Abstract
Scientists and academics increasingly work on collaborative projects and write papers in international research teams. This trend is driven by greater publishing demands in terms of the quality and breadth of data and analysis methods, which tend to be difficult to achieve without collaborating across institutional and national boundaries. Yet, our understanding of the collaborative processes in an academic setting and the potential tensions associated with them remains limited. We use a reflexive, autoethnographic approach to explicitly investigate our own experiences of international collaborative research. We offer systematic insights into the social and intellectual processes of academic collaborative writing, identifying six lessons and two key tensions that influence the success of international research teams. Our findings may benefit the formation of future coauthor teams, the preparation of research proposals, and the development of PhD curricula.
Keywords
There has been a universal shift in nearly all branches of science since the middle of the 20th century insofar as teams—rather than individuals—are increasingly dominating the production of impactful published works (Börner et al., 2010; Wuchty, Jones, & Uzzi, 2007). In fact, almost all original research papers are now produced by multiple authors (cf. Whitfield, 2008). In line with this, there has been a notable recent trend of increasing academic collaboration, in terms of the internationality and size of collaborative research projects, for a number of reasons (Ellis & Zhan, 2011; Jones, Wuchty, & Uzzi, 2008; Straub & Anderson, 2010; VandenBos, Jamison, & Fernandes, 2011; Walsh, 2011). Due to tenure and promotion requirements, the general pressure to publish has significantly increased (Adams & Thornton, 1986; Petry & Kerr, 1982), and this has led to a growing demand for large, internationally relevant data sets (Jordan & Henderson, 1995), often made widely available through the Internet. Collaborative research may offer a solution to meeting the requirements of A-level journals in terms of novel approaches to problems and high theoretical and technical standards (Barnett, Ault, & Kaserman, 1988; Mohrman & Lawler, 2011; Peterson, 2001, 2009). The increased visibility that comes from being part of a coauthor team with well-established researchers is an additional attraction to engage in collaborative research (Beaver & Rosen, 1979; Rynes, 2011), as is the potential to develop complementary skills. In addition, an effective division of labor facilitates fitting in research alongside teaching and administrative duties (Belanger & Brockman, 1994). Collaborative research with international peers may also be viewed as more fulfilling and intellectually stimulating (Belanger & Brockman, 1994; Peterson, 2009; Thomas, Tienari, Davies, & Meriläinen, 2009). Perhaps most importantly, collaborative research contributes to the cumulative growth of international scientific knowledge, with an ultimate aspiration to improve the quality of research through collaboration (Rynes, 2011). Indeed, Collins (1998) argued that long-lasting research findings are always developed in social entities, such as networks, intellectual groups, or rivalries (see also Mohrman & Lawler, 2011).
The importance of such collaboration has been acknowledged by a number of scholars and a variety of research streams shed light on different collaboration-related issues in organizations, teams, and networks (see for example, Daft & Lengel, 1986; Emerson, Nabatchi, & Balogh, 2011; Hackman, 2011; Monge & Contractor, 2003; Olson & Olson, 2001; Wilson, O’Leary, Metiu, & Jett, 2008). However, apart from advances in the creation of communities of practice, or epistemic communities, in academia (e.g., Knorr Cetina, 1999; Wenger, 1998, 2000), university–industry collaborations (e.g., Ambos, Mäkelä, Birkinshaw, & D’Este, 2008; Shani, Mohrman, Pasmore, Stymne, & Adler, 2008), the science of team science (e.g., Börner et al., 2010; Stokols, Hall, Taylor, & Moser, 2008), and the role of star scientists (e.g., Zucker, Darby, & Armstrong, 2002), we know relatively little about the team processes and dynamics associated with academic research collaboration. This is unfortunate because, although there is ample evidence to show that diverse perspectives and backgrounds have the potential to produce better results in teams, these potential diversity gains may easily be neutralized by process losses unless processes are managed well (Hackman, 2011; Stahl, Maznevski, Voigt, & Jonsen, 2010).
In fact, academic research may be particularly challenging in terms of collaboration compared with other types of teamwork such as corporate teams: researchers tend to be independent, autonomous thinkers (Busse & Mansfield, 1984) who are difficult to manage (Mintzberg, 1998; Stephan & Levin, 1992). Academic collaborations are often loosely coupled with evolving structures and objectives. In this regard, managing large academic author teams resembles an exercise in herding cats (Hellawell & Hancock, 2001). 1
We—management scholars and academic writers—primarily study how others work rather than how we work ourselves. The team of authors who wrote this article are all members of a network that aims to initiate new research collaborations and exchange pedagogical ideas. When we first got together in Uppsala, Sweden, in November 2010, we focused on just that: how we collaborate as researchers. For the purposes of this article, we use a broad definition of collaboration: “a process through which parties who see different aspects of a problem can constructively explore their differences and search for solutions that go beyond their own limited version of what is possible” (Gray, 1989, p. 5). We posit that collaboration dynamics are interesting because they prompt us to question some taken-for-granted routines of our daily lives (see Garfinkel, 1967), such as writing papers together. We use a reflexive, autoethnographic approach to examine how international collaborative author teams in management research work together and achieve research outcomes. In a world where management studies using in-depth approaches such as ethnography are still rare (see Chapman, Gajewska De Mattos, & Antoniou, 2004; Rosen, 1991), the highly standardized methods and practices for producing academic knowledge and journal publications (e.g., Gopen & Swan, 1990) may prevent us from questioning our own assumptions relating to the dynamics of how we work together and what kind of social and intellectual processes we are involved in. We contribute to the literature by identifying six learnings and two key tensions that affect the collaborative process and its outcomes. We consider our primary audience to be management scholars who work in international collaborative research teams.
The article is structured as follows. We first review literature on collaborative academic writing. We then introduce our methodology of reflexive autoethnography, which combines autobiography and ethnography, to observe three large collaborative research projects. Autobiographic writing (Ellis, 2004; Ellis, Adams, & Bochner, 2011; Ketelle, 2004; Reed-Danahay, 1997, 2001) is a “narrative inquiry” in which the author(s) embark on “theoretically structured introspection” (Fiske, 1990, p. 85) by retrospectively and selectively writing about past personal experiences using hindsight (e.g., Bruner, 1993; Schultze, 2000). The article presents consecutive stages of reflection and introspection as they evolved chronologically and concludes with insights from the field that help us structure and exemplify steps in collaborative processes.
Patterns of Collaborative Writing
Collaborative research draws on a combination of the different academic skills and subjectivities of individual group members in interpreting the field and research results. A collaborative research process includes several different social processes including (a) division of labor, (b) psychological support, and (c) the construction of meaning through intragroup dialogue (Antal & Richebé, 2009). To our knowledge, the intersections between these three processes are underexplored.
Starting from the first set of processes, if a collaborative research project simply divides tasks among individuals, the purpose may be purely to economize resources. Specific competences of individual researchers are used for different parts of the research process such as research design, data collection, data analysis, and writing. Senior researchers may undertake tasks that require experience, whereas junior researchers may engage in tasks that are more labor intensive.
Ede and Lunsford (1990) surveyed 1,400 members of seven professional associations such as The American Consulting Engineers Council and The Modern Language Association, and identified seven patterns for the division of labor among collaborators (see Appendix A). This survey suggests that several factors determine patterns of collaboration: how team members organize their participation in the project (sequentially or in parallel) and whether team members work independently or interdependently at various stages of the project (planning, draft writing, revising, and submitting). Planning or idea generation takes place in teams in about half of all collaborations, whereas in the other half, one individual is responsible for the first idea and for organizing the collaboration. Thus, in the drafting stage, either one team member writes the entire draft, or several team members produce parts of it (which are then compiled by an individual or a group); there is no interdependent work. Where one person is responsible for planning the project, other team members become active in later stages, namely, writing the draft or providing feedback and revising the paper. This suggests that the advantages of diversity are mainly realized in the planning and revising stages rather than during the writing, which seems to be the most independent task in research collaboration. The fact that the team/group wrote the draft together in only 3% of the projects is a good indication of how difficult collaborative writing is.
Such a division of labor may, of course, not serve all the needs of individual contributors. Some researchers may seek environments that provide encouragement, support, and inspiration. More importantly, a division of labor does not in itself necessarily contribute to the quality of the final product. Instead, creative dialogue between researchers combined with interactions in the field that include a construction of meaning is what enables the research team to contribute (Lee & Bozeman, 2005). As Geertz (1979) explained it, researchers are
hopping back and forth between the whole conceived through the parts which actualise it and the parts conceived through the whole which motivates them, we seek to turn them, by a sort of intellectual perpetual motion, into explications of one another. (p. 239)
This hopping back and forth between one’s own position and the position of others in the designing and publishing of joint research is a process of creating collective meaning that may result in a product of higher conceptual quality than what any of the participants could have achieved alone (Lee & Bozeman, 2005).
Collaborative Writing in an International Context
International collaborative research can sometimes be seen as a specific response to the call for division of labor in the form of replication studies across cultural contexts (Boyacigiller & Adler, 1991; Peterson, 2001, 2009). It allows research teams to tap into specific cultural knowledge by accessing local collaborators for data collection (Boyacigiller & Adler, 1991; Peterson, 2009; for example, GLOBE study; House, Hanges, Javidan, Dorfman, & Gupta, 2004). Whether they are to produce replication or original studies, international collaborations tend to involve several cultural contexts (Easterby-Smith & Malina, 1999; House et al., 2004; Von Glinow, Drost, & Teagarden, 2002), in which no single researcher can be an insider (Boyacigiller & Adler, 1991). Despite these benefits, a number of challenges are likely to unfold during the different phases of a research project. Based on their experience with an international human resource management (HRM) best practice research consortium involving researchers from 18 countries, Teagarden et al. (1995) identified four successive phases in the life of a research consortium: (a) forming the research consortium, (b) generating questions and constructing a survey, (c) doing the research, and (d) making sense of findings. These authors found that developing clear ground rules such as authorship and data ownership guidelines helped the consortium’s writing process. Trust was developed most rapidly through “working sessions in which each member could see each other’s abilities and work together to resolve challenges” (Teagarden et al., 1995, p. 1271). Interaction was concluded to be the sine qua non for overcoming relational issues, including a shared vision, which was important for constructive collaboration.
We agree with Teagarden et al.’s (1995) argument that a core issue in doing research is to align research objectives with the right approach and methods and to make sense of findings, and in international research teams, the challenge of alignment is likely to be manifold because it means bringing together members’ potentially very different interests, backgrounds, and perspectives. In other words, the differences between international research teams and domestic ones have to do with complexity—more diverse interests, cultures, and backgrounds; more different institutional environments and objectives; and issues of distance and face-to-face interaction. In other words, potential issues become amplified when more complexity is introduced.
Thomas et al. (2009) called for more reflexive analysis of collaborative research to understand its dynamics, and Karra and Phillips (2008) proposed that “researchers need to learn to see their own culture, and then they need to translate what they see for publication in international journals” (p. 542). Reflecting on cross-cultural research, Easterby-Smith and Malina (1999) made a case for a more reflexive approach to ongoing research collaboration. By working together and observing each other in action, international research teams not only build trust but also come to understand and appreciate different approaches to doing research. By extension, they develop a more reflexive view of their own research practice, which may previously have been taken for granted. Although many of these insights are often only crystallized and formulated in the aftermath of a project, the authors suggest that facilitating conscious and open dialogue during the project would provide a “space in which to exchange expectations, assumptions and feelings” (Easterby-Smith & Malina, 1999, p. 84).
Learning From Our Own Field
We adopted a reflexive autoethnographic approach that seeks to generate “theoretically relevant descriptions of a group to which one belongs based on a structured analysis of one’s own experience and the experiences of others from one’s group” (Karra & Phillips, 2008, p. 547). Reflecting on one’s own actions has been recognized as a factor that positively influences not only individual but also team performance (Pieterse, van Knippenberg, & van Ginkel, 2011; West, 1996) as it can be seen as the basis for learning. Reflexivity has been explicitly studied, discussed, and encouraged in various research contexts (Alvesson, Hardy, & Harley, 2008; Easterby-Smith & Malina, 1999; Geertz, 1973; Hammersley & Atkinson, 1983; Petriglieri, Wood, & Petriglieri, 2011; Polanyi, 1958), including international research teams (e.g., Tartas & Mirza, 2007; Thomas et al., 2009).
Autoethnography refers to “an approach to research and writing that seeks to describe and systematically analyze personal experience” to examine and problematize established rules and patterns of doing research (Ellis et al., 2011, p. 1). More specifically, autoethnography involves a combination of autobiographic writing, in which the author(s) retrospectively and selectively write about past personal experiences using hindsight (Bruner, 1993), and ethnography, which has to do with gaining in-depth understanding of a culture’s values, practices, and systems of meaning (Van Maanen, 1988). Our primary focus here is on research culture (rather than national culture) in the academic author context, referring to the shared practices and systems of meaning that emerge within research teams to achieve desired outcomes. We also blend insights from autobiography, ethnography, and autoethnography while maintaining our focus on “theoretically structured introspection” (Fiske, 1990, p. 85).
By using a reflexive, autoethnographic lens (Ellis, 2004; Ellis et al., 2011), we are able to gain a process view from inside and a deeper understanding of collaboration-in-practice. Schultze (2000) argued that the ethnographer can use his or her actions as a thinking agent who brings his or her own subjectivity to bear on the construction of knowledge in a self-reflexive and self-revealing fashion. This could lead to a critical scrutiny of existing practices because authors as well as readers are constantly challenged to examine their own practices and assumptions while learning about others. A reflexive approach has the potential to yield insights not easily observable from the outside. When we use our own work as the object of our inquiry, our observations and conclusions will have a high degree of authenticity (Marvasti, 2004), although clearly we know more than we are able to communicate (Polanyi, 1958). To understand our own “native” processes, we need to be open-minded, accept the subjectivity inherent to the method (Garfinkel, 1967; Moore, 2011), and anticipate accusations of perceived narcissism (Alvesson et al., 2008).
We use a number of techniques to analyze three collaborative research processes to cover the what, the how, and the why (Lincoln & Guba, 1985). First, we focus on epiphanies, or key moments that are remembered and perceived to have significantly affected the research process (Denzin, 1989). Many of these moments manifest themselves as crises within the collaboration, which can clearly be seen from the descriptions below. Second, these epiphanies have been drawn by the three authors themselves from reflexive, personal narratives—or “confessional tales” (Van Maanen, 1988)—describing how the participating authors’ sensemaking of their respective collaboration process has evolved. Other researchers have valued and drawn from personal documents, such as letters, diaries, journals, auto/biographies, and memoirs (Cooper, 1987; Harrison, 2009; Stanley, 1993; Znaniecki, 1934). Third, we use an iterative approach of layered accounts, alternating and juxtaposing between the authors’ own individual empirical accounts, reflections between the three authors of the project cases, previous literature and theoretical analysis, and critique across the whole coauthor team (Dubois & Gadde, 2002; Ellis et al., 2011).
Three Cases of Collaborative Research
In this section, we provide three cases of international collaborative research to illuminate how the social dynamics and processes intersect. These cases drawn from our own experiences represent different stages in collaborative research processes and different types of projects: a conceptual paper development project, an instrument development project, and a data-collection project. Our objective here is to make sense of our own collaborative experiences in line with the reflexive, autoethnographic approach. As we intended to represent the firsthand view of the researchers involved without “contaminating” the data (Moustakas, 1994), we kept the original voices of the authors who contributed the accounts when writing up the study (Nagel, 1956). Author 1 was the lead author in a paper project of 17 authors, which focused on writing about key scientific insights in the international business field. Author 2 was a key member of a 6-person English–French collaborative research project focusing on instrument development for ethical decision making in organizations. Finally, Author 3 was a core member of a research group of 9 people (expanded from 6 in the early phase) undertaking a large multilevel data-collection project on HRM in multinationals.
Author 1: The “Scientific Insights” Project
Step 1: Prepared brainstorming
As a loosely connected management research group, we gathered for an annual meeting in May 2009 to look at life’s important questions and future trends. Our subgroup of approximately 17 highly competent people (professors from 15 different universities around the globe—spanning multiple regions, cultures, languages, and religions—all with PhDs from a variety of fields, born and trained in different “systems”) discussed everything from globalization to cognitive neuroscience, reaching a discussion of what future research in international business should look like. A few people had prepared by skimming the research literature for essential pieces that would serve as inspiration, including special issues of the British Journal of Management and Management International Review. We anchored our discussions in an article published in the Journal of International Business Studies because we considered it to be rigorous, although burdened with epistemological foundations that we questioned. With this challenging point of departure in mind, we brainstormed about what should be the leading themes in international business studies and why they are important.
Step 2: Structure and assignment of subgroups with timelines
In the weeks that followed, I spent a considerable amount of time trying to structure all the input from the brainstorming (from the large face-to-face meeting) into initial categories that would make sense for the future work which, I presumed, would mainly take place in a virtual setting. The structuring was not only meant to help organize the subsequent group work and division of labor but also would serve as the basis for the writing process. In this process, people pitched in mostly by adding thoughts and their expertise and areas of interest via email, for example:
Hey Author,
Bravo and thanks for taking a substantial lead on this. I have no problems taking on the epistemological debate with anyone who would like to join me and in helping in any other way.
Step 3: Content creation and literature review(s)
Each subgroup (category) consisted of three or four people with one leader/coordinator, and each group was asked to do a literature review and make recommendations—for example, “Are we asking the right research questions?” In this review section, a few scholars who were considered influential were identified, and their work was synthesized and extrapolated (e.g., Adler & Harzing, 2009; Bell, 2010; Van de Ven, 2007). In some cases, dialogues were opened directly with these authors to better understand their work and thinking. That really worked.
The major part of the dialogue took place inside each subgroup, through phone meetings, but primarily by email. Additional dialogues took place between the subgroup and me; I was supposed to have an overall view of the project. I didn’t always, as it was emerging to some extent, but I was also reluctant to communicate the structural mess, as the intention was to sort it out later. This, however, led to some tension along the way, for example:
Hi Author and Nickname
I may be completely off the mark here—Please correct me if I have missed something or misread the article. I don’t think we have achieved what we intended.
Our argument (if I understand this correctly) is that we redefine the “key contributors” and expand the domain in response to the emerging themes. And that this is a valid contribution because otherwise we are “drinking our own Kool-Aid” and missing key trends. The problems that I see with our article are . . .
I was happy that I could filter emails like these so they did not go out to the large community of authors, as this kind of frustration can easily escalate or spin out of control. This specific case was “solved” within 1 week by a third person (coauthor) who performed a stellar piece of “structural repair,” which preserved the content, improved the paper, and led the instigator to comment as follows:
This is why I love working with this group. We allow for constructive conversations!
I look forward to the next round, Cheers.
Step 4: Patchwork, revisions and an internal review process
Six contributors were dedicated upfront as informal “reviewers.” That is, based on individual expertise and skills, experiences, and interests (which I knew by then), people sequentially reviewed the emerging paper, some by focusing on content (i.e., what was missing and what was inadequate) and some on the structure of the paper and the internal logic. This “patchwork” resembled a drugless trip (Glaser, 1978) and it seemed to work rather well.
The revisions were plentiful over the short time frame in which the steps described above took place. The back and forth between initial content providers (subgroups), informal reviewers, and myself was extensive and sometimes exhausting. The discussions were mostly held electronically, using “track changes” for corrections and text suggestions and “comments” for exchanges of meaning, as in the following:
This comes across as sour grapes!
All of these need to be in the form of a research question, not a prescriptive statement.
Both of these comments were discussed back and forth and led to substantial changes in the way that we worked with the paper, as they clarified that we were in the business of creating and questioning, not criticizing or crying out loud. The entire paper was produced in less than 6 months. During this time, I sent out frequent updates to all group members (like a heartbeat and not just related to critical incidents); I believe that this kept most people not only well informed but also interested and engaged.
Author 2: The “Cross-Channel” Project
A few years ago, a Lille-based French philosopher–entrepreneur together with French academic and practitioner colleagues developed an instrument to explore ethical decision making in organizations. One opportunity to connect with English-speaking academics and perhaps to gain exposure to a wider audience came through an accidental introduction at a conference to my French dean who suggested contacting a scholar at his school, Ivystone University. 2 The scholar happened to be a close friend and colleague of mine.
We were already under considerable and increasing research pressure owing to the UK government’s rapidly approaching Research Excellence Framework (REF) exercise. Without any further intellectual investigation, my colleague leapt to the conclusion that an English-language validation of this French decision-making tool might offer a minimum-effort route to a high-quality published article. From the Ivystone side, division of labor was the overriding reason for entering into this collaboration with “persons unknown.” The prospect of a quick publication was undeniably appealing to me, as I felt “behind” on my own publications following a period of ill health. Moreover, I brought with me strong French language skills, a research interest in cross-cultural management, and the possibility of providing psychological support for my Ivystone colleague who was under a heavy administrative load. I also felt tremendous pressure to provide that support in exchange for having been invited to participate in the project.
Step 1: Skipping brainstorming
With the French and English sides together, we lurched into virtual collaboration with no real understanding of the skills, abilities, constraints, resources, motivations, or expectations of one another as a team. In effect, we jumped onto a “moving train,” which means that not only was the brainstorming stage skipped but also any straightforward discussion of the mechanics of the instrument. Our cross-channel collaboration was further complicated by issues of language (English and French sides), workload (English side), intellectual property ownership (French side), and funding (English and French sides).
The first few weeks of our collaboration passed in increasing frustration for us, the English partners. As the original contact with the French side, my Ivystone colleague led further contacts, which were carried out exclusively by email. We had expected swift receipt of the online instrument for translation. The Ivystone lead co-opted a third colleague to begin work on a literature review. This development had occurred in isolation from our French colleagues and was presented to me as a “fait accompli” in view of our short-term funding deadlines, other heavy work obligations, and our desire for quick publication. The response from the French side seemed puzzling. Days might pass before an email response was received and then it didn’t seem to address the content of our message. With our exclusively instrumental focus, worries about wasted time and potential loss of funding began to appear. Was this collaboration worth pursuing? We hadn’t even got to the point of understanding clearly what the instrument consisted of. My Ivystone colleague was ready to give up on the relationship. I suggested that we might be having communication problems, possibly stemming from language difficulties, and that maybe a face-to-face meeting would help to clarify things.
Step 2: Thinking about structure, timelines, and roles
We invited the French side to visit us at Ivystone. Only the sociologist was able to attend. We discovered that he was a “real” academic like us—not an entrepreneur as we had imagined. He was not only associated with one of France’s prestigious research centers but also looking to generate a program of research funding and English-language publications. He also proved to be an extremely likeable person. In short, we “hit it off” and so the seeds of psychological support began to germinate. Things were looking up. We seemed to have hit the jackpot as far as working with these “persons unknown” was concerned.
Although we spoke English at that meeting because my Ivystone colleague is a non-French speaker, the French sociologist’s English was not fluent and I noticed that he missed parts of what we were saying. At times, I translated difficult ideas into French. He suggested that we apply for joint research council funding with the deadline less than 2 months away. This suggestion was a step away from the initial transactional- and task-based relationship to a more transformational and idea-creation approach. Although excited by the possibilities, my heart sank—given my already full workload—at the enormity of the task, especially coupled with the language challenges (and possibly the unspoken expectation that I would write the entire English-language proposal on my own). What an unequal division of labor. Instead of feeling elated, I felt truly overwhelmed by the situation.
My Ivystone colleague seemed oblivious to the challenges. She had, in her current administrative role, little time to focus on the project. She indicated that she hoped that I would carry things forward. In the days following Meeting 1, I realized through a careful reading of the French-language funding documents subsequently written and sent to me by the French side that the decision-making instrument was not a psychometric one, as my Ivystone colleague, a psychologist, had imagined, but rather an action-research one. Hence, a quick online experiment or two was not going to be possible. I also quickly discovered that the division of labor between my Ivystone colleague and me really meant that in practice, I was to handle all the project management. This project was not about team idea creation, development, and writing but about independent work fuelled by a sense of desperation set within a laissez-faire organizational research environment and a highly pressurized external research environment. I was now in the position of letting down a good friend (not to mention a senior colleague) and feeling rather short of the psychological support that might have made the collaboration worthwhile. How could I have let myself become involved in this mess? Should or could I now get myself out of it? A first step seemed to be to get off the moving train at the next station and sit for a while together to decide where to travel as a group.
Step 3: Aiming for meaning creation
With difficulty, I managed to convince my Ivystone colleague that the funding bid was simply too much too soon. We had not produced any joint work to date and we certainly hadn’t constructed much meaning together. Moreover, we desperately needed meaning to move forward. Although disappointed, the French sociologist accepted our decision not to speed up the collaboration and invited the three of us from Ivystone to Meeting 2 in Lille. Although I discussed our misunderstanding of the nature of the tool with my Ivystone colleague on a number of occasions, she was so preoccupied that it was not until she could question the French side about it directly in Lille that she finally listened, understood, and accepted that for the French firm, the focus of the project was on constructing the meaning of ethical decision making itself. Our instrumental focus on division of labor for a quick publication had led to much lost time and energy. At the same time, it opened a new phase in the team’s collaboration in which psychological support became a more significant feature. Importantly, we agreed to the development of a short conference paper led by the French side. A whole-team construction of meaning was still not a focus though it seemed to me that perhaps it should be. The French side was in control intellectually and it seemed impossible to change this. The paper was written by the French side and reviewed individually by the three Ivystone partners. Nonetheless, the paper was duly presented at the chosen conference with partners from both sides in attendance marking the beginnings of the public face of our collaboration. Although I was plagued by worries about the lack of open discussion regarding where the project was heading, I hoped for the best.
Step 4: Toward collaboration at last
Our English-language pilot didn’t happen until the following year, much to the French side’s disappointment. My Ivystone colleagues and I realized that owing to other commitments, we couldn’t realistically continue with the project and continue to let down our French partners. I had long desired a more realistic division of labor, but until this point, my Ivystone colleague had not seemed to want to recognize this, perhaps because no solution to the issue of resources had presented itself. As discussed with our French colleagues (Meeting 4) at Ivystone the following spring, we needed to bring in some more human resources to work primarily on the ask, but also on idea creation (or at least identification). Industry funding would also be necessary to keep us going. Psychological support for the team as a whole seemed to be germinating now with open discussions of workload issues and institutional political realities. Moving toward some collaborative construction of meaning seemed a possibility. The strict cross-channel division of labor also seemed to be lessening. Moreover, I felt myself breathe a sigh of relief.
Author 3: The Multilevel Data-Collection Process
The third example is a 3-year, large-scale, multilevel data-collection process that focused on HRM issues in multinational corporations and was a collaborative effort between two Finnish universities involving the joint efforts of six senior and three junior researchers from various backgrounds and nationalities. The collaboration involved 12 Nordic multinational firms, and the data-collection process consisted of the design and administration of four different questionnaires.
All of us had done previous work on HRM and various issues related to multinational corporations. As an opportunity arose to obtain financing for a major research project, we decided to put together a research plan. We agreed right from the beginning that we wanted to create an ambitious, internationally cutting-edge project and, after considerable reflection, decided that multilevel data collection was the right approach, because it was the best way to examine the issues at hand and because there had been recent calls for rigorous multilevel research in our field.
Step 1: Searching for meaning
However, the clarity soon vanished when we set to work. Our first task was to agree upon a detailed research design. We wanted to start from the end result: tangible paper ideas and frameworks/models. We soon agreed that we needed data from multiple actors in the organizations studied—which meant four different questionnaires. The size of the task felt terrifying indeed, and although the task itself was clear, its details were not. We, the six senior members of the team, all had our own (and rather varying) research interests, and although we all appreciated each other’s interests, we knew we had to produce a much more focused design. We started a brainstorming process in which different subgroups (which were fluidly assigned based on research interests and expertise) first came up with concrete paper ideas and theoretical frameworks in parallel. We alternated this parallel subteam work with frequent whole-team meetings, which at times felt rather frustrating to everyone as ideas competed, were dropped, and resurfaced in what felt like an extremely messy process. Looking back, it was collective construction of meaning and the coming together of different ideas, opinions, and experiences; without such a process, we could not have arrived at a plan that everyone was (reasonably) happy with. In hindsight, even the messiness was probably necessary as the process enabled everyone to buy into the plan and feel ownership. Psychological support within the team was soaring.
Step 2: Creative interactions—and destructions
With a broad plan emerging, we then set off to design the first three questionnaires, one for corporate HR managers, one for subsidiary general managers, and one for subsidiary HR managers. These had to work together because independent and dependent variables had to come from different sources to avoid common method bias. In addition, a forthcoming employee-level questionnaire had to be kept in mind.
The process of initial brainstorming in parallel subteams had worked well in the previous stage, and we decided to continue with it. The task was to turn the tentatively agreed-upon paper ideas into concrete constructs and questionnaire items. When the subteams came back, we realized that we were back at square one. The paper ideas that we thought we had agreed upon had changed, new ones had appeared, and others had evaporated. This “creative destruction” resulted partly from looking back on previous ideas with fresh eyes and realizing that they were difficult to accomplish in practice, and partly from the very process of thinking, which brought new ideas. Nevertheless, we continued to work on the revised and iterated paper ideas, with the goal of devising actual questionnaire items.
Commitment to the companies meant that we had a task to accomplish, and preferably sooner rather than later. We also realized that too many constructs were on the table to fit into manageable questionnaires. As the end result became more real, people became more competitive about “their” items. This had an obvious effect on the psychological support within the team. Despite the process becoming increasingly frustrating, we were able to keep on track, resolve differences, and arrive at a set of questionnaire items. The data collection itself went more smoothly than we expected, partly due to the extensive preparation and planning, but also because of the hiring of three excellent research assistants, two of whom subsequently joined the project as PhD students.
Step 3: Dialogue continues—and tensions arise
Having successfully completed Step 2, we now started planning web-based questionnaires for some 1,200 employees. The design of this questionnaire became an even more painful process in which different interests, concepts, and measures were debated fiercely. A primary reason for this was that the number of questions in this form had to be minimized so as to reach a sufficient response rate; for reasons of validity and data analysis, a low response rate was seen as a crucial pitfall to be avoided. Underlying tensions surfaced and people were becoming increasingly defensive about “their” constructs. The debate could have gone on and on, with one of us in particular pushing hard for a set of questions on a particular subtopic; others (me included) were growing tired of having to discuss endless versions of a long, constantly changing list of constructs and questions. The group dialogue didn’t seem to work as well as it had in the previous phases. The feeling of “we are not getting anywhere” was becoming increasingly strong. People started talking in the corridor about how to get out of the situation, rather than around the joint meeting table. Simultaneously, pressure to launch the questionnaire was intensifying. As the majority was of the opinion that enough is enough, our mind-set quickly shifted from joint egalitarian creation to majority rule. We simply had to complete the task, compromises or not. We agreed that a final, streamlined proposal of the remaining questions would be agreed upon in the next meeting, no matter what; this happened and we closed the questionnaire. The actual data collection went relatively smoothly.
Reflexive Analysis: Sigtuna Group Reflections
The idea to reflect on our own experiences with team collaborations in research was the spin-off of our first meeting as a group in Uppsala in November 2010. In August 2011, we met again in Sigtuna to discuss the three sets of case data during 3 consecutive days. An agenda had been sent out in advance. It included a talk by a guest speaker on the topic of autoethnography. All of the participants also received the data (detailed case descriptions) in writing prior to the 3-day workshop. In the workshop, we collectively discussed and analyzed data in several two-hour sessions during the 3 days, categorizing our observations of the collaborative dynamics in the three cases, juxtaposing the reflections of the case participants against critique, and probing from the larger author team of the paper. We further crystallized the meaning of our case-specific learnings and reiterated with the relevant literature, alternating between dialogue and individual reflection and analysis. We learned several lessons through this process of iteration, some identified based on data from all three cases, whereas other findings were inspired by one or two cases and recognized in the cross-comparison of the cases in the collective team discussion. The lessons were written up subsequently and circulated across all coauthors for a multiple round of revisions in the weeks after the face-to-face meeting.
Through this iterative process, we were collectively able to (a) reflect on and categorize our experience and the collaborative dynamics we had observed and pursued as individual researchers, including how we felt about the ongoing processes; (b) crystallize the meaning of our experiences (cases); (c) engage in explanatory cross-case analysis, juxtaposing the reflections of the original case authors against each other and critique from the larger author team; and (d) iterate the focus of the paper and cross-fertilize literature recommendations. We identified a number of critical issues in the three cases and the literature we reviewed, as we consulted literature in parallel to our analysis and reflections. The essence of our discussion is summarized in our learnings with regard to six process features that particularly influenced research outcomes in the three projects and two key tensions that challenged the team processes. The six process features are presented first followed by the two tensions in the next section:
Team identification was a key component in all projects, understood as the degree to which individual team members identify themselves with the team. Although we acknowledge its importance, we did not “measure” or analyze it per se, yet even long after the teams have been dissolved, team identification still remains in some form. Van Der Vegt and Bunderson (2005) pointed out that collective team identification increases the relationship between expertise diversity and team learning. Such a relationship is important because collaborative projects could be described as “learning by writing.” Moreover, identification is a crucial process that can help mediate the “far-but-close” paradox, insofar that it helps geographically dispersed authors perceive close proximity (Wilson et al., 2008; see also Maznevski & Chudoba, 2000).
We noted how important task interdependence is for the team as a raison d’être. Despite the fact that academic collaborations are often loosely connected, as in the cases we reviewed, interdependence creates identity, mutuality, and reciprocity, all in a positive sense. Interdependence is defined as a state of being in which a person is determined, influenced, or controlled by some other person (DeSanctis, Staudenmayer, & Wong, 1999), and thus it reflects the extent to which group members are dependent on one another to perform and complete their individual jobs and is related to effectiveness of projects (Kirkman, Mathieu, Cordery, Rosen, & Kukenberger, 2011).
The initial brainstorming stages were a collective act, meant to maximize the benefits of variety and creativity, but with an initially unclear and unstated objective. For example, the term a messy collection of ideas was used in the description of Project 3 and was reflected throughout all three project cases. We believe that this lack of clarity early on may be caused by the fact that certain objectives are implicit and tacit, given our striving for publications. Nevertheless, several other objectives can be on the agenda and evolve over time, that is, knowledge exchange, friendship, and teaching practices. Thus, our experiences differ somewhat from the team research literature in which clear objectives and difficult goals often lead to better performance (see the recent meta-analysis by Kleingeld, van Mierlo, & Arends, 2011); on the other hand, initial ambiguity is an inherent quality of academic research, often in contrast with many corporate projects that are—with the exception of R&D—often more defined from the start.
Given the difficulties experienced by the Ivystone side of the cross-channel project, their “jumping on a moving train” with little information on the path to follow led to frustrations and almost to withdrawal from the project. This suggests that the initial stage-setting of a new project is vital, for developing ideas and for structuring the project and building trust. This was the case in the crucial first step of the scientific insights project and partially for the “multilevel data-collection project.”
The narrative about the cross-channel project explicitly mentions the importance of face-to-face meetings. Where the project seemed to be doomed to fail and the Ivystone partners almost withdrew, a personal meeting between the French side and the English partners developed trust and laid firm foundations for developing psychological support. In fact, the personal meetings seem to be related to steps in the progress of the project, suggesting that personal meetings are also helpful for structuring the project in each phase of the process. Furthermore, the data-collection project and the scientific insights project started out with a meeting of the participants during which idea generation was central. This allowed the participants to fall back on trust developed at the first meeting through periods of little face-to-face interaction, perhaps because they already knew each other from past professional and social engagements. This suggests that personal meetings, which persist in the workplace because it is a long-established gold standard for establishing and nurturing human relationships (e.g., Olson & Olson, 2001), are less important for the later stages of a project if the project begins with an interactive meeting and/or if trust has already been established (i.e., when authors have previously worked together; see also Jonsen, Maznevski, & Davison, 2012).
Our experiences 3 seem to support the following: one person was mainly responsible for the processes (i.e., organizing the writing or keeping milestones and deadlines). This suggests a pattern of collaboration where subteams are each represented by a person who has a larger role in the planning, writing, and revising stages. Collaboration is carried out sequentially in which a single individual is responsible for the progress and process of the project. Earlier work on international research collaborations (Ede & Lunsford, 1990) found that some teams use the input from multiple team members in early stages of the project, like in the scientific insights project, whereas other teams benefit from the variety of team members’ expertise in the later stages of feedback, reflection, and revision. However, in all but one of the patterns of collaborative writing reported by Ede and Lunsford (1990; Appendix A), some stage of the project was organized by a single individual: be it the initial idea or the planning, writing a first draft, or putting together the submission. This suggests that research collaborations and writing projects can benefit from knowledge generation in diverse teams in different stages, but also that a clear division of labor (and roles and responsibilities) is preferable in such teams.
Individual Post-Sigtuna Reflections
Following the meeting in Sigtuna and collaborative development of lessons learnt, it was decided that the individual case authors should go back and reflect on their own narratives in the light of what had already been discussed as a group. This evaluation included the degree to which they experienced the six collaborative lessons learnt in their respective projects as having positive, negative, or neutral perceived effects on their project. It was carried out alone, without consulting the other two case authors or any other members of the paper team. In addition to the summarizing table (see Appendix B), we present the reflections of the individual authors below:
Reflections: Author 1
As process owner, I took clear “leadership” in some phases (e.g., structure, timing, and submissions processes) and took a laissez-faire rule approach in others (e.g., how the subgroups were working). We were indeed dealing with an amazing amount of diversity (including informational diversity) in this project. The structure, processes, and division of labor were suggested early on in the project. The aim was to make use of the diversity in content and review processes (continuous improvements), but it also turned out that the paper became a highly cross-level and cross-disciplinary contribution, precisely because the group entailed such a broad range of expertise. Instead of suppressing divergent opinions, we made an effort to keep a “fluid structure” that could entail different viewpoints and thus enrich the paper. Single participants were given the opportunity to gradually internalize some perceptions and views from other participants as well as to provide opportunities for other researchers to internalize parts of their perceptions of the final product. We also discussed the extent of agreement on the fundamentals and those areas in which there was less agreement. Nevertheless, the objective of contributing to “science that matters” was never at risk. As we had agreed on the superordinate importance of making science relevant to practitioners and stakeholders beyond academia, “positivists” could work harmoniously together with “constructivists.” We were all guided (I believe) by a rather humanistic approach to business and research; perhaps we were even committed passionate scholars (Adler & Hansen, 2012; Bourdieu, 2001) with many shared tacit components (Polanyi, 1958). We provided each other with psychological support and encouragement throughout the process, either via face-to-face or phone conversations or through emails.
A lot of the meaning creation took place by e-dialectic efforts and by critiquing each other’s work through reviewing (formalized) or daily e-interaction. We only met once as a large author group. The rest of the interaction was virtual. Thus, I believe it was crucial to manage which discussions should be held between a few people only and what communication should take place more widely. As process owner, I could filter and redirect communication and negotiations. I don’t mean that I acted in a manipulative sense, but rather like a “socially sensitive router.” Disputes were thus solved by dyadic or triadic communication, although admittedly there was a tacit acceptance (across cultures) of the “power” of our informal reviewers to change, delete, or resolve issues. Moreover, I was very much aware of how (especially negative) emotions can be contagious in groups, in both senses, and easily lead to escalation of unwanted conflicts (see Barsade, 2002; Hatfield, Cacioppo, & Rapson, 1994; Jehn, Rispens, Jonsen, & Greer, 2013). I imagined that this could easily happen in a virtual setting as well. In essence, I needed to (a) restrict the skepticism (and negative emotions) to dyads or microenvironments, yet still deal with it, and (b) actively promote an information stream of a more positive or neutral character (typically process updates).
Finally, and returning to using the power of diversity, we applied a mix of brainstorming groups and nominal groups at different times in the process (see McGlynn, McJurk, Sprague Effland, Johll, & Harding, 2004). As the chosen “task” was inherently cross-disciplinary, high on modularity and medium on complexity (specialty insights), we benefited mostly from the diversity by using it in a large brainstorming group early in the process rather than in nominal groups (working individually) later on in the process. It could be interesting to test this approach further, that is, whether diverse perspectives are used in concert early on in a process and divided (by space and time) later on. If we had presented the article to the large group of coauthors and brilliant independent minds midway in the process, it might have taken us two steps forward and three steps back. Using the expertise sequentially in e-dialectic approaches worked for us. Fortunately, although the authors came in with different backgrounds, experiences, research cultures, and thus expectations, the argumentation was never about winning but merely about contributing viewpoints in the spirit of dialectics. Perhaps that is why a learning collaboration like this gives me energy and satisfaction.
Reflections: Author 2
Language and workload issues, coupled with our initial virtual “jumping on a moving train” collaboration, certainly proved significant obstacles to true collective working. We nevertheless persevered. It seemed that processes of psychological support, friendship, and compatibility mixed together with a certain amount of desperation (for an English-language connection on the French side and a business connection on the British side) were essential to keep us engaged. It had taken a year, but we understood each other much better with mutual support developing in spite of everything. With a little luck and work to get our students settled, we could say that we had finally developed something of a real partnership with a sense of team identification at its heart.
After a shaky start, the cross-channel team enjoyed a number of small successes to be proud of and build on at the end of 12 months. Although we had met face-to-face only four times, we had managed to translate eight scenarios into English, submit and have accepted an abstract for a conference, present our first collaborative work at the conference, and conduct a small pilot study with Ivystone executive MBA students. Although the pilot study was small, the results looked promising in that considerable variation was apparent in the reactions to the scenarios. Further work leading to a full academic paper now seemed possible as did extending the collaboration to doctoral students and other researchers and seeking industry funding.
At that stage, I learned the “hard way” through a range of positive, neutral, and negative experiences that attention to shorter term social and operational processes are essential to a prospering relationship. Without such attention, the longer term conceptual and strategic collaboration is likely to flounder and possibly (perhaps probably) collapse. With an eye on the longer term, we had largely overlooked the shorter term without conscious reflection on the interdependence between the two.
Reflections: Author 3
Such a large-scale and labor-intensive data-collection process was a major challenge and a huge learning process—a much more complex endeavor than any of us had undertaken before. The project plan started as a relatively loose and—at least in retrospect—messy collection of ideas that were sold to the participating companies. Like the first example, this was also an emergent and creative sensemaking endeavor, and at the beginning, I certainly didn’t guess (or could not have guessed) where it would end up. Looking back, our “ignorance” was a blessing. Had we been aware of the complexity of our undertaking beforehand, we might never have started it in the first place. Nevertheless, the project was completed and proved a success. When I reflect on it, I think there was a reason. We were all extremely committed to making it work and despite our different interests and working styles, we all took responsibility and overcame the obstacles. This commitment, I believe, was crucial to our success in rising above substantive differences in interests. In fact, I don’t think I’ve ever worked in such a mutually committed team before. We were all willing to acknowledge and accept different perspectives in a process that involved everyone jumping between their own position and those of others, which was significantly helped by all of us being highly committed to the accomplishment of the project; despite the enormous complexity of the task and the differing interests, we felt like a team and the team’s interest came before personal ones for all of us.
Although I note in Appendix B that the lack of clear initial objectives was inhibitive for the success of the project, I’ve also come to think that the at times very frustrating and disorganized dialogue was necessary for the outcome to emerge: Instead of suppressing different viewpoints, we engaged in a fluid construction of meaning that took advantage of diversity in ideas and opinions and enabled a supportive and trustful spirit to grow. Had we had a more effectiveness-oriented approach, it is very possible that the whole process would have broken down in Stage 1. At the same time, all this takes a lot of time, and toward the end, we became more instrumental. This was probably necessary, as it brought the project to its goal—Otherwise, we might still be thinking about those questionnaire items. The process was another enabler. Although many of the above described stages took place in an emergent and self-organizing way, the project as a whole was very objectives driven with clear milestones (such as meetings and reports to the case companies) that provided nonnegotiable deadlines and the necessary discipline. One person took leadership of process issues, making sure that we were on track with the process and deliverables. The creative part, on the other hand, took place in an egalitarian and self-organizing way. The interplay between complexity and differing interests versus ownership and trust on one hand, and being objectives driven but self-organized on the other, was to me the key to the success of the project—not that it always felt that way at the time.
Cross-Case Reflections
Returning to our Sigtuna meeting, our group reflections supported extant research and brought new challenges, which we will briefly touch on. We found that several of the identified themes, such as “team identification” and “task interdependence,” were supported by earlier research. When it comes to “clear objectives,” our findings were not all together consistent with previous research, insofar that objectives may develop and emerge over time, and starting out with predefined goals may hinder more than help. In groups where most people do not know each other’s competencies and capabilities, this may be particularly important; it allows processes that can grasp the unique potential of interaction where this may surface.
Three issues—“stage-setting,” “face-to-face meetings,” and “division of labor and roles”—supported what we know from earlier research, but importantly also provided us with new input into the implications of these issues for academic collaboration projects. Perhaps not surprisingly but less addressed in the extant literature is the “stage-setting” of projects as critical for the ensuing process. This was evident in the “scientific insights project” and the “multilevel data-collection project” in a positive way, whereas the lack of stage-setting seemed to have affected the “cross-channel project” negatively in more than one way. This led us to discuss that international research (or writing) collaborations benefit from early interaction between many members of the team in terms of clarifying the division of labor and building psychological support and safety (see Edmondson, 1999). Related to this is the issue of “face-to-face meetings” where we found, as shown in earlier research, that the lack of such meetings had negative effects. However, we also learnt from the “cross-channel project” that it is not necessarily too late to introduce these, and that they can have turn-around effects as trust is established and objectives become clearer. For the “multilevel data-collection project,” these meetings functioned as critical for making decisions and for taking action when the project had stalled. Seemingly, “face-to-face meetings” can take on many roles and we believe that international research collaborations benefit from face-to-face meetings throughout the project’s stages, in particular if interdependence and trust is not established in an early stage (see also Maznevski & Chudoba, 2000).
For division of labor and roles, our experiences support earlier research, but we also suggest that the collaborative project is a pattern of sequential stages where a single individual needs to be responsible for the progress and process of the project (i.e., organizing the writing or keeping milestones and deadlines). Depending on the scale and scope of the project, the collaboration could also include subteams, which are each represented by a person who may take on a larger role in the planning, writing, and revising stages. Important here is, perhaps, what Richard Hackman labels refereeing (Coutu & Beschloss, 2009), which is a stronger role than that of a convener (who identifies and brings the legitimate stakeholders to the table and helps facilitate discussions; Gray, 1989). Refereeing is often needed in a large group of diverse minds and personalities who are making decisions on the content and structure of a paper. The refereeing can of course be of high or low quality, which can affect the “success” of the paper, yet it should ideally not create process conflicts (Jehn, 1997), thus discussing this early on can be beneficial to the outcome.
Furthermore, on reflection of our own experiences, we suggest that academic collaborations can be reidentified as 4 communities of practice (Wenger, 1998), a specific form of interorganizational networks defined by Wenger and Snyder (2000, p. 139) as “groups of people informally bound together by shared expertise and passion for joint enterprise,” as they share many of the key characteristics: for example, a common purpose of developing members’ capabilities to build and exchange knowledge with knowledge as the primary output (Wenger & Snyder, 2000). We found that the key effects of boundary processes (Wenger, 2000) were prevalent in all three cases. These include coordination, transparency, and negotiability—in particular openness to questioning current practices (Oborn & Dawson, 2010). As we academics investigate “who we are” (a moving target perhaps), we can then begin to learn from a decade of “communities of practice” discourse. For example, Probst and Borzillo (2008) found a number of reasons for “communities of practice” failures: lack of a core group, a low level of one-on-one interaction between members, rigidity of competences, and lack of identification with the communities of practice.
From the reflections of the individual authors on their respective cases, three specific issues emerge. These are seemingly common in our experience, yet relatively unnoticed in the literature. The authors also supplied accounts for how these issues were overcome effectively, which enabled the collaborative process to continue. In the first case, negative emotion and skepticism came into light, and as this can easily lead to and escalate conflicts, it needed to be handled. The author referred to this as being a “socially sensitive router,” not manipulative but restricting negative interaction to dyads or microenvironments while promoting positive collaborative process updates and outcomes. In the second case, a pertinent underlying issue was the lack of support not just from the “other side” but also from the author’s “own side.” This frequently paralyzed the process and caused the author to feel overloaded and alone, and to consider dropping the project altogether. When the author met the other party in a face-to-face meeting, there was trust and renewed hope for the research collaboration. However, for the process to dramatically improve, it needed a realistic realization from a fellow participant that the author could not carry the whole workload. Finally, in the third case, the salient issue was how to handle the situation (and emerging conflict) when research collaboration participants focused on their own research questions, “fighting” to keep certain items in the questionnaire and becoming increasingly reluctant to let go of their favorites. Commitment to the research collaboration was raised as a contributing factor for overcoming these differences, but the looming deadline for reporting and informing the case firms also played a part in intensifying the effort to agree. These three findings—(a) the coordinator acting as a “socially sensitive router,” (b) the equal need for internal and external support, and (c) the effect of internal commitment and external obligations to overcome individual differences—provide us with novel contributions to the literature in addition to what we have discussed above.
Moreover, we found in our discussions that international researcher-team collaboration processes and, possibly, trade-offs among them are influenced by a combination of two specific tensions that interact and create tension: (a) freedom for interinstitutional collaboration and (b) specific experiences of collaborative writing (see Figure 1). The process of working toward a collective product is shaped through a combination of the various experiences people bring and the perceptions and expectations of different international group members regarding the final outcome. These may be gradually aligned through interaction and dialogue among participants. Bourdieu (1990) suggested a duality in which structures in the surrounding world guide our actions on one hand, while our actions maintain or modify those guiding structures on the other: guiding structures being those social norms and values that support individuals in choosing what actions to take and what perception to have of their surroundings. This duality creates a dynamic between the individual and the collective, and between the internal and the external. Such a dynamic is relevant for all research projects and becomes integral for successful international collaborations that are associated with more complexity. We build on this notion by creating a conceptual space in which this tension can unfold.
Degree of collaborative writing experience versus freedom for interinstitutional collaboration
The first item was particularly observable in Cases 1 and 2, in which the members of the research teams had to juggle demands from and comply with the research cultures of their own institutions to a much greater extent than researchers in Case 3 did. In other words, Case 3 researchers were able to dedicate a greater proportion of their time to the project versus other responsibilities compared with members of the other two case projects. Perhaps it is, at least partly, this relative escape from internal tensions that allowed us (and perhaps others) to enjoy so much working with people from other institutions. We reasoned that, apart from gaining diverse perspectives and benefit from variance across institutions, interinstitutional collaboration is satisfying due to a fairly “levelled playing field” (Polanyi, 1958, refers to equal footing and mutual recognition as scientists). In this environment power, boundaries and status are merited primarily in the specific collaborative context (or experience relevant to it), as opposed to intrainstitutional work where existing hierarchy, issues, and power structures often remain and spill over between projects and assignments. Importantly, it has been noted that between-university collaboration is associated with the production of better science (Uzzi, cited in Whitfield, 2008), and that the multi-authorship advantage of higher quality intellectual output disappears when co-authors are from the same institution (Chung, Cox, & Kim, 2009).
The second item refers to differences in writing experience, and it was observed in all three case projects. If some people on a team have little or no (academic) writing experience, and the condition is one mentioned above with “equal footing,” then this may create tension precisely because there is no established power and hierarchy (as is the case in intrainstitutional collaboration) to guide decision making. On the positive side, it may also allow more diverse perspectives to surface and be used compared with intrainstitutional work.
Limitations
Several conditions for effective teamwork have not been explicitly dealt with in this text.
First, task complexity: Tasks that are more complex typically require higher levels of interdependence (McGrath, 1984). However, we have assumed here homogeneity in the level of task complexity for collaborative science.
Second, time: In general, the longer a team is together, the smoother and more automatic its processes become; this is considered useful in reducing conflict (e.g., Jehn & Mannix, 2001). Although our focus was on newly formed teams, this was evident in our third case, in which some of the researchers in the multilevel data-collection project had collaborated before on other projects; strong ties enabled the development of a solid trust base that allowed joint creation of meaning despite challenges. Time also allows heterogeneous teams to create a common identity, which contributes to their performance (Earley & Mosakowski, 2000). It allows a shift in the group’s focus from surface-level characteristics to a deeper understanding and appreciation of group members’ underlying or deep-level psychological characteristics.
Third, team size: As groups grow in size, they often experience increasing problems in many areas, including communication, coordination, lower cohesion, social loafing and free riding, and diffusion of responsibility. These effects, including that of a “bystander” (Latane & Darley, 1970), indeed put extra production pressure on the first author of one of our cases. We did not, as a matter of fact, experience that the contribution of the first author diminishes as the size of the author team increases, rather it remains constant. In addition, it is widely accepted that increased group size can reduce performance and productivity in terms of the quality and quantity of the output (e.g., Mullen et al., 1991), and encounter coordination issues. Although we recognize this, we have not focused our study on these aspects and we also acknowledge that larger author teams (such as the GLOBE study, connecting 170 scientists; see House et al., 2004) may have advantages such as resource allocation and smaller fault-line effects (Thatcher & Patel, 2011).
Fourth, different kinds of teams: The team literature is rich on different kinds of teams and the conditions for their success. In a recent review, Hollenbeck, Beersma, and Schouten (2012) presented an insightful typology of 42 different kinds of teams. Analyzed using a multidimensional scaling approach, the authors find three underlying constructs that differentiate teams: skill differentiation, authority differentiation, and temporal stability. In this article, we have touched upon several aspects embedded in the above, yet we have not anchored our findings in the team literature as such. This approach applies to the growing body of team literature research that deals with virtual teams (see Hinds, Liu, & Lyon, 2012; Jonsen et al., 2012, for recent reviews). Both the degree of virtuality and the richness of the media used can affect team dynamics and performance.
Fifth, team leadership: Recent research on team leadership is rich on examples of plurality, whereby leadership changes and emerges over time and situations, including shared leadership, pooled leadership, and spreading leadership over time (see Denis, Langley, & Sergi, 2012, for a recent review). This line of research analyzing patterns and roles of distributed leadership may indeed influence academic author collaboration, especially because these teams have no formally hierarchically appointed leaders.
Finally, we have focused more on tensions than creation in our collaboration analysis. We believe this may be an unintentional result of using epiphanies, or salient events, whereas much of the creativity and creation may happen in a more evolutionary way. It is worth noting that several of the items mentioned above are important factors when applying diversity research lenses (e.g., Akkerman, Admiraal, Simons, & Niessen, 2006; Stahl et al., 2010), yet not explicitly exploited in this article.
Concluding Reflections and Implications for Future Research
Previous research on collaboration has been useful in helping us identify some of the key issues related to collaborative research and writing. However, when trying to make sense of our own academic collaboration, we found that extant research tends to take a more detached, post hoc look at factors facilitating or inhibiting effective collaboration. The autoethnographic lens adds a rich account of the processes and team dynamics involved, including the events and uncertainties that take place as the collaboration unfolds. The “hanging out of our dirty laundry” leads to a better understanding of these messy realities, and we found this to be particularly informative when trying to make sense of academic collaboration, “backstage” à la Goffman, rather than “frontstage.” We conclude in the paper that our main contribution is to offer systematic backstage insights into academic collaborative processes as experienced by “insiders” and compared with existing research from pertinent fields. The findings have questioned conventional wisdom in a few areas, related collaboration to existing knowledge in other areas, and proposed new ones to consider and explore further. Thus, we aim to bring stimulus to the topic of collaborative research processes in an academic context and hope to be a resource for learning, reflection, and dialogue.
Just as our own starting point as a network of academics was to foster and facilitate new collaborations, we hope it may benefit teams to discuss the aspects we have raised in this article prior to engaging in collaborative processes. We propose that this topic, with its detailed discussions, becomes part of PhD curricula, where students may also themselves apply some of the methods we have used in this article based on their own experiences with group work. Moreover, an added value of this article can be as an integrated part of research proposals, without attempting to be prescriptive per se (e.g., see Peterson, 2009, for recommendations for international project teams). By this, we mean that proposals may be stronger when they include conscious reflections on how the multiple author processes can work and what to consider prior to commencing. It has indeed been suggested that funding agencies may benefit from assessing collaboration factors prior to allocating resources (Whitfield, 2008).
By reviewing collaboration research and using autoethnography (Karra & Phillips, 2008; Scherdin, 2007), we reflected on our own experiences and derived lessons from three research projects that may facilitate and stimulate further discussions among future scholars collaborating in teams. We have focused on how research collaborations operate in practice, an issue on which the literature has been rather silent (Salmi, 2010). Our use of a group-based autoethnographic method seemed appropriate because meaning creation derives from interactive interpretation by multiple persons, not simply from the cognition of a single individual (Schutz, 1967).
Previous studies have pointed out how differences between research approaches and contexts in different countries can affect collaboration across borders (Easterby-Smith & Malina, 1999; Peterson, 2001, 2009; Teagarden et al., 1995). We do not dispute this, but some of our experiences were somewhat different—perhaps because most of the scholars involved were internationally “groomed” and experienced in cross-cultural collaboration. For example, even though the backgrounds and institutional contexts of the collaborators of the scientific insights project varied, all authors had previously written (accepted) papers for the Academy of Management Conference and/or journals of international standing. Throughout the introspection into our own collaboration practices, we foregrounded rather than repressed the diversity of our team, starting by perceiving each other as real “others” and paying heed to dialectic thinking as a logic to expand one’s perceptions (Starbuck, 2006). Such cross-examination between multiple investigators, forms of data, or theoretical perspectives promotes a more informed understanding of the phenomenon and does justice to a more complete representation of different perspectives and viewpoints (Piekkari & Welch, 2011).
The issues and lessons we observed do not, primarily, relate to national boundaries, cultural assumptions, or educational traditions, but rather to group cohesion and social capital, including aspects related to interaction mode and frequency and trust, for example. This finding thus seems to echo a recent study by Mäkelä, Andersson, and Seppälä (2012), who found that social capital can powerfully overcome cultural and linguistic barriers (see also Davenport & Daellenbach, 2011). However, in the case of small and medium-sized teams with geographically dispersed members and variability in international experience (such as the cross-channel project), we see how this may sometimes play a role. In such cases, we recommend that these constraints be offset by one or more early, in-depth, face-to-face exploratory, and expectation-setting meetings of the whole team.
This sensemaking journey of different practices and processes can, we hope, help and guide future scholars by transforming tacit knowledge into explicit knowledge about collaborative writing for international author teams in management. Although remaining mindful of why we do scientific research and for whom (see Jonsen et al., 2010), we hope this article contributes in particular to the how. Nevertheless, we highly encourage more reflections and research on how we collaborate in management research, given that research is often grounded in axiomatic knowledge and assumptions, and at times entails incommensurable philosophical positions.
Footnotes
Appendix
Assessment of Lessons Learned by Case Authors
| Lesson | Author 1 | Author 2 | Author 3 |
|---|---|---|---|
| Team identification | Positive | Negative | Positive |
| Task interdependence | Positive | Negative | Positive |
| Clear initial objectives (i.e., explicitly stated ones) | Negative | Positive | Negative |
| Stage-setting | Positive | Negative | Neutral |
| Face-to-face meetings | Neutral | Positive | Positive |
| Single individual responsible for process | Positive | Neutral | Neutral |
Authors’ Note
The authors wish to thank the three anonymous reviewers and Professor Helen Sword for helping us improve this manuscript, and FAS (The Swedish Council for Working Life and Social Research) and Uppsala University for making this project happen.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.