6
STUDYING GLOBAL WORK GROUPS
IN THE FIELD
STANFORD UNIVERSITY
GEORGE MASON UNIVERSITY
Together and separately, we have been studying geographically and globally distributed work teams for over a decade. Fortunately, our work (and the work of others) has evolved considerably over that time and we enthusiastically share in this chapter some of the methodological insights we have had and experimentation that we have tried along the Selecting the research team for these way. Our focus in this chapter is on qualitative field research in real organizations. We are not talking about studies using quantitative methods (which open up a whole different set of issues), nor about laboratory experiments or studies of student teams. All of those are valid methods (and between us, we've used them all), but we focus here on qualitative field studies designed to build theory about global work.
Why does research on global teams warrant its own chapter in this book? In our experience, the study of global teams presents unique methodological challenges in four ways. First, by definition, team members are not located together. If one wants to study the entire team, one cannot use traditional methods of going out to a single field site and remaining there for the duration of the study. Even if one travels to multiple sites, it is impossible to be in both (all) places at the same time, thus observations are not conducted at exactly the same time (or phase) in a project and cross-site comparisons become suspect. Second, on global teams, team members at different sites are almost guaranteed to have quite different perspectives. Members at one site may feel insulted and devalued as a result of their distant team members’ behavior without anyone at the distant site being aware of it. If one wants to conduct team-level research on the dynamics of global teams, we believe it is imperative to understand the perspectives of team members at all sites. Third, globally distributed teams are generally culturally diverse. This isn't unique by itself, but coupled with the geographic distribution, it becomes more challenging because team members are not only from different cultures, they are embedded in different cultures. Taking culture into consideration is particularly difficult because there is so much variation within a given culture and there is interpenetration of the influence between culture and other factors in the workplace. Finally, to understand intercultural collaboration, researchers are inevitably hampered by their own cultural lenses. As a result, the research team and the data collection approach must be sensitive to these cultural nuances.
We describe the method that we have developed over the last 5 years to study globally distributed teams. It is not perfect and we talk about the challenges later in the chapter, but this approach yields extremely rich data that we believe captures relatively well the experience of the teams and team members we study. Using these methods, we have gained insights into the dynamics of intercultural collaboration and adaptation on teams (Cramton & Hinds, 2007, 2009), the role that site visits play in global work (Hinds & Cramton, 2007), how emotionally charged language issues can derail a team (Beyene, Hinds, & Cramton, 2005), what factors affect team members’ sense of having power and influence on the team (Hinds & Cramton, 2008), and how teams deal with the pressure to design global solutions by reconciling local differences (Xu & Hinds, 2009).
Our method is in many ways traditional qualitative research (ethnographic, if we're permitted to use that term loosely), but with our own pragmatic spin. We conduct interviews in situ by traveling to the locations of our informants and talking with them. We are generally hosted by a local manager who has agreed to invite us into the site, provide us with badges and work space, and get us acquainted with the location. We also conduct observations of team members as they go about their day-to-day work. Where our method departs from traditional ethnographic research in organizational behavior is in: (a) our use of a multicultural research team to conduct field interviews and provide insight into the data collected; (b) our conduct of concurrent observations in which we have a researcher at each of the sites during the same time period; and (c) the length of time we spend in the field, which is necessarily short due to the resources required to travel to and remain in multiple countries.
Although one might study global work at the individual level (e.g., people's experience of collaborating with distant colleagues) or dyadic level, our approach is to identify project teams that are split across at least two locations and focus on the dynamics of the team. Our interview protocols focus on team dynamics and our observations are explicitly focused on the social dynamics on the teams.
We have conducted one large field study together. This included 12 software development teams split between two locations: three teams split between the US and Germany, three split between Germany and India, and six split between India and the US. We interviewed 181 informants during a four-month period, observed six of the teams for the next four months, then returned the following year to interview a subset of the team members again. This was the big project on which we “cut our teeth,” but we have continued to conduct field research over the last few years. Pam, for example, led a study of global product integration activities in six teams spread across the US, Germany, Sweden, India, and Australia. She is also in the midst of another study focused on design practices in different regions of the world. The insights and anecdotes we present in this chapter reflect the intersection of all of the studies on which we have worked together and separately. We hope that you enjoy reading it as much as we enjoy doing the research.
A General Description of the Method
Selecting the research team
Selecting the research team for these in-depth, long-term field studies is one of the most important decisions to be made. We knew, for example, at the beginning of our joint project that we needed to have research team members who could speak the local languages and understand the cultures of those countries in which we would be conducting research. This is in keeping with the embedded intergroup relations perspective championed by Alderfer and his colleagues (e.g., Alderfer, 1987; Alderfer & Smith, 1982) in their studies of race relations. Alderfer argued that composing teams of researchers that encompass the diversity of the groups being studied helps to establish a more empathic understanding of the underlying group dynamics. In the cross-cultural domain, we projected that working as a multicultural research team would increase the likelihood of our being included in sensitive, culturally nuanced communications and arriving upon culturally balanced understandings. In our joint research project on global teams, we had an idea, based on our initial conversation with the field sites, that Germany and India were likely countries for the teams we would be studying. We were extremely fortunate to find outstanding graduate students from Germany and India and a third “multicultural” student who had lived in multiple countries and spoke many languages. In all cases, we strive to compose our research teams so that they mirror the cultural backgrounds of the team members in our studies.
The cultural mixes in our teams are imperative and orchestrated. The personality mix, however, is equally important and can be harder to orchestrate. What makes it work? Deep interest in culture and understanding others’ perspectives, thirst for adventure, adaptability (to time zones, ambiguity, food, etc.) are all necessary ingredients. In our studies, the research team has to live and travel together for an extended period of time. We also have to collaborate together across our own cultural differences as we collect and interpret data from the field. Traveling together is stressful, we are often exhausted (but invigorated), and a resilient team with a good attitude makes an enormous difference to the experience and to the quality of the data that are collected.
Selecting (and being selected by) the field site
Selecting the right field site is also key in this type of work. This is true of any organizational field study (as described in Pratt & Kim, this volume), but it is especially so in studies of global teams because there is less time on-site to build rapport and negotiate access due to the duration of the visits and cultural differences make the probability of a cultural blunder much higher. The groundwork has to be in place prior to arriving so that the time in situ can be used most productively. In most cases, the companies that we have studied have approached us asking for our insights about globally distributed work, so we have had the good fortune to work with receptive field sites. In our experience, studies go most smoothly if there is an insider who acts as host at each site and can ensure that the research team knows how to get to the field site, has access to the building, has a desk at which to sit, has access to the internet, knows where to meet people, etc. Some of these conveniences may seem minor, but it is possible to spend an enormous amount of time and energy sorting out these details if there is no insider to help.
As with all field studies, the commitment at all levels of the organization opens doors and makes the research possible. In advance of our visits, we make sure that teams are identified and contacted by the informant granting access to the organization. We then contact the team leaders and make sure that they understand the study and are committed to having their team participate, and find out if there are any issues or concerns about which we should be aware. Building rapport, as much as possible, using email ahead of time paves the way for our visits. Finally, we try to schedule interviews prior to arriving on the site.
We have developed a simple scheduling application that allows participants to schedule and reschedule themselves in the interview appointments that we have made available. We found that this is enormously helpful because there are inevitably changes that are needed and we have more certainty that the schedule reflects the actual availability of informants. We also found that getting access to the organization's internal scheduling system ahead of time can make a huge difference. We can then send invitations through the internal company system which legitimizes our requests and makes it easier for informants to accept our requests. One little snafu with this, however, was that we once forgot that we were in Europe and sending invitations to team members in India who were on a different time zone. We, for example, made available a 10am appointment, but the calendaring system automatically converted this to 1:30pm Indian time because it assumed that we would remain in our respective locations. We ended up with about 30 percent of our interviews scheduled 3.5 hours later than we expected during one of our visits. This would have been fine except for the fact that we were using two systems concurrently, one that adjusted (the company's internal calendaring system) and the other that didn't (our home-grown interview scheduling system). In future projects, we learned to use a single system and be more aware of the automatic time-zone conversions embedded in today's advanced calendaring systems.
A final point about the field sites is about our responsibility to the organizations we study. Again, as with most field studies, there is a significant cost to the organization in hosting us and allowing us to interview and observe team members. We generally offer to present our preliminary insights to the organization multiple times in order to enable participants in multiple time zones to hear about our findings. We have found that written reports do not get read and have little hope of affecting change in the organization, so presentations, generally via a system like WebEx, seem to do the trick. We also try to return to all field sites and make in-person presentations of our results. This not only fulfills our responsibility to the informants, but allows us to enter into dialog with them, find out what has changed since we were last there, and obtain their insights about our preliminary findings.
Conducting interviews in situ
A key element of our approach to the study of global teams is to talk with people. We are generally interested in team dynamics, including attitudes, perceptions, and behaviors. To obtain information about how people feel about their team dynamics, their perceptions of and concerns about how the team is working together, and their attributions about the sources of challenges and opportunities, we have to talk with them. We use unstructured interviews (see Spradley, 1979), but spend a great deal of time researching the prior literature, identifying questions of interest, and thinking about how we might follow up on, or probe, thoughts interviewees express in order to learn more (see more on this in Pratt & Kim, this volume). We do this for several reasons. First, it helps us to think more deeply about how to approach the interviews. Second, as a research team, it forces dialog about our research goals and how this translates into the interviews themselves. Third, it helps us to identify potentially inappropriate questions or approaches for the cultures we plan to study. When preparing for our interviews in Germany, for example, our German research assistant informed us that asking about the nature of the “relationships” between team members in Germany would be tantamount to suggesting that they were intimate with each other. We quickly changed our wording to “relations” for that setting.
With unstructured interviews of members of global teams, we worry about having our research team calibrated and avoiding systematic bias based on the researcher conducting the interview. We deal with this in several ways. First, we always begin our studies by interviewing in pairs so that we can get a sense of others’ interviewing styles. We generally conduct 3–4 interviews (or more, if needed) in pairs and discuss the interviews afterward to calibrate with one another. Second, we plan carefully for who will be interviewing which team members in which locations. So, for example, in our joint study we considered having one or two team members interview all members of a team. That would have had the advantage that those researchers would know the project, the language, and have a sense of the issues pertinent to that team. Despite its potential advantages, however, we did not do this because it makes cross-team comparisons potentially difficult due to concerns about systematic differences in the interviewing styles, interests, and interpretations of the interviewer. Similarly, having interviewers segmented by site or country has the advantage of being logistically much less complicated (and less expensive), but also makes it difficult to know if any differences by site are merely a result of the perspectives of different interviewers. Instead, we opt whenever possible for the more complicated and resource-intensive approach of having all researchers interview members of multiple teams and at multiple sites, often in the course of an intensive data collection trip.
On occasion, we have also had an insider accompany us on the interviews. On the whole, we have found this to work extremely well because the insider knows the language of the organization (e.g., acronyms, projects, people, etc.), can ask more sensitive questions that we may not know to ask, and can reaffirm the importance of the study to the organization. In Pam's study of global product integration teams, for example, our four-person research team included two insiders from the research laboratory in the field organization. The insiders worked in tandem with Pam and a PhD student in the data collection as well as the early data analysis and were responsible for making the results actionable for the organization.
Armed with an interview protocol and probes, a well-calibrated team, and a schedule, we sit down with each informant in a private office or meeting room (or on some occasions in break areas or a corner of a cafeteria), and proceed to talk with them about their experience of being a global team member. Although we always have a broad research question that shapes our study design (e.g., who we talk with, number of sites, cultures represented, etc.), we allow the informant to use our open-ended queries to communicate what they think is important. In Pam's study of global integration teams, for example, we started with questions about the role of the informant in the team and the goals of the team. We also generally asked about the extent to which there were different requirements for the project in different regions and how compromises affected the way things were done locally. We use our protocol as a guide, varying the order in which questions are asked and dropping questions if an informant has seized upon a particular topic and is showing us a new dimension of the situation. In our joint study of global teams, however, we wanted to be sure to ask about cross-cultural learning, so we had a single question at the end of the protocol that we usually posed ten minutes prior to the end of every interview. The interview could proceed in any direction, but we always brought it back at the end to ask if they had learned anything in the process of working with people from the other countries as represented on the teams. This achieved a nice balance between the two objectives of finding the particular informant's deep knowledge and getting a well-rounded picture of a particular critical topic.
Finally, on the logistics side, we always record the interviews (with the permission of the informant, see Pratt & Kim, this volume, for more on Institutional Review Board considerations) and we always use two recorders just in case of equipment failure. On numerous occasions we have had batteries die, found that we neglected to turn on a recorder, had the adjustments set so that the recording was poor, accidentally deleted a recording, etc. So far, using our overly conservative method of dual recorders, we have managed to acquire at least one high-quality recording of every interview we have conducted. Back in the days of tape recorders, we had to carry around suitcases filled with tapes and we always had the completed tapes carried by two researchers on different flights to avoid possible loss. We now use DAT (digital audio tape) recorders, download our files at the end of each day, and send our DAT files nightly through a password-protected website to our transcriptionist who can start transcribing our files right away. A good transcriptionist will also provide immediate feedback about the quality of the recordings, problems with accents, background noise, etc., so that we can immediately correct any problems. This approach, however, has not been without problems. In Pam's recent work in China, we discovered that our secure transmission website was blocked by the Chinese government and the bandwidth available in the hotel was such that it took hours to upload a single file. Our attempts to transmit our files regularly to our transcriptionist were therefore stymied. Instead, we downloaded files on to external hard drives for safe transport back to the US.
A final comment regarding interviews is that language is inevitably a challenge. Even if the lingua franca of the organization is English, we quickly learned that informants may be uncomfortable expressing themselves, may not be as forthcoming, and may be unable to understand the subtleties of the questions when interviewed in English. In planning the interviews, we try to make sure that, even if the informants are fluent in English, they have the option of being interviewed in their native tongue. We then translate the interviews after transcribing them. Pam has also experimented with having an interpreter present in the interview, but this is not as straightforward as it may seem. Nonsimultaneous (sequential) interpretation refers to cases where one person at a time speaks, so the interviewer asks a question in English, it is interpreted into the interviewee's native language, then their response is translated for the interviewer, and so forth. Such an approach poses several difficulties for ethnographic interviews. First, the interview takes far longer, so much less ground can be covered in the time available. Second, and most importantly, it can disrupt the flow of the interview and the ability of the interviewer to build rapport with the informant because the informant and the interpreter end up talking directly with each other while the interviewer is the “odd man out. “ Simultaneous interviewing holds much more promise, but it requires a highly skilled interpreter who is able to “fade into the background” so that the interviewer and informant have the feeling of talking directly. Such expertise, however, can be hard to find and tends to be expensive, thus pushing the costs of the project even higher.
Conducting concurrent observations
The second prong of our data collection strategy, and perhaps the most exciting for us, is conducting concurrent observations at distributed field sites. Concurrent observations refer to observing and recording the activities of team members at multiple sites during the same time period. Although observing sequentially at different sites is valuable for understanding the perspectives of all team members, concurrent observations enables us to see how events and perceptions are similar and different during exactly the same time period. Thus, we are assured that the differences we are seeing are not the result of different project phases, or reorganizations or external events that occurred between observation periods. During the observations, we take detailed field notes paying particular attention to what is salient, the rhythms of the day, the information team members at each location have, and how cross-site meetings are prepared for, engaged in, and interpreted at each site. We record how events that affect both sites are interpreted and handled. During observation in the US of a US–Indian team, for example, Catherine mentioned to a US-based team member the Indian subgroup's impending move the next day into a new building. The US team member had not known his colleagues were moving and wondered aloud whether this then explained why they had not been responding to his emails that day. During another observation period of our joint project, a massive reorganization, the largest in the history of the firm, was announced. At the headquarters location in Germany, this was a dominant topic of conversation and speculation during the observation week. At the Indian site, however, this topic never came up in conversation among the team members that Pam observed. Toward the end of the week, she asked the local team leader and he said that he knew about it, but didn't want to stir up the team, so he didn't discuss it with them. As a result, we were able to learn about how differently this reorganization was handled at the two sites and to gain insight into how the information flows varied.
One of the ways that we manage concurrent observations is to have regular communication between the locations being observed. We try to communicate to one another the topics of conversation that seem to be important so that the observer at the other site is aware of these issues and can make sure to notice their presence or, as interesting, their absence. We also notify each other of important cross-site meetings that will be taking place so that we can observe these from both locations. Finally, although we rarely have time to read them while on site, we have a practice of sending completed field notes to the entire research team at the end of each day. Continual updating across sites is a way of keeping the research team calibrated and making sure that we are capturing comparable sets of field notes.
There is, of course, a practical challenge in conducting team-level observations. The main one, we think, is that there are multiple people of whom to keep track. Rarely do our informants conveniently remain in a single location (even the sub-teams that are collocated) so that we can observe them all at the same time. We handle this by stationing ourselves at a desk as close to the team members as possible and wandering around to pick up on any action that is occurring. Oftentimes, we are observing knowledge workers, so they are sitting at their computer screen for a significant portion of the day. We ask the team members we are observing to invite us to meetings and we stroll the halls looking for social interaction involving the team members we are studying. In some locations, this is quite easy. India, for example, tends to have a work culture that invites interruption and frequent discussion. Germany, in contrast, tends to be more “heads-down” work with scheduled conversations. As a result, we found it more challenging to get insight into what people were doing in Germany and had to resort to different methods. For example, while observing the German members of a team housed in several adjoining offices, Catherine set up shop in different offices each half day and learned to read body language concerning when German team members were receptive to questions and conversation.
In preparing for our observational work, we anticipate having far more to observe than we can ever capture in our field notes. Although we want to remain open to surprises, we find that determining the level of the observation a priori (at least provisionally) is essential to ensure that our distributed research team takes similar approaches. We make decisions, for example, about the priority of capturing individual work vs. social interaction, within team interaction vs. interactions outside of the team, leader vs. team member behavior, and whether or not to capture general rhythms or focus on the minute details of day-to-day work. Our field notes are inevitably a mixture, but setting priorities ahead of time helps to focus us when we have to make choices in the field about what to capture (see Pratt & Kim, this volume, for further discussion of observation strategies).
Analyzing the data
Once the data are collected, we begin our analysis. Although we send our DAT files to our transcriptionist nightly, we rarely see the transcripts before we return home and never before we go into the field the next day. This presents a challenge to ethnographic research, particularly grounded theory building, which relies, in part, on ongoing analysis of the data and adjustment of the research strategy and sampling based on the insights that are emerging (Lofland, Lofland, Snow, & Anderson, 2005; also Charmaz, 2006; Strauss & Corbin, 1990). When we only have 1–4 weeks at each site, we simply cannot approach the study in this way. To help alleviate this (if only a little), we hold meetings with the research team each day and talk about what we are learning and what we are seeing. Based on these meetings, we get a sense of the themes that are emerging and are able to talk with the members of the research team from the local culture about their interpretations. We continually adjust our interview approach based on these meetings. In our joint study of global teams, for example, we noticed early in the interviews that there were numerous informants talking about language issues. One of the informants in Germany, for example, told us about how stigmatizing it was to ask her management's approval to attend English training because she was expected to be fluent in English as a prerequisite for the job she held. As a team, we talked about this interview and were more attentive to probing if similar issues surfaced in other interviews.
Once we get home and start receiving interview transcripts, we begin our analysis by open coding (see Strauss & Corbin, 1990). We generally have multiple members of the research team open code all of the data so that we have better insight into what themes are robust across readers (especially readers from different cultures). We then generate team-level summaries by having one researcher read all of the material for a single team. The summaries typically include a description of the team, the dynamics of the team, and insights about particular themes identified in open coding (e.g., power dynamics, communication patterns, etc.). Working with our German research assistant, Catherine checked for cultural bias in interpretation by conducting and comparing independent reviews of all the materials from several of the teams, particularly those teams that appeared to experience the greatest cultural tensions. From that point, we begin delving deeper into the data and iterating with the literature to develop grounded theory on topics that emerge.
Strengths, Surprises, and Challenges
Strengths
There are several strengths of this method. One is that we are building on a method that is well developed. Although we have our own spin on it, the foundation for what we do has been established elsewhere by management scholars (e.g., Barley, 1986; Orlikowski, 2002). There are classes that students can take on ethnographic methods, books that researchers can read (see Pratt & Kim, this volume, for suggestions), and symposia that deal with the opportunities and challenges of the method. Further, the method is an established approach in many fields of study, so gaining acceptance is not problematic. More importantly, the method yields incredibly rich data. We find that we have more data and insights than we can possibly manage after a study of the type we describe. We learn an amazing amount and, because of the method, we trust that what we learn has meaning.
Surprises and challenges
There have been numerous surprises and challenges in data collection, data analysis, and writing papers based on the method we describe. One obvious challenge is that it requires a great deal of traveling and careful planning. Although traveling has the ring of adventure, it takes time away from family and other obligations, so it predictably takes its toll on all members of the research team. Globetrotting also brings with it jetlag, unusual food, medical considerations (e.g., malaria medication, shots, etc.), and the inevitable missed flight or lost luggage. In a nutshell, it is exciting, but it is also stressful and exhausting. Finally, with regard to travel, it is extremely expensive, so one must have a large research budget to do what we describe. In addition, the carbon footprint is significant and the cost of transcriptions, translations, interpreters, and so forth, mounts quickly. The bottom line is that it is a resource-intensive approach.
Another challenge is the intensity of the work on site. Because of the short duration of our stays at each location, we try to accomplish as much as we possibly can while there. This translates into long hours of intense work. Catherine, for example, ended up with tendinitis as a result of writing and typing observation notes in the field during our joint study. There is a delicate balance to be reached between the resources expended for this type of research and the burden on the researchers. Each additional day of observation requires, for example, another day on site (and away from home) for at least two researchers.
Another challenge is when concurrent observations breed identification with the subteam being observed. As with most ethnographic work, identification with members of the field site is to be expected (Lofland et al., 2005). This means that, in the case of concurrent observations, research team members may become identified with the people or subteams at the particular site they are observing. We prefer to have observers at a given site who are fluent in the local language, and, as a result, observers are also likely to either be natives of or have a particular affinity for the country in which they are conducting observations, thus exacerbating the likelihood of identification. Tension between the distant subgroups being observed also can intensify identification on the part of the researchers at each site. Because the researcher is seeing the world through the eyes of the particular site she or he is observing, she or he can form attributions about the distant team members in the same way as the team members she or he is observing. Although this may generate tension within the research team at some points, we also recognize it as an opportunity to better understand team dynamics, consistent with the notion of “parallel process” in the work of Alderfer and Smith (Alderfer & Smith, 1982; Alderfer, 1987; Smith, Simmons & Thames, 1989; Smith & Zane, 1999).
This happened to Pam and one of our research assistants in our joint study. Pam observed in the US and our German research assistant observed in Germany during the same week. Pam found herself defending the US-based team against the criticism of the research assistant who had observed in Germany and the researcher assistant defended the seemingly aggressive behavior of the Germans to Pam. Once we realized what was happening, we were able to use ourselves and our internal dynamics to better understand the perspectives of the different parts of the team.
The stability of the teams is also a challenge in organizational research. In our joint study, we were surprised (perhaps naively) by a massive reorganization that took place mid-way through our study. Several of the teams that we were studying were no longer intact and we were not able to (easily) return to continue our research with them. We believe that reorganizations are particularly challenging for team researchers, especially those hoping to do longitudinal research on intact teams, because the teams become disrupted. Fortunately, in our joint research project, we had more or less finished the first wave of data collection when the reorganization occurred. As a result, it affected our ability to obtain longitudinal data, but not the first phase of interviews and observations. We were also lucky because we had two organizations in that study, one of which did not undergo reorganization. Having a second organization, it turns out, is a good risk-mitigation strategy for longitudinal field studies of teams.
One of our biggest challenges in all of this work is the massive amount of data that we collect and analyze. Because we are studying teams, and not individuals, our sample is at the team level. So, we may have only 12 teams (our joint study) or six teams (the study of global product integration teams), but that amounts to 15–20 times that number in individual interviews. In addition, because we risk getting into the field and finding that a team does not match our requirements, we err on the side of larger samples, further increasing the number of interviews required.
We have been told that we do too many interviews for a grounded theory method because we interview beyond saturation, but we feel that we do not have a choice because we need to do all of the interviews while we are on the road. It just is not feasible to interview, analyze, and go back to India the next week if we have more questions or need more material. We could, of course, follow up with telephone interviews, which we do on occasion, but time-zone differences, language and accent considerations, reorganizations, and the difficulty of building rapport over impoverished media can make this less appealing. Grounded theory really demands iterative analysis and our method reduces opportunities to develop and test theory along the way. We see this as the weakest link in our approach.
In terms of the most frustrating aspect of our research, it is without question the amount of time it takes to analyze and write papers based on the mass of data that we have. It is not unusual for it to take three, four, or five years to submit a paper from our data. Reading data, coding, recoding, summarizing, analyzing, reanalyzing, and writing dominates our research time. When we are fortunate enough to be asked to revise and resubmit a paper, it can take a year or more to recode and reanalyze the data and revise the paper. In all seriousness, given how research activity is rewarded, at least in the US, we would not recommend this style of research to someone working toward tenure. For all of its rewards, it is time-consuming and risky. In addition, it is rare that a granting agency will fund multiple years of analysis work, so the analysis and writing ends up, ultimately, getting squeezed into smaller and smaller amounts of available time as we feel pressure to find funding.
A final challenge is fulfilling our responsibility to the field site. Organizations generally are not happy waiting three to five years to get feedback on the studies in which they participate. We have found that a presentation following the open coding stage seems to work well for all involved. Our insights at this point may not be deep and penetrating, but they hopefully provide some new knowledge for the organization and the informants. We sometimes follow up a year later with additional insights and try to stay in touch with the organizations, but the preliminary feedback is when we find that our organizational contacts are most likely to hold the same positions and the organizations are most receptive to hearing our results.
Appropriate and Inappropriate Uses
The method we describe is appropriate for any study in which a researcher cannot observe the entire group in one location. This includes distributed teams, global teams, and cross-organizational teams where the goal of the study is to understand the dynamics of the team and thus it is vital to understand the perspectives of all team members. As with much qualitative work, it is best for theory building, not theory testing. It requires a significant commitment from the research team and from the organization, although the number of teams can be relatively small. Our studies have not yet spanned more than four sites and we have only conducted concurrent observations with teams split between two locations. As the number of locations increases, it may not be practical to employ concurrent observations and it may become cost prohibitive to visit all locations. Thus, where consistent with the research goals, limiting teams to fewer locations is preferred.
Opportunities, Latest Innovations, and Ethical Issues
In this chapter, we have described what we have done thus far in our research. As the study of global teams is relatively new and evolving, so is our approach. Although we have yet to leverage these opportunities, we envision several ways to improve upon these methods and address some of the challenges. First, there is an obvious opportunity to leverage local researchers who have language skill and cultural insight. Although we have not used this approach to study a globally distributed team, Pam is currently working with researchers in China and The Netherlands in a similar way. She has developed research relationships, for example, with faculty and industry researchers in China and is jointly conducting cross-cultural research. These types of research relationships have great potential to enable studies of globally distributed teams with research teams that mirror the characteristics of those being studied. In such arrangements, however, we advise that the research team conduct some joint data collection activities prior to working in a distributed fashion so that the team can calibrate its efforts. We think that some overlap between researchers and sites helps in detecting and minimizing biases. For data analysis and interpretation, it is important that each researcher have a good understanding of all contexts in which the research is being conducted.
Another opportunity that we have yet to integrate into our studies of global teams is a diary method. Diaries involve informants keeping track of specific behaviors, thoughts, or feelings over a period of time specified by the researcher. Sometimes they involve triggers (such as emails, phone calls, or pages) for the informant to record a particular behavior or answer a question provided by the researcher. So, for example, a diary on a distributed team might ask team members to record every interaction they have with a distant team member or the researcher might email the informants asking them to describe their biggest challenge that day, depending on the goals of the research. Diary methods enable the researcher to acquire records of informant behaviors, attitudes, and thoughts over a longer period of time than it may be feasible to observe. Combined with interviews, diaries can be a powerful method for collecting detailed behavioral data (Conrath, Higgins, & McClean, 1983). Members of a team can be invited to keep logs of their behavior for a specified period of time so that the data collection period, as in the concurrent observations, is held constant. Diary methods are challenging because the researcher must compel the informants to complete the diaries and must specify precisely what is desired, but if done correctly, may be a less resource-intensive way to gather data during the same time period from multiple perspectives on global teams.
There are several ethical considerations in our research on globally distributed teams. One, which we have already discussed, is the resource requirements. We believe it is important to explore less resource-intensive approaches to gain the same (or nearly the same) insights. Diaries, for example, as we discussed above, is one possible way to increase the amount of behavioral data collected with less travel. As with all ethnographic studies, we also feel we have an obligation to represent the interests of the informants to the organization. This is not always straightforward because organizations have competing interests, for example, not paying for team members to travel when our findings suggest that it is necessary to healthy team dynamics. Still, our ethical responsibility includes gaining access to executives to make sure that our informants’ voices are heard.
In Closing
Our goal in this chapter has been to describe how we approach research on global teams in a way that enables us to gain empathy for the team members at all sites. In developing our research, we think it is important to give the considerable time that is necessary to produce a generalizable scholarly product that remains faithful to the complicated reality that our informants encounter in their day-to-day work lives. Our research suggests that trying to study global teams without visiting the distant sites is likely to be biased. Although there are many challenges, we find this research to be incredibly rewarding. We gain a deep understanding of the perspectives of team members and the dynamics of these teams and we have yet to embark upon a study where we were not surprised and excited by what we learned.
References
Alderfer, C. P. (1987). An intergroup perspective on group dynamics. In J. Lorsch (Ed.), Handbook of organizational behavior (pp. 190–122). Englewood Cliffs, NJ : Prentice Hall.
Alderfer, C. P., & Smith, K. K. (1982). Studying intergroup relations embedded in organizations. Administrative Science Quarterly, 27, 35–65.
Barley, S. R. (1986), Technology as an occasion for structuring: evidence from observations of CT scanners and the social order of radiology departments, Administrative Science Quarterly, 31, 78–108.
Beyene, T., Hinds, P., & Cramton, C. (2005). Language challenges in international work: The impact of uneven proficiency in the lingua franca. Presented at the Academy of Management Meeting. Honolulu, HI.
Charmaz, K. (2006). Constructing grounded theory: A practical guide through qualitative analysis. Thousand Oaks, CA: Sage Publications.
Conrath, D. W., Higgins, C. A., & McClean, R. J. (1983). A comparison of the reliability of questionnaire versus diary data, Social Networks, 5, 315–322.
Cramton, C., & Hinds, P. (2007). Intercultural interaction in distributed teams: Salience of and adaptations to cultural differences. Academy of Management Best Paper Proceedings, Philadelphia, PA.
Cramton, C., & Hinds, P. (2008). Being in the right place with the right people: How influence dynamics are disrupted in offshore collaborations. Paper presented at the European Group for Organizational Studies (EGOS), Amsterdam.
Cramton, C., & Hinds, P. (2009). The dialectical dynamics of nested structuration in globally distributed teams. Academy of Management Best Paper Proceedings, Chicago, IL.
Hinds, P., & Cramton, C. (2007) Situated knowing who: Why site visits matter in global work. Presented at the Academy of Management Meeting, Philadelphia, PA.
Lofland, J., Snow, D., Anderson, L., & Lofland, L. (2005). Analyzing social settings: A guide to qualitative observation and analysis. Belmont, CA: Wadsworth Publishing.
Orlikowski, W. (2002). Knowing in practice: Enacting a collective capability in distributed organizing, Organization Science, 13(4), 249–273.
Smith, K. K., Simmons, V. M., & Thames, T. B. (1989). “Fix the women:”An intervention into an organizational conflict based on parallel process thinking. Journal of Applied Behavioral Science, 25(1), 11–29.
Smith, K. K., & Zane, N. (1999). Organizational reflections: Parallel processes at work in a dual consultation. Journal of Applied Behavioral Science, 35(2), 145–162.
Spradley, J. P. (1979). The ethnographic interview. Belmont, CA: Wadsworth Publishing.
Strauss, A., & Corbin, J. (1990). Basics of qualitative research: Grounded theory procedures and techniques. Sage Publications.
Xu, C., & Hinds, P. (2009). How do differences matter? Toward a theoretical framework of subjective diversity in work groups. Presented at the Academy of Management Meeting, Chicago, IL.