This research was made possible with funding provided by the Carnegie Bosch Institute for Applied International Management Research, the University of Wisconsin Initiative for World Affairs and the Global Economy, and the National Science Foundation Grant #SBR 96-31748. The researchers would like to acknowledge the time and effort extended by all of the respondents in this research together with their associated staff. Special thanks to Rauol Zapata in Puerto Rico, Joylie Agustin and Ricardo Lim at the Asian Institute for Management in the Philippines, and Michael Segalla at the Hautes Etudes Commercials in France. We would also like to express our appreciation for the administrative support, translations, and transcriptions provided by Paula Bassoff, Ryan Billingham, Florence Brunell, Joan Donovan, Kerry Jung, Francisco Lloveras, Rachel Ritterbausch, David Robinson, Carol Troyer-Shank, and Richard Zapata at the University of Wisconsin. 
INTERCULTURAL ANALYSIS OF THE MEANING OF TEAMWORK: 
EVIDENCE FROM SIX MULTINATIONAL CORPORATIONS 
 
ABSTRACT 
The meaning ascribed to the concept of "team" across cultures was investigated using interview transcriptions based on structured interviews conducted with over 125 team members. Teams represented six multi-national organizations and four cultural regions. One facility of each company was included in each cultural region. Working inductively, we used comprehensive textual analysis to identify common schema for teamwork in each culture and each organization. Statistically significant differences in frequency of occurrence of schema were found across cultures and across organizations. Propositions are posed concerning the formation and effects of teamwork knowledge schema. Implications for managers in multinational organizations are also delineated.  Key Words: Teams, Cognition, Schema, Multinationals 

Function: Management Regions: U.S., Latin America, Southeast Asia, Western Europe 

OVERVIEW 

The use of teams in educational, humanitarian, and business organizations is increasing. Findings from several large scale research programs suggest that team-based systems help organizations of many different types respond to performance pressures for speed, cost, quality and innovation (Lawler, Mohrman & Ledford, 1992; Osterman, 1994; Mohrman, Cohen & Mohrman, 1995). Evenso, these research programs have failed to address the interests of international organizations with facilities in several countries. Organizations face special challenges when implementing team-based systems across their global facilities. Intercultural theories of organizational behavior suggest that teams must be implemented in a manner that is consistent with the cultural context in which each of the international organization’s facilities are embedded; however, to date, empirical research investigating this process is non-existent. 

This paper begins by describing the theoretical framework, methods, and preliminary results for a four-year research program designed to investigate the implementation of teams in multinational organizations. The underlying objective of the research is to develop a contingency theory of team effectiveness which takes into consideration the cultural context. Key foci in developing this model are social cognitive processes in teams, including those social processes that are related to the acquisition, storage, transmission, manipulation and use of information for the purpose of creating a group-level product or service (Larson & Christensen, 1993). The four year research program investigating team implementation at various geographical facilities addresses three broad objectives: (1) explore the team system currently being employed at each geographical facility; (2) identify the sources of knowledge utilized by teams in each facility; and (3) delineate factors influencing variation in team effectiveness at the various facilities. 

The first year of the research program involved a series of intensive exploratory interviews with over one hundred individuals representing fifty-nine teams. Members of six major pharmaceutical and medical product companies were interviewed: Merck, Inc., GE Medical Systems, SmithKline Beecham, Kodak Health Imaging, Pfizer, Inc., and Johnson & Johnson. The teams are located across four geographic regions: the U.S., France, Puerto Rico, and the Philippines. The researchers traveled to each country and met face-to-face with each interviewee. A pre-established series of questions were posed pertaining to the function of the teams, teamwork knowledge, the management of the teams, and the context in which the teams work. Interviewees were also asked to discuss which factors they felt were the most important facilitators and inhibitors of team effectiveness. Working in an inductive mode to develop theory, interviews were content analyzed to determine central themes. These analyses were then utilized to develop specific propositions and instrumentation to investigate these propositions quantitatively in subsequent phases of the research. 

In discussing the results obtained in the first year of our research program, we focus here on one of our most intriguing preliminary findings. Going into the project, like others before us, we made certain assumptions regarding how the topic of "teams" would be received across cultures and organizations. We assumed, for example, that people would know what we meant when we used the words "team" and "teamwork." Stated another way, we assumed that there would be basic agreement regarding the meaning of "teamwork" in each of the facilities we visited. 

Our observations during the first year of our research lead us to question these assumptions. Members of multinational organizations reported that "teams" have become a pervasive element in many locations across the world; however, there was little agreement regarding what the concept of teamwork represents. Contrary to our belief that the team concept would be fairly universal, we found important and systematic differences in the ways our respondents defined teamwork. Because the teamwork concept is the most fundamental building block for our research program, we carefully examined three issues before moving on to the quantitative phases of the research: (1) how did team members define teamwork? (2) why did they define it this way? and (3) what does this imply about social cognitive processes in teams? The results of this inquiry are reported here. 

USE OF WORK TEAMS ACROSS CULTURES. 

The use of work teams within organizations has increased dramatically over the past decade. Research conducted in the early part of the decade (Wellins, Byham & Wilson, 1991) suggested that only about one-quarter of the surveyed organizations were using teams, involving only a small portion of the workforce. More recently, a study by Osterman (1994) determined that 54.4 percent of U.S. organizations have at least some of their employees working in teams; 40.5 percent of U.S. organizations have over half of their workforce involved in the team effort. Additional evidence suggests that this trend will continue to accelerate. When the 313 organizations in a comprehensive study conducted by Lawler, Mohrman, and Ledford (1992) were asked how they planned to use teams in the future, 60 percent responded that they would increase or greatly increase their use of teams; 37 percent said their use of teams would stay the same, and only 3 percent said they would reduce or discontinue the use of teams. 

For the most part, this increase reflects a belief that teams are an appropriate mechanism for implementing strategies formulated to deal with performance demands and opportunities presented by the changing business environment. Organizations are experiencing dramatically increased pressures for performance. They are being required to develop and deliver products and services at lower costs but with higher quality and increased speed. A great deal of the research and popular literature proffers teams of various sorts as an appropriate response to these performance pressures for speed, cost, quality and innovation (Mohrman et al. 1995). Much of the literature on quality management, for example, recommends teams to make improvements in organizational processes (Deming, 1986; Juran, 1989). This recommendation is based on the understanding that organizational processes cut across organizational departments and that a process cannot be optimized without examining it in its entirety (Davenport, 1993). Teams are also a favorite design choice in the literature examining speed, cycle time, and time-to-market (Stalk & Hout, 1990; Wheelwright & Clark, 1992; Myer, 1993). And finally, literature on the processes of innovation and learning points out that innovation occurs when different perspectives and knowledge bases are joined within teams, resulting in the reframing of problems and solutions that would not have been likely or possible within one perspective (Kanter, 1983; Pinchot, 1985; Senge, 1990). 

In addition to the trend of increasing use of teams, the United States is also part of an increasingly global economy. About 100,000 American organizations operate overseas and it is estimated that a sixth of the nation’s jobs are derived from international organizations (Cascio, 1989). International facilities must operate outside their local environment and must compete against unfamiliar players. The level of ambiguity and risk increases, and there is a growing need to adjust organizational practices to these new environments. While a number of programs of research have investigated the implementation of work teams in the U.S., a key facet of team implementation that has been ignored are the special concerns that face international organizations as they design team-based systems in facilities located in multiple countries. 

Only a handful of studies have compared the dynamics of teams in the U.S. to team dynamics in other countries. Furthermore, it should be noted that much of this research has been conducted in the laboratory or with temporary student groups. A few notable exceptions have examined work teams within an organizational context across different cultures. The research conducted by Ayman and Chemers (1983; 1986), for example, demonstrated that sensitivity to group norms was a more critical component of leader behavior in Iran and Mexico than in the United States. In a similar vein, Earley and Erez (1987) found that the same goal setting intervention implemented in Israeli and American work teams produced markedly different group interactions. Earley (1994) also uncovered a differential tendency in Chinese, Israeli, and American work teams to decrease performance due to group membership, a phenomenon often referred to as social loafing. 

More recently, Gibson (1996) found that the relationship between team beliefs and team performance differed in American and Indonesian work teams. In collectivistic work teams (typically found in Indonesia, discussed more below), strong positive beliefs enhanced team performance. In individualistic work teams (typically found in the U.S.), strong positive beliefs inhibited team performance. In a similar vein, Kirkman (1997) found that amount of resistance to working in teams varied depending upon the cultural orientation of employees in the U.S., Finland, Belgium, and the Philippines. Respondents with individualistic values (typically found in the U.S. and Finland) resisted working in teams more than respondents with collectivistic values (typically found in the Philippines and Belgium). Furthermore, respondents who valued power distance (typically found in the Philippines and Finland), reported higher levels of resistance to self-management than did those low in power distance (typically found in the U.S. and Belgium). 

TEAMWORK KNOWLEDGE AS AN EXPLANATORY VARIABLE 

These studies indicate that culture has implications for team effectiveness. Why might this be the case? To answer this question, we believe it is important to first look inside teams to better understand the social cognitive processes that occur as teams work together in an organizational context. Advancements in the study of social cognition suggest that teams share cognitive processes, and that cognition can be meaningfully understood at the group level of analysis (Brooks, 1994; Carley, 1997; Cannon-Bowers, Salas & Converse, 1993; Converse, Cannon-Bowers, & Salas, 1991; Kamoche, 1995; Klimoski & Mohammed, 1994; Larson & Christensen, 1993; Orasanu & Salas, 1993; Rentsch, Heffner, & Duffy, 1994; Rouse, Cannon-Bowers, & Salas, 1993). 

Traditionally, the term social cognition has referred to the content of individual cognitions regarding social behavior in interaction with other people. More recently, however, interest has grown in a different type of social cognition, in which the word "social" denotes how cognition is accomplished, not its content (Larson & Christensen, 1993). The focus becomes the processes involved in the acquisition, storage, transmission, manipulation and use of information for the purposes of creating a group intellective product. This definition of social cognition is also consistent with an emerging trend in the small group and interpersonal relations literature to view intragroup communication as a form of information processing (e.g., von Cranach, Ochsebein & Valach, 1986; Wegner, 1987). Cognitive phenomenon associated with this process have also been referred to as collective cognition (Gibson, 1996), transactive memory (Wegner, 1987), team mental models (Rouse, Cannon-Bowers, & Salas, 1992; Klimoski & Mohammed, 1994), and teamwork knowledge schemas (Rentsch, Heffner, & Duffy, 1994) . 

Teamwork knowledge as an outcome of social cognition. The concept of teamwork knowledge is relatively new in the managerial literature. Teamwork knowledge is made up of what each team member understands about team work. Teamwork knowledge is an example of a team schema (Rentsch, Hefner, & Duffy, 1994) or team mental model (Klimoski & Mohammed, 1994). A schema is a knowledge structure developed from past experience and used to organize new information and to facilitate understanding (Poole, Gray, & Gioia, 1990). Although team members may have teamwork knowledge within a specific domain, (e.g., teamwork on a football team, or on a sales team), it is likely that they will also acquire and develop a core teamwork knowledge structure that applies to most types of teams. This core teamwork knowledge might include knowledge for enhancing the quality of team members’ interactions, communications, relationships, and so on (Salas, Montero, Glickman, & Morgan, 1988). Core teamwork knowledge provides individuals with an understanding and a structure for working together as a team. 

It is during a group’s discussion that social cognition occurs and teamwork knowledge schema develop. Indeed, the group’s discussion is social cognition. The interactions that take place during group discussion serve three distinct functions (Sniezek & Henry, 1990). First, interactions help bring problem-relevant information to light. Second, they serve as a means of influencing the individual-level cognitive processes that take place within each group member. That is, by highlighting certain items of information, drawing attention to faulty logic, presenting arguments in support of particular conclusions, etc., group members can affect one another’s perceptions, judgments, and opinions (Stasser & Davis, 1981). Finally, social interaction also serves as the vehicle by which group members’ perceptions, judgments and opinions are combined in order to generate a single group solution. 

Content of Teamwork Knowledge Schema. There appear to be two basic notions of the content of teamwork schema: perceptual (situational) and behavioral (action). In accordance with the perceptual perspective, teamwork schema capture assessment of stimuli in situations in which team members are trying to conceptualize what is going on around them. In the language of cognitive psychology, team members are developing and using categories in their interpretation. Even within the perceptual perspective, there is some disagreement about what aspects of knowledge are shared. While some researchers suggest beliefs (Axelrod, 1976; Bonham et al. 1988; Innami 1992; Walsh & Fahey, 1986) or understanding (Daft & Weick, 1984; Floyd & Woolridge, 1992) as the entities which decision makers hold in common, others refer to shared frames of reference (Isabella, 1990; Panzano, 1992) or shared categories (Fiol, 1993; Panzano, 1992). 

For example, Cannon-Bowers et al. (1993) are very explicit about stating that it is expectations that are held in common in a team, while Orasanu and Salas (1993) posit that organized knowledge is what is shared among team members. Also illustrating this perspective, Weick and Bougon (1988) and Gray, Bougon, and Donnellon (1985) suggest that team schema center around concepts and/or relationships among concepts. Gray et al. (1985) also recognize that team schema may involve values and ideologies. Fiol (1983) breaks the notion of team schema into two interpretive dimensions: content and framing. First, consensus around interpretations is embedded in the content of communications, which is reflected in the categories or labels that define what is expected (e.g., threats and opportunities). Second, meaning also resides in the framing of communications, which refers to the way in which view points are expressed, regardless of content (e.g., rigid or flexible perceptions of an issue). Therefore, there can be agreement not only with regard to what is said but how it is said. 

A second, very different perspective suggests that schema capture what team members are to do about what they see or comprehend. Team schema are often treated as internalized decision-choice/behavioral routines or "scripts." For example, Cannon-Bowers et al. (1993) describe the content of team schema as task work or team work. Models of task work refer to an understanding of the activities and action sequences that both the individual and the team collectively must carry out to perform the team task. Models of team work would include a conceptualization of the need to communicate with one another, compensate for one another's weaknesses, the proper amount of mutual performance monitoring to carry out, and the type of internal coordination strategies needed for the team to function effectively. Using this same perspective, Weick & Roberts (1993) propose that the interrelating of social activities embody collective schema. In this sense, teamwork schema seem to include what has been traditionally been termed "role expectations" (Katz & Kahn, 1978) or group norms (Feldman, 1984); however, teamwork schema are distinct from the notions of culture or norms. Teamwork schema are emergent characteristics of the group which reflect organized knowledge and the tendency of individuals to categorize what they "know." 

Klimoski and Mohammed (1994) argue for a conceptualization of teamwork schema, which they refer to as team mental models, that combines both perceptions and actions. They suggest that teamwork schema reflect a variety of content. Schema may represent efforts to simplify events or responsibilities in order to make them more tractable. They reflect organized knowledge, usually in the form of a set of concepts stored and retrieved from memory in relationships to one another. Such organized knowledge may derive from presumed cause and effect linkages or they merely may reflect learned patterns. More to the point, while the organized patterns may be strictly spatial in nature and origin, in all likelihood, such knowledge is organized semantically. Moreover, teamwork schema reflect internalized beliefs, assumptions, and perceptions. They are really how the group members as a collectivity think or characterize phenomena. They are also an emergent characteristics of the group, which is more than just the sum of individual models. 

Several researchers have argued for the notion of multiple teamwork schema. For example, Klimoski & Mohammed (1994: 432) state, "there is probably no value in continuing to think of "a" team mental model. There can be (and probably would be) multiple models co-existing among team members at a given point in time.....the nature of these models will be in some ways context dependent." Orasanu (1990, cf. Orasanu & Salas, 1993) has suggested that teams must develop shared situational schema for the specific problem, which are grounded in the team's stable schema for the system, task, and team, but go beyond them. These situation schema include shared understanding of the problem, goals, information cues, strategies, and member roles. Communication is used to build shared situation schema. Once shared schema have been created, they provide a context for interpreting directions or information requests, and allow for volunteering of information or actions at appropriate times. They also provide a basis for predicting behavior or needs of other team members. 

Empirical support for the concept. Several studies provide support for the existence of collectively-held schema. For example, Panzano (1992) conducted a field study investigating how top decision makers in Ohio mental health boards interpret the Mental Health Act of 1988. He found mean rater reliability statistics which supported the notion that organizational frames exist as shared perceptions among members. Interestingly, convergence was higher for framing variables such as threat and opportunity than other perceptual measures (e.g., slack). According to Panzano (1992), these results suggest that managers on mental health boards share interpretations about how key issues impact their organization. 

Strong evidence concerning shared cognition was also provided by Hutchins (1991). Using a computer simulation model and a connectionist framework for thinking about cognitive phenomenon at the group level of analysis, Hutchins (1991) demonstrated that the cognitive properties of groups can differ from those of their participating members. Specifically, the simulations suggested that, even when holding the cognitive properties of individuals constant, groups as a whole may display different cognitive properties, depending on how communication is organized within the group over time. The cognitive properties of groups were produced as a result of an interaction between structures internal and external to individuals. 

Using multidimensional scaling and concept maps, Rentsch et al. (1994) found that individuals with more teamwork experience had more concise and general teamwork knowledge. For example, high experience individuals described teamwork primarily using three categories of adjectives: individual team member characteristics, leader qualities and negative team characteristics. Low experience individuals used many more categories of adjectives and within each category, generated more adjectives. 

Implications of teamwork knowledge. At this early stage in the development of the concept of teamwork schema, there is limited empirical support for these types of impacts. However, one study sheds light on the implications of teamwork schema. Walsh, Henderson & Deighton (1988) used realized coverage (the breadth of perspectives voiced during a discussion) and realized consensus (shared representations) to operationalize their notion of a negotiated belief structure. The analysis of 713 product decisions made by 29 groups of graduate students in a complex, simulated business environment indicated that coverage and consensus were systematically related to product and firm performance. Specifically, shared agreement around a few schematic dimensions was associated with superior brand performance. Walsh and his colleagues concluded that the linkages between the coverage and consensus variables and decision performance point to the validity of the negotiated belief structure construct. 

An important function that can be served by social cognition is the discovery and resolution of differences in how various group members conceptualize a problem at hand (Doise & Douglas, 1978; Hutchins, 1991; Ono, Zimmerman, Stasson & Davis, 1985). A variety of actions might be taken by group members to accomplish this. For example, they might ask one another to share their individual conceptions of the problem expressly for the purpose of surfacing discrepancies. Or one member might take the initiative and try and "sell" her/his own conceptualization to others. This person might be a formal leader of the group, an informal leader, or perhaps a subject matter expert. 

Leaders, both formal and informal, may take such action in part to fulfill role expectations for providing structure (Hollander, 1985). Experts, on the other hand, might be more motivated by their access to problem solutions. That is, experts can be thought of as carrying around with them certain solutions and the means for attaining those solutions. Access to these means and solutions gives experts leverage and confidence in solving certain kinds of problems. However, it is also likely to bias them toward conceptualizing ambiguous problem situations in ways that favor the use of those means and solutions. If a problem situation can be conceptualized in a way that is commensurate with an available solution, then the expert has a rational basis for persuading others that her/his preferred conceptualization is, in fact correct. Thus, groups are likely to adopt problem definitions that are consistent with available solutions (cf. Cohen, March & Olson, 1972; March & Olson, 1976). 

Potential variations in teamwork knowledge. Shared teamwork knowledge schema refer to organized knowledge shared by team members (Orasanu & Salas, 1993). The observations obtained in our interviews indicated that some of this knowledge is broadly shared by members of a culture; while some knowledge is limited to members of a restricted group, such as members of an organization or a profession; and some knowledge is particular to a situation. Previous cross-cultural research on teams (e.g., Earley, 1994; Gibson, 1996; Kirkman, 1997) suggests that important variation in aspects of teamwork can be at least partially explained by cultural values, cultural differentiation, and organizational culture. It is reasonable to believe that these same phenomena impact teamwork knowledge, resulting in variation in the concept of teamwork across cultures and across organizations in our sample. 

Furthermore, intercultural research has established that organizations are more effective when management practices in a work unit are congruent with national culture (Hofstede, 1980; Shackleton & Ali, 1990; Chow, Shields & Chan, 1991; Newman & Nollen, 1996). For example, Newman & Nollen (1996) found that work units that are managed in a manner consistent with the values of the external culture are more profitable than work units in which the fit is less well achieved. Given that multinational organizations often operate across several cultures, these organizations face great challenges in managing teams. However, at this point, research is unable to give more comprehensive guidance regarding the management of teams to the multinational organization. 

Summary. We began a four year program of research investigating social cognitive processes in teams across cultures by conducting exploratory interviews with over one hundred members of teams in six multinational organizations. Our observations led us to question the basic assumption that employees agree on the meaning of the word teamwork. Literature regarding the concept of teamwork knowledge schema, intercultural values, and organizational culture was reviewed to develop potential mechanisms that explain this variation. Working in an inductive mode, we analyzed interview transcripts to better understand how the concept of teamwork knowledge varied in multinational organizations. We report the results of these analyses below and then outline specific propositions to be investigated in future phases of the research. 

METHODS 

Sample. The cultural contexts selected for this research were the U.S., France, Puerto Rico, and the Philippines. According to research conducted by Hofstede (1980), employees in the U.S. and France tend to be relatively low in power distance and collectivism, whereas employees in Puerto Rico and the Philippines tend to be relatively high in power distance and collectivism. Furthermore, research summarized by Wapner and Demick (1991) suggests that the U.S. and France are relatively high in cognitive differentiation; whereas Puerto Rico and the Philippines are relatively low in cognitive differentiation. 

Multinational firms listed in the Corporate Families and International Affiliates Directory encompassed the general pool from which organizations were selected. Three key factors were taken into consideration in selecting organizations: (1) industry, (2) geographic representation, and (3) use of permanent teams in each of two functional areas (Manufacturing/Production and Sales/Marketing). Procedures associated with each of these three factors are discussed below. 

Firms were first screened using the Standard Industrial Classification (SIC) coding system to control for potential industry effects. The SIC system has been widely adopted as a standard for defining and analyzing industries and is periodically revised by the U.S. government. The SIC system assigns lines of business in a firm to one of ten major categories and then assigns a four digit code to each line. The first two digits describe the general nature of the work activity being conducted, the third and fourth digits describe the specific activity. 

Two major SIC divisions were included in order to incorporate different stages of the production cycle - the Manufacturing Division and the Wholesale Trade Division. Within Manufacturing, the four-digit codes 2833 (Medicinal Chemicals and Botanical Products), 2834 (Pharmaceutical Preparations), and 2899 (Chemicals and Chemical Products) represent pharmaceutical-related activities. Within Wholesale Trade, the four digit codes 5047 (Medical, Dental, and Hospital Equipment and Supplies), 5122 (Drugs, Drug Proprietaries, and Druggists Sundries), and 5169 (Chemicals and Allied Products) represent pharmaceutical-related activities. To be considered for this study, a firm was required to have a least one line of business coded as pharmaceutical manufacturing (2833, 2834, or 2899) and at least one line of business coded as pharmaceutical wholesale trade (5047, 5122, or 5169). 

Locations of facilities were examined next. Only firms with facilities in each of four regions (the U.S., Latin America, Southeast Asia, and Western Europe) were considered. Furthermore, the facilities in each of these regions had to represent lines of business coded as described above using the SIC coding system. Using this procedure yielded a total of ten multinational firms: Abbott Laboratories, Baxter International, Eastman Kodak Company, GE Medical Systems, Johnson & Johnson, Merck and Company, Pfizer, Sanofi Winthrop, SmithKline Beecham, and Warner-Lambert. 

Human resource professionals in each of the ten firms were contacted by phone in order to provide a brief introduction to the research and to screen for their use of teams. Based upon the work of Mohrman, Cohen, and Mohrman (1995), our definition of work teams was, "a group of individuals who work together interdependently in a continuous manner to produce products or deliver services for which they are mutually accountable." Only firms that utilize work teams in both manufacturing/production and sales/marketing across each of the four geographic regions were invited to participate. Six of the ten multinational firms listed above qualified. The identity of these firms has been disguised here using code numbers. 

Design. The research program investigates teams in multinational organizations using the method of triangulation described by Jick (1979). This method involves studying a phenomenon of interest utilizing three distinct methods and analyzing the degree of convergence in the findings across methods. The investigation will proceed in six phases over a 4-year period (see Table 1). 
__________________________________________ 
Insert Table 1 About Here 
_________________________________________ 
Phase I of the project consisted of Exploratory Interviews. A written project proposal was submitted to each of the six firms that qualified for inclusion in the sample. Written correspondence was followed by phone contact in order to gather further information about the teams being utilized and answer potential questions about the research. Meetings were conducted with human resource professionals in order to schedule interviews. Human resource professionals in each multinational firm were asked to randomly select three to four teams per facility across the four geographic regions for the interviews. 

A total of 111 individuals were interviewed. These individuals represent 59 teams. Between one and eight individuals were interviewed from each team. In the United States 45 individuals representing 11 teams were interviewed; in France 16 individuals representing 12 teams; in Puerto Rico 25 individuals representing 11 teams; and in the Philippines 26 individuals representing 12 teams were interviewed. Sample statistics are presented in Table 2. 
__________________________________________ 
Insert Table 2 About Here 
__________________________________________ 
Four types of teams were included. The majority of the teams in the sample (59%) were work teams. Work teams are ongoing teams responsible for producing goods or services (Cohen & Bailey, 1997). Their membership is typically stable. The second most prevalent team type (18%) was project teams. Project teams are time-limited and used for a one-time output such as a new product or service. They are typically cross-functional. Two other team types represented in the sample are parallel teams (6%) and management teams (6%). Parallel teams pull people from many different work units to perform a job that the regular organization is not well equipped to do -- they exist "in parallel" to the formal structure. Management teams laterally integrate sub-units, provide direction and are responsible for the overall performance of a business unit. 

The researchers traveled to each region of the world and conducted in-depth personal interviews with a wide variety of members from each organization. These individuals ranged from first line supervisors of manufacturing and distribution teams to vice presidents and business unit managers. The researchers posed a series of predetermined questions pertaining to concepts of teamwork, the function of the teams, team motivation, leadership in the teams, feedback and reward systems, sharing of knowledge and practices across teams, metaphors for teamwork, and the impact of culture on teams. Interviewees were also asked to discuss which factors they felt were the most important facilitators and inhibitors of team effectiveness. A complete list of the interview questions is contained in the Appendix. The researchers encouraged the discussion of any additional issues interviewees felt were pertinent to the use of teams in multinational companies. 

ANALYSIS 

Database Preparation & Analysis Plan. All interviews were tape recorded and transcribed by a professional transcriptionist to prepare them for qualitative analyses. The transcriptions resulted in a text database consisting of over 1,000 pages of single-spaced text. We then conducted a qualitative analysis of the text database. Qualitative analysis involves the process of making sense of data that is not expressed in numbers and is especially useful in the exploratory stages of theory development (Tesch, 1990). The analysis was conducted in a manner consistent with that recommended by both Strauss and Corbin (1990) and Gephart and his colleagues (Gephart, 1993; Wolf, Gephart, and Johnson, 1993). 

Two general stages of analysis were performed. In the first stage, we sought to roughly identify themes or metaphors used by members of the sample to describe teamwork. As noted earlier, it was while conducting interviews for the larger research program that we noticed that there appear to be meaningful differences in the way that the word teamwork is conceptualized, particularly across cultures and across organizations. Before proceeding with a detailed, in-depth content analysis, we wanted to confirm our intuitive sense that we were hearing different conceptualizations of teamwork during the interviews. We therefore conducted a preliminary analysis to delineate common themes for the meaning of teamwork in the transcripts. This preliminary analysis was followed by a second stage involving inductive creation of word lists and categories. Next, the entire interview database was subject to comprehensive textual analysis to determine whether variance in teamwork knowledge existed across countries and across organizations. The content analysis program TACT (Bradley, 1989; Hawthorne, 1994; Popping, 1997) was utilized in order to facilitate the comprehensive text analysis. We describe the specific steps in our two stages of analysis below. 

Step One: Word Processor Analysis. A word processing program was used to search for text excerpts that describe teamwork in the words of interviewees. This analysis consisted of using a word processor to search for words such as "metaphor," "images," and "mental picture" (please see Appendix for a complete list of interview questions). We highlighted these words with colors to allow us to readily locate these passages in the interview transcripts. Interviewees’ answers to these questions were selected, printed and carefully read. 

Five different themes emerged from this stage of analysis. Segments reported here were selected from the interview transcripts because they were coherent passages demonstrating team members’ conceptualization of teamwork. The first theme identified was "team as family" which the following excerpts from Puerto Rico illustrate: 

Textual Analysis. Driven by the identification of different metaphors, the next step involved a more thorough examination of the language used to describe teamwork using the technique of text analysis (Gephart, 1993). Since there is no theoretical or empirical precedent, we continued to work inductively and used text analysis to facilitate data exploration (Jehn & Doucet, 1996). Computer-aided text analysis programs are very helpful in this endeavor as they allow examination of very large amounts of text in a relatively efficient manner. Computer-aided searching allows the categorization and comparison across texts from various interviews to examine content concerning teamwork. 

To prepare for the text analysis, the interview transcripts were input into the computer text analysis program TACT. The input process requires that a structure and coding scheme be defined for the text. We determined a coding scheme which allowed us to distinguish speakers, teams, organizations and nations. As a result of this input process, TACT creates a special data file called a "personal database" that contains information about the position and structure of every word in the database. The total number of words in our database was 266,905 words. 

Following Gephart and others (Gephart, 1988, 1993; Gephart & Wolfe, 1989; Jehn & Doucet, 1996) the first step in comprehensive text analysis is to develop a list of words and/or categories of words that are thought to capture the construct of teamwork. If there is adequate theoretical background, such a list may be derived from previous research, extant scales, dictionary lists and thesaurus lists. However, for this research, we wanted to avoid this deductive list creation for two reasons. First, the construct of teamwork schema or teamwork mental model is very new and very little extant research exists. Second, and more importantly, this sample is multi-cultural, and the very center of our research question is whether such schema or models vary across cultures. Therefore, developing word lists from western-based theories, dictionaries and thesauruses would be inappropriate. These sources may miss important culturally-embedded terms. Furthermore, the sources may apply alternative or inappropriate meanings to words other than those intended by non-western speakers. Therefore we chose to follow a two-step process designed to develop word lists which captured cultural nuances in meaning. The first step was a term-selection process using multiple raters. The second step was a categorization of these terms. These two steps are described below. 

The first analysis run was the creation of an alphabetical list of every word in the interview database. The number of times each word is mentioned in the database was computed. This resulted in an alphabetized list of every unique word used in the interviews. The alphabetical word list for our interviews contained 6,661 unique words. Copies of this list were given to two raters from each country. Raters were asked to independently circle all terms they felt were related to teamwork. Raters were told to broadly define "teamwork" and were not given a definition of teamwork. When the raters had completed their task, the researchers created a spreadsheet with a column devoted to each rater. A complete list of all unique terms identified across raters was developed from these individual lists. This list contained 1,740 words. 

The next step was to collapse this list of words into categories. The words were printed out onto cards and five raters representing the countries involved in the study sorted them into categories. The raters took the slips of paper and worked together to put them into categories of like terms. They were encouraged to talk to each other during the process and ask questions about why a given word was placed into a category. They were also encouraged to express unique ways in which a term is used in their native culture. For this reason, some words were placed in more than one category if it was determined that the word held multiple meanings. This process resulted in 27 different categories. The raters developed names for each category. Categories ranged from those that captured team outputs to those that captured forms of team leadership, to those that captured concepts, symbols or schema for teamwork. For the purposes of this paper, we chose to focus on this last set of categories which captured schema for teams. Five categories developed by our raters represented schema for teamwork. These categories were: (1) family; (2) sports; (3) community; (4) associates; and (5) military. The remaining 22 categories generated by the raters will be used in future analyses outside the scope of this paper. 

These five teamwork schema categories contained on average approximately 30 unique words each, as designated by our raters from each country. The raters defined the categories broadly and placed words in them that they felt elaborated on the category from their native language point of view. Before running our main analyses, however, we needed to determine if these same words in the context of our interviews were being used to describe teamwork. This process is often referred to as "in-context verification" (Gibson, 1994; Gephart, 1993). 

TACT allows the creation of categories with multiple words. Each of the five categories was defined with the words identified by the raters. TACT was instructed to pull excerpts from the text database which contained each term in each category (a total of approximately 150 words). We extracted five lines of text before and five lines of text after the term was used. This created a "subtext database" consisting of approximately 400 pages of single-spaced excerpts. The researchers read each excerpt to verify the manner in which each term had been utilized. In some cases it was very clear that the term was not being used in conjunction with teamwork. For example, the word "depends" had been selected for the family category by the raters. In the database, "depends" was used only in phrases like "it depends on..." to denote contingency. These comments did not relate to teamwork. For this reason, the word "depends" was removed from the category, as were other words that were clearly not being used to describe teamwork in our interview database. The final contents of each category after screening for in-context verification are displayed in Table 3. 

After verifying the categories, the main analyses were performed. TACT was instructed to create five "category databases" containing excerpts of text which included any word in a category. A frequency distribution was produced for each word in a category across all interviewees in the sample. These frequencies are reported in brackets next to the words in Table 3. Next, a frequency distribution for each category by country was produced. Finally, a frequency distribution for each category by company was produced. Results are reported below. 
 
__________________________________________ 
Insert Table 3: Categories 
__________________________________________ 

RESULTS 

Text Excerpts. From the list of text excerpts, several representative passages for each category were chosen to illustrate how these categories were used with respect to teamwork. These excerpts are listed in Table 4. 
_________________________________________ 
Insert Table 4: Excerpts 
__________________________________________ 
Weighting for Sample Size. The total number of interviewees varied across countries and these differences may cloud our results. For example, in the U.S. the total number of interviews is considerably larger (44 interviews) than in France (15 interviews). Purely due to chance, we might expect to see more occurrences of any given category in the U.S. than in France. Therefore, it was necessary to weight the frequencies to account for differences in absolute numbers of interviews across countries. 

Frequencies Across Countries. For each schema category, a distribution of occurrences across countries was created using TACT. This frequency represents how many times a word in that category occurred in the transcripts for that country. These frequencies are summarized in Table 5. Chi-square analyses were run to test the null hypothesis that no difference in frequency should be observed in the use of a specific category across countries. In this test, a significant chi-square statistic indicates that the categories do not occur with equal frequency across countries. The results are reported in the last column of Table 5. The results indicate statistically significant differences for each of the five schema categories (for Military c ² = 6.911, p > .10; for Family c ² = 11.094, p > .05; for Sport c ² = 27.422, p > .01; for Associates c ² = 42.307, p > .01; and for Community c ² = 18.399, p > .01). 

These analyses suggest that each schema category occurred more frequently in some countries than in others. For example, military schema occurred most frequently in the Philippines and least frequently in France. Family schema occurred most frequently in the Philippines and least frequently in Puerto Rico. Sport schema occurred most frequently in France and least frequently in Puerto Rico. The associates schema occurred most frequently in France and least frequently in Puerto Rico. Community schema occurred most frequently in the Philippines and least frequently in Puerto Rico. 

Next, we investigated within-country differences and ran chi-square tests to examine whether the schema categories occurred with equal frequency within a given country. Again, the null hypothesis was that the categories would occur equally within a country. In these tests, a significant chi-square statistic indicates that the categories do not occur with equal frequency in that country. Results are reported in the last row of Table 5. In three out of the four countries (Puerto Rico, the Philippines, and France), the five schema categories occurred relatively equally. The chi-square statistic was only statistically significant in the U.S. (c ² = 17.263, p > .01), indicating that U.S. respondents used some categories more frequently than others. Examination of Table 5 demonstrates that U.S. respondents used the associates schema most often and military schema least often. 
__________________________________________ 
Insert Table 5: Frequencies by Country 
__________________________________________ 
Frequencies Across Companies. For each category, TACT was instructed to produce a frequency across companies. For the comparisons across companies, frequencies were weighted by number of interviewees per company. These frequencies are reported in Table 6. Chi-square analyses were run to test the null hypothesis that the schema categories would occur equally across companies. In these tests, a significant chi-square statistic indicates that the categories do not occur with equal frequency across companies. The results of these analyses are reported in the last column of Table 6. The chi-square statistic was statistically significant for four of the five categories (for Family c ² = 14.483, p > .05; for Sport c ² = 10.527, p > .10; for Associates c ² = 12.043, p > .05; and for Community c ² = 25.412, p > .01). 

These analyses indicated that four of the schema categories occurred more frequently in some companies than in others. For example, military schema occurred most frequently in Firm 15 and least frequently in Firm 12. Family schema occurred most frequently in Firm 11 and least frequently in Firm 10. Sport schema occurred most frequently in Firm 15 and least frequently in Firm 11. The associates schema occurred most frequently in Firm 15 and least frequently in Firm 12. The community schema occurred most frequently in Firm 12 and least frequently in Firm 11. 

Finally, we investigated within company differences and ran chi-square tests to examine whether the categories occurred equally within a given company. Again, the null hypothesis in these tests is that the categories will occur equally. A significant chi-square statistic indicates that the categories do not occur with equal frequency in that company. The results of these analyses are reported in the last row of Table 6. In five out of the six organizations, the categories were used relatively equally. The chi-square statistic was significant only in Company 12 (c ² = 10.543, p > .05), indicating that respondents in Company 12 used some schema categories more frequently than others. Examinations of the frequencies in Table 6 demonstrates that respondents in Company 12 used community schema most frequently and military schema least frequently. 
__________________________________________ 
Insert Table 6: Frequencies by Company 
__________________________________________ 
Summary of findings. These results suggest that interviewees use a variety of themes to describe teamwork. These themes, or teamwork knowledge schema, appear to vary across cultures and across organizations. Certain themes are more prevalent in certain cultures. Certain themes are more prevalent in specific organizations. However, additional research is necessary with this text database, and others, in order to better understand how and why these themes differ. Our inductive analysis also suggests additional research is necessary to determine the impact of these themes. 

DISCUSSION AND PROPOSITIONS FOR FUTURE RESEARCH 

In this section, we pose several propositions for future research investigating variance in teamwork knowledge schema. These propositions are based on the inductive analysis, frequency distributions, and careful examination of the text excerpts for each teamwork schema category reported here. We make predictions concerning the specific content of teamwork schema, concerning the degree of variation we might expect to see in the use of these schema, and concerning the impact of such schema. The relationships we propose are depicted in Figure 1. 

Cultural values. With regard to differences across national cultures, a common theme in previous studies is that many of these differences are due to variations in cultural values. Although variations within countries do exist, people within a given country often share common values and these values can be utilized to distinguish one country’s culture from another (Hofstede, 1980; Shweder & LeVine, 1984; Triandis, 1989). More specifically, two cultural values that appear to be especially critical in work organizations are power distance and collectivism. Power distance represents the degree to which members of a culture accept and expect that power in society is distributed unequally. Cultures low in power distance will try to minimize inequalities, favor less autocratic leadership, and favor less centralization of authority. Research suggests that the Scandinavian countries, Austria, Israel, Ireland and the United States are low in power distance (Hofstede, 1980). On the other hand, cultures high in power distance will be characterized by greater acceptance of inequalities and preference for authoritarianism. Research suggests that Malaysia, the Philippines, Panama, Guatemala, and Puerto Rico are high in power distance; Japan, France, Belgium, Italy are moderate in power distance (Hofstede, 1980). 

Power distance may have implications for teamwork knowledge, in that it might shape team members expectations regarding hierarchical relations in teams. We might expect to see schema for teamwork which contain information about hierarchical relationships in high power distance cultures. Across our five teamwork schema categories, the military and family categories contain the most information about hierarchical relationships. Our chi-square analysis suggested that there were significant differences in the use of both the military and family categories across countries. Respondents in the Philippines used these schema the most often. Previous research suggests that Filipinos are typically high in power distance. In contrast, across our five teamwork schema categories, the sport and the associates categories contain the least information about hierarchical relationships. Our chi-square analysis suggested that there were significant differences in the use of both these categories across countries. Respondents in France used these schema most often. Previous research suggests that the French are typically moderate in power distance. These findings lead us to pose the following propositions: 

In line with the cultural congruence argument (e.g., Newman & Nollen, 1996 discussed earlier), we would expect that teams which have military or family schema will be more effective when managed in such a way that recognizes status relationships and incorporates these into the team structure. Thus, our full argument would be that teams in cultural contexts which value power distance will use teamwork knowledge schema which contain implicit information about hierarchy. The meaning they ascribe to teamwork will include prescriptions for status relationships. When managerial practices coincide with these schema, teams will be more effective. The following proposition summarizes this notion: 

Collectivism may have implications for teamwork knowledge schema in that it may impact the content of the size of the entity considered a team. Workers in collectivistic cultures are likely take a broader, more encompassing view of teamwork. In these cultures, we might expect that teamwork schema will refer to large social aggregates. Across our five teamwork schema categories, the community category refers to the largest social aggregate (e.g., a community typically consists of more people than a sport team). There were significant differences in the occurrence of this category across cultures. It occurred most frequently in the Philippines. Previous research suggests that Filipinos are highly collectivistic. In contrast, across our five teamwork schema categories, the sport category arguable refers to the smallest social aggregate (e.g., a sport team typically consists of fewer people than a community). There were significant differences in the occurrence of this category across cultures. It occurred most frequently in France. Previous research suggests that the French are low on collectivism. These findings lead us to propose the following propositions for future research: 

In contrast, high differentiation is often evidenced on perceptual tasks, when objectives are perceived as discreet from their perceptual contexts (Carli et al. 1986). An associated tendency is the inclination to develop knowledge, language and customs that are context-free (Triandis, 1989). In general, the Swiss, Germans, French, Scandinavians and Americans tend to be high in differentiation (Witkin & Berry, 1975; Gundykunst, Ting-Toomey & Chua, 1988). For example, in the U.S. or the Netherlands, people tend to believe that each stranger should share in rule making, and the fewer initial structures there are the better. 

Cultural differentiation may have important implications for the degree of variety in teamwork knowledge schema. We might expect that in high differentiation cultures, we will see only one or two dominant schema for teamwork. In contrast, in the low differentiation cultures, we will see a variety of teamwork knowledge schema that are used interchangeably depending on the context. In our interview sample, the Americans are typically the highest in cultural differentiation. Our chi-square analysis demonstrated that the U.S. was the only country in which significant differences occurred in the frequency of category usage. In the U.S., associates category dominated; the family category was used least often. In the other three countries, respondents drew upon several schema categories, and did not favor any one category more so than another. These findings led us to propose the following propositions for future research: 

Empirical evidence of differences in organizations’ cultures has been demonstrated (e.g. O’Reilly, Chatman & Caldwell, 1991) and even organizations that appear to be highly similar, such as large public accounting firms as a group, may vary widely in their underlying value structure (Chatman & Jehn, 1994). More recent conceptual advances have suggested that dynamic industries may have greater inter-organizational culture diversity than stable industries (Gordon, 1991). In addition to industry differences, organizational culture is also thought to vary based on founders and leaders of the organization (Schein, 1993). 

The frequency comparisons made by company suggest that the occurrence of teamwork schema also vary across organizations. Chi-square results indicated that there were statistically significant differences in the use of the family, sport, associates and community categories across organizations. The sport and associates categories were used most frequently by interviewees from Firm 15. The family category was used most frequently by interviewees from Firm 11. The community category was used most frequently by interviewees from Firm 12. Variations in organizational culture may lead to variations in the shared knowledge of teamwork held by organizational members. These variations in culture may be identified by variance on underlying dimensions of organizational values. O’Reilly, Chatman and Caldwell (1991) identified eight dimensions of organizational culture based on values held by organizational members. To illustrate how these organizational culture dimensions may relate to variation in teamwork knowledge schema, consider the dimension "aggressiveness." Organizations with highly aggressive cultures tend to value competition, opportunism, and aggression, and tend to devalue social responsibility. These values may have implications for teamwork knowledge content: 

The research and propositions presented in this paper represent advances in research methodology, theory and practice. Previous research has investigated team implementation in U.S. organizations, but the vast majority of these investigations have remained within our national boundaries. Very little insight into the functioning of teams in multinational organizations has been provided by previous research. As demonstrated by the results described here, we have made every attempt to be sensitive to intercultural variation in conducting this research. We feel this improves the robustness of our results. Evidence that this is true is demonstrated by comparing the results we obtain in our stage one preliminary analysis of the interviews with the results we obtained in our stage two comprehensive content analysis. Our preliminary analysis led us to believe that five themes existed in the interviews: team as family, team as orchestra, team as social circle, team as village, and team as sport. However, when we conducted the second stage of the analysis using raters representing each of the cultures involved in the study, we found a different set of categories emerged. No support was obtained for the team as orchestra theme and an entirely new theme, team as military, developed. Furthermore, the "team as village" theme was expanded to become the "team as community" category and the "team as social circle" theme was modified to become the "team as associates" category. Thus our two-stage process eventuated in the identification of five schemas for teamwork knowledge: team as family, team as sport, team as military, team as associates, and team as community. Had we not conducted the second stage of the analysis, our preliminary results would have represented a Western bias that we hold as American researchers. Instead, through a comprehensive culturally-sensitive process, we were able to develop schema categories that are relevant across cultures and organizations. 

Beyond advancements in research methodology, our research will also advance theory in at least two regards. First and foremost, theoretical developments based on this research will make a contribution to the body of literature regarding work teams. Several reviews of this literature (e.g, Sundstrom et al. 1990; Goodman, Ravlin & Schminke, 1990; Bettenhausen, 1991) suggest that what has been lacking is a focus on the cultural context in which teams are embedded. The research program described herein directly addresses the influence of context, proposing and empirically investigating cultural phenomena that act as mechanisms through which the cultural context influences teams. Second, this research adds to the body of theory that has been referred to as international and intercultural management. Theories in this area strive to understand the managerial implications of values, beliefs, and ways of acting shared within a society. Foundations are borrowed from the disciplines of social psychology, sociology, anthropology and organizational behavior. 

More specifically, this research will advance theory surrounding the social cognitive processes in teams. The results of the first-year interviews, for example, contribute to the development of the concept of teamwork knowledge schema. By way of review, teamwork knowledge is made up of what each team member understand about teamwork. Teamwork knowledge schema are knowledge structures developed from past experience and used to organize new information and to facilitate understanding. One debate we mentioned earlier in the literature on teamwork schema is whether these schema contain perceptual information or behavioral information. Many theorists have suggested that the schema contain one or the other; Klimoski and Mohammed (1994) argue that teamwork knowledge schema contain both. They also suggest that teams have multiple teamwork schemas, and that these might be shared to a greater or lesser degree. 

Our exploratory results begin to shed some light on these issues. For example, careful examination of the interview excerpts surrounding the five teamwork knowledge schema identified in our interviews suggested that teamwork knowledge schema contain information about both perceptions (e.g., beliefs, judgments, opinions) and about action (e.g., activities, action sequences, interactions). Furthermore, our quantitative results suggest that within three out of the four countries, multiple teamwork knowledge schema existed. Only in the U.S. were significant differences found in the use of the teamwork schemas. Within five out of the six organizations, multiple teamwork knowledge schema existed. Only in Company 12 were there significant differences found in the use of the teamwork schemas. Thus, our results provide preliminary support for Klimoski and Mohammed's (1994) arguments. 

In terms of practice, this research is also expected to advance the practice of international management. Due to the lack of empirical research regarding teams across cultures, leaders within international organizations have, up to this point, been forced to make educated guesses as to the most appropriate method of implementing teams across their various geographic facilities. Among other key insights, the research described here will provide guidance regarding the motivational processes that must be managed in order to successfully implement teams in interculturally. 

More specifically, this research will shed light on the impact of meaning systems on the quality of the work experience and the effectiveness of teams. One important aspect of knowledge schema that can be incorporated into training programs is the degree to which team members share teamwork knowledge. Actions taken by members to ensure that everyone in the group shares a common conceptualization of the problem at hand serve not only to coordinate subsequent problem-solving activities, they may also increase the likelihood that the conceptualization finally adopted is actually a suitable one (Larson & Christensen, 1993). Evidence suggests that group problem solving is significantly improved when members are encouraged to examine the way in which they have defined the situation, and to consider whether or not they are solving the "right" problem (e.g., Bottger & Yetton, 1987; Maier, 1970, cf. Larson & Christensen, 1993). 

Larson & Christensen (1993) use the term "meta-knowledge" to refer to information that provides clues about the types of problem-relevant information that particular group members are likely to have. They call the total pool of such information that exists within a group the group’s meta-knowledge base. Stasser and his colleagues (cf. Larson & Christensen, 1993) found that groups bring out a much larger percentage of their shared information that their unshared information during problem-solving discussions. Given the considerable difficulty that groups have in bringing out unshared information during discussion, it is of interest to speculate about what can be done to improve the situation. One possibility is to increase the group’s meta-knowledge base. Stasser (1991) found that when members knew which other members of the group had expertise in which specific knowledge domains, the amount of unshared information they actually discussed increased significantly. Such knowledge serves an essential integrative function in that it enables members to make more efficient use of each other’s personal store of problem-relevant information (Wegner, 1987). To an extent, widely distributed meta-knowledge may serve as a substitute for leadership in problem-solving groups (cf. Kerr & Jermier, 1978). 

Mitchell (1986) found that working relationships among team members improved when internal frames of reference were shared among team members. According to Mitchell (1986), self-disclosure of internal frames of reference provided team members with "predictive" and "explanatory" knowledge about each other. They could then use this knowledge to anticipate and understand each other’s actions. Work teams whose members understood and respected one another's alignments or internalized frameworks seemed to be able to accurately predict and explain other's beliefs and actions, as well a develop expectations concerning other team members. Testing the relationship between similar knowledge structures and team performance directly, Rentsch (1993) found significant and positive correlations between the degree to which team members had similar team knowledge structures and level of team effectiveness. Sharing teamwork knowledge may therefore help increase team effectiveness. This knowledge could be shared by having teams define teamwork, prior to working together as a team. Together team members could then construct a concept map, therefore articulating the team’s consensual teamwork knowledge prior to beginning the team’s work. The team could then make reference to the map as they worked and, together, modify their shared teamwork knowledge structures (Rentsch et. al. 1994). 

On a final note, the investigation represents a first step in developing an intercultural contingency framework for team effectiveness. In the long run, it is expected that this will improve the long term viability of international organizations. It should also be noted that analyses, interpretations, and implications drawn as a result of this research will be provided to each team participating in the study. It is expected that this process will be extremely helpful in their continued efforts to implement teams effectively. Furthermore, as the results of the program of research are disseminated in publications and at conferences, it is anticipated that the lessons learned within these organizations will provide much needed guidance to other organizations attempting to successfully adjust their team-based systems to international environments. 

522-553. 

++ Number of interviewees in Philippines = 26. ** p > .05 

+++ Number of interviewees in France = 15. *** p > .01 

++++ Number of interviewees in USA = 44. 

  
  
TABLE 6 
Frequency of Occurrence of Categories by Company 
 

Category	Company
Military	Firm 10+	Firm 11++	Firm 12+++	Firm 13~	Firm 14~~	Firm 15~~~	c 2, df = 5
Total Occurrences	5	2	3	6	1	10
Frequency per individual	.16	.25	.09	.46	.125	.60	2.417
Family	Firm 10+	Firm 11++	Firm 12+++	Firm 13~	Firm 14~~	Firm 15~~~
Total Occurrences	6	13	27	9	10	24
Frequency per individual	.19	1.63	.85	.69	1.25	1.41	14.483**
Sport	Firm 10+	Firm 11++	Firm 12+++	Firm 13~	Firm 14~~	Firm 15~~~
Total Occurrences	19	2	22	17	16	38
Frequency per individual	.61	.25	.69	1.31	2	2.24	10.527*
Associates	Firm 10+	Firm 11++	Firm 12+++	Firm 13~	Firm 14~~	Firm 15~~~
Total Occurrences	19	12	16	12	19	54
Frequency per individual	.61	1.5	.50	.92	2.38	3.18	12.043**
Community	Firm 10+	Firm 11++	Firm 12+++	Firm 13~	Firm 14~~	Firm 15~~~
Total Occurrences	22	3	41	13	9	21
Frequency per individual	.71	.38	1.28	1	1.13	1.24	25.412***
c 2, df = 4	5.183	1.436	10.543**	0.724	1.295	6.130

+ Total number of interviewees Firm 10 = 31. * p > .10 

++ Total number of interviewees Firm 11 = 8. ** p > .05 

+++ Total number of interviewees Firm 12 = 32. *** p > .01 

~ Total number of interviewees Firm 13 = 13. 

~~ Total number of interviewees Firm 14 = 8. 

~~~ Total number of interviewees Firm 15 = 17. 
  
FIGURE 1 
Proposed Model 

Interview Protocol 

1.] Could you tell us a little about what you do and the teams you work with? 

2.] Who is on the teams? How are these members selected? How are responsibilities divided? 

3.] What is the function of the teams [what outputs do they provide]? 

4.] Who is the team’s "customer" [internal or external]? 

5.] Who receives the teams’ work [who is directly downstream in the process]? 

6.] How is performance monitored and rewarded? 

7.] What kind of feedback do teams receive about performance? 

8.] How do you know when you have done a good job? 

9.] Do you believe the teams are effective? Why or why not? 

10.] Do the teams have leaders? What are the responsibilities of the leader? 

11.] Who does the team report to? Does it interact with other teams? 

12.] Would the teams benefit from more direction? Who should provide it? In what format? 

13.] What are the key factors that contribute to and/or inhibit the success of the teams? 

14.] How are practices shared in this organization? 

15.] To what extent does headquarters dictate practices? 

16.] Is individual achievement or collective achievement more important in this organization? 

17.] Is individual achievement or collective achievement more important in this country? 

18.] What facets of the culture here impact teams, either positively or negatively? 

19.] What metaphors [or mental images] do people use for teams in this country? 

20.] Do you have anything else you would like to add?