of regulatory objects such as the mission statement and terms-of-reference to provide guidance on the aims and objectives of the community—establish a code for voluntary participation and defining acceptable behavior at CoP meetings; and (c) the need to invest in the “hardware” and “software” enablers for CoP participation. Extracts from our notes of the major items discussed at the inaugural event show HPES’ commitment to the literature to build a long-lasting CoP (per Thompson, 2005; Wenger, 1998; Wenger & Snyder, 2000); namely:

  • Ensuring voluntary participation—Assurance that CoP membership would be strictly voluntary with no coercion or economic incentive to participate.
  • Using artifacts and connection protocols—Request for clarification and debate by CoP attendees on aspects of the mission statement and terms-of-reference.
  • Maintaining professional interests—Members to be surveyed after each session via an online voting system. Both quantitative satisfaction scores and qualitative comments were sought regarding the presentation and the meeting topic. This information, together with attendance data, would be instrumental in identifying new topics for future events.
  • Improving CoP experience—Future sessions to begin with a presentation on a “hot” topic delivered by an invited speaker or a CoP member acting as subject matter expert. An open voice and messaging Q&A session would follow for a more detailed discussion.
  • Creating CoP awareness—Request to attendees to spread the word through their networks about the launch of the virtual CoP.
Interim Report 3

Managers and workers as participants are powerful collaborators whose commitment to the CoP project will determine its capacity and longevity. The purpose of the third interim report was to provide independent feedback on the early events. We made this report a quarter of the way into the monthly meetings during Phase 4 (Action Taking 2). The rationale for this decision was to allow the convening team time to advertise the new CoP and to have delivered a number of events for realistic feedback.

A mix of data collection and analysis techniques was utilized to produce an overall view of how well the CoP was functioning by membership, event attendance, extent of participation, and indication of required changes to maintain interest. We started our analysis by collating our individual observation notes for an overall view. The convener was able to produce summary data from the CoP data bank which we could use for our membership and CoP attendance analyses. These included number of registered members at Event 1 and Event 4, number of attendees at each meeting, geographical location of attendees, number of meetings attended by employee number (names were not released to us), most well-attended event topic and future topic requests. We revisited events by accessing HPES’ video and audio-recording system to examine the extent of engagement during meetings. A summary participation analysis matrix for the four events was produced. It showed the average number of attendees, number of contributions, and duration of active engagement by discussion topic and number of intervention by the convening team.

We were also able to conclude from our video analysis and participation matrix that topic discussions were fully intelligible between the attendees and discussant from the very beginning, indicating common professional interest, language, and understanding. The conversation flow was becoming more fluid and open with time, requiring very limited intervention by the convening team. Finally, we interviewed three “regular” participants (who attended at least three meetings) and three “irregular” members (who had registered their membership but attended no more than two meetings) to garner their views on all aspects of the online CoP, including what they had liked at the events, what improvements or changes were needed, and if they felt that they were (already or becoming) a part of the broader PMO community.

Phase 5: Evaluation and Specifying Learning

The CAR process was able to serve the dual purpose of supporting action taking by the business partner (HPES) while allowing the research institution (Cranfield) to derive new knowledge of practice. By the end of the CAR project, HPES had taken action and full ownership of the virtual CoP. We had also been able to learn and conceive our theory on the possibility of establishing a community of busy, dispersed professionals—first, from the relevant literatures and then from the study evidence. We were in a position to produce our final evaluation and learning report.

HPES’ expressed motive of staff welfare over return-on-investment was significant for the project’s evaluation strategy. Our third interim report of the intervention was predicated on the three key success and longevity criteria of membership, attendance, and participation. Using a similar range of data collection and analysis techniques as the interim report, we evaluated the membership by checking the list of the LinkedIn group within the CoP and produced a visual graph of rising membership numbers to demonstrate growing interest. Membership was at 150 shortly after the inaugural meeting. Six months on, the system log showed 230 of which 180 can be considered as “active” and by the close of the CAR project, membership was >750 with a number of logged requests to join from the wider project and program management community. This broadening appeal of the CoP to others within the organization is a further indication the CoP was seen by many as a useful platform for engagement.

We then reviewed event attendance for all the events. Using the system log which showed joiners and leavers automatically and additional summary information produced by the convener from the data store, we were able to estimate average number of members at each event (rising from 30 to >50) and who stayed on throughout the Q&A session.

Participation was chosen as an indicator of engagement beyond mere membership or attendance. The format of each meeting tended to be a presentation on the tabled topic followed by an open voice and messaging Q&A session. Issues and opinions were polled using an online voting tool to encourage broader input from the community. Once again, we were able to identify highly rated topics for future scheduling and to indicate participant behavior during events by the number of times the convener had to intervene via a participation analysis matrix.

In sum, our analyses indicated that the intervention by HPES had been successful and was potentially long-lasting.

An important element of this final phase is our ability to generate theory from the empirical evidence. In this study, it was the notion that a non-situated, enduring CoP could be engendered by way of a deliberate intervention. We were able to surmise the emergence of such a functioning community by referencing our observation notes against Wenger’s (1998) community characteristics of shared styles and discussion, ongoing participative activities, mutual relationships, personal identification with the group, and continuous professional learning (see Table 1 in Lee-Kelley & Turner, 2014). Aside from revisiting events and observational data, we also contacted active members for their individual assessment of the benefits and effects of that community, their intention of continued association, and suggestions for the future directions of the community. Responses to the question of improving performance and enabling capability development were positive although we were unable to show a “before” and “after” CoP impact on performance as project activities were not linked to community development.

Lessons for Future CAR

  • Connecting theory with practice: Our first “takeaway” from the case is that when researching the “real” world, a carefully crafted research strategy can bridge the theory–practice gap. Our summary of the scoping meeting shows that senior management (as the sponsor of the study) had a view of the problem and had decided that a CoP would resolve the problem. However, to produce results of true practical and theoretical value, it is important that the a priori assumption is “tested” by the CAR team to ensure (a) the business problem and research question are explicit and accepted by all participating parties, (b) a research strategy is developed that incorporates independent verification of the state and scale of the problem, and (c) they are allowed adequate time and access to carry out the research.
  • Choosing the inquiry team: Another lesson is the composition of the inquiry team. CAR researchers must gain the trust of their business partner by demonstrating adequate industry experience and subject matter knowledge. Given the dynamic nature of the corporate intervention and the dispersed history of those under study, it was vital that participants were not only identifiable and accessible but also that they were willing and candid in their responses.
  • Maintaining independence in collaboration: This case illuminates the difference between a hands-off, third-party research and an in-company, collaborative project. In CAR projects, the institution can find itself working to a given solution by their business partner. Crucially, provided the business partner is satisfied that they would receive the requisite advice and support from the academic researchers, the latter should be given a free hand to select an appropriate methodology for the inquiry.
  • Designing and conducting the CAR project: A further lesson relates to the mixed-methods design of the inquiry process. A cross-sectional, mono-method approach (e.g., via a survey, or observation or interview) is inadequate for exploring and elaborating why already experienced and busy project professionals would want to register as a remote member of an online CoP. Furthermore, the participating organization was expecting regular feedback from the study team to make improvements to their CoP. Researchers and managers using CAR should design a process that will adhere to the established AR principles of committed sponsorship, extended exposure, phased inquiry, close-quarter observation, periodic evaluation, and report for improved learning and practice.
  • Evaluation of the CAR project: A criticism by a reviewer on a paper submitted for publication was the lack of objective metrics to measure performance improvement. Researchers should certainly take on board this valuable feedback when designing their future CAR project. The emergent nature of the case CoP owing to the organization’s staff welfare focus and the lack of access to a corporate directory of PMO personnel meant that it was not possible to define success by the size of the membership. Although the team found alternative ways of evaluating the founding of the CoP, including providing qualitative evidence on the emergence of a community and reporting a number of innovations through within CoP sharing, it was not possible to assess the direct impact of CoP socialization on individual performance or project outcome. We evaluated the CoP formation only by reference to the theoretical characteristics highlighted in the literature. An important lesson for the team here is the realization that we should have worked harder to negotiate access to “hard” data to improve the content and depth of our CAR evaluation.
  • Academic sharing: Another important consideration for the researchers is subsequent publication and impact of prior agreement that permission of the organization would be required but not unreasonably denied for dissemination of the case findings (e.g., Lee-Kelley & Turner, 2017; Lee-Kelley, Turner, & Ward, 2014). Future CAR projects should account for the duality of goals in any business-academic research: Industry would want a practical solution to some problem, whereas the research results must be capable of theoretical extension and publication.


The aim of this case study is to increase appreciation by the research community of the need to close the research–practice gap. Business organizations allowing access to their corporate intervention efforts will want the results to have practical significance. The ability, therefore, to craft a research strategy that can support industry’s need for useful and actionable solutions yet preserving academic rigor is a subject deserving attention. Our case study demonstrates that CAR can be a viable strategy for studying dynamic interventions close-up. Post-project reflection of the project has yielded some valuable lessons that will inform future collaborative research.

Exercises and Discussion Questions

  • Collaborative action research (CAR) is a variant of action research (AR). How would you distinguish CAR from AR?
  • Think of a change project in your organization. How would you collaboratively develop a CAR project for studying it?
  • How would you decide the collaborative roles and contributions for the CAR project?
  • How would you decide the duration of the research?
  • How should the collaboration team be chosen?
  • How would you decide what research methods and analysis techniques would be appropriate for the change project?
  • What are the possible shortfalls for your CAR project?

Further Reading

McNiff, J. (2013). Action research. Abingdon, UK: Routledge.
Reason, P., & Bradbury-Huan, H. (Eds.). (2013). The SAGE handbook of action research: Participative inquiry and practice. London, England: SAGE.
Creswell, J. W. (2014). A concise introduction to mixed methods research. Los Angeles, CA: SAGE.
Plano Clark, V. L., & Ivankova, N. V. (2015). Mixed methods research: A guide to the field. Los Angeles, CA: SAGE.


Baskerville, R. L., & Wood-Harper, A. T. (1996). A critical perspective on action research as a method for information systems research. Journal of Information Technology, 11, 235246.
Brown, J. S., & Duguid, P. (2001). Knowledge and organizations: A social practice perspective. Organization Science, 12, 198213.
Koliba, C., & Gajda, R. (2009). “Communities of practice” as an analytical construct: Implications for theory and practice. International Journal of Public Administration, 32, 97135.
Lave, J., & Wenger, E. C. (1991). Situated-learning: Legitimate peripheral participation. Cambridge, UK: Cambridge University Press.
Lee-Kelley, L., & Turner, N. (2017). PMO managers’ self-determined participation in a purposeful virtual community-of-practice. International Journal of Project Management, 35, 6477.
Lee-Kelley, L., Turner, N., & Ward, J. (2014). Intentionally creating a community-of-practice to connect dispersed technical professionals. Research Technology Management, 57, 4452.
Lewin, K. (1946). Action research and minority problems. Journal of Social Issues, 2(4), 3446.
Lindkvist, L. (2005). Knowledge communities and knowledge collectivities: A typology of knowledge work in groups. Journal of Management Studies, 42, 11891210.
Ma, M., & Agarwal, R. (2007). Through a glass darkly: Information technology design, identity verification, and knowledge contribution in online communities. Information Systems Research, 18, 4267.
McNiff, J. (2016). You and your action research project (
4th ed.
). London, England: Routledge.
McNiff, J., Lomax, P., & Whitehead, J. (2001). Action research in organizations (Routledge Studies in Human Resource Development). London, England: Routledge.
Reason, P., & Bradbury, H. (Eds.). (2008). The SAGE handbook of action research (
2nd ed.
). London, England: SAGE.
Robinson, V. M. J. (1993). Current controversies in action research. Public Administration Quarterly, 17, 263290.
Silva, L., Goel, L., & Mousavidin, E. (2009). Exploring the dynamics of blog communities: The case of metafilter. Information Systems Journal, 19, 5581.
Susman, G. I., & Evered, R. D. (1978). An assessment of the scientific merits of action research. Administrative Science Quarterly, 23, 582603.
Thompson, M. (2005). Structural and epistemic parameters in communities of practice. Organization Science, 16, 151164.
Vaast, E., & Walsham, G. (2009). Trans-situated learning: Supporting a network of practice with an information infrastructure. Information Systems Research, 20, 547564.
Wenger, E. (1998). Communities of practice: Learning, meaning and identity. Cambridge, UK: Cambridge University Press.
Wenger, E. C., McDermott, R., & Snyder, W. M. (2002). Cultivating communities of practice. Boston, MA: Harvard Business School Press.
Wenger, E. C., & Snyder, W. M. (2000). Communities of practice: The organizational frontier. Harvard Business Review, 78, 139145.
locked icon

Sign in to access this content

Get a 30 day FREE TRIAL

  • Watch videos from a variety of sources bringing classroom topics to life
  • Read modern, diverse business cases
  • Explore hundreds of books and reference titles