عدم تقارن در انطباق ساختاری : تاثیر افتراقی متمرکزسازی ساختارهای تصمیم گیری در مقابل تمرکززدایی ساختارهای تصمیم گیری تیم

This study tested predictions derived from Structural Adaptation Theory (SAT) on the longitudinal effects of centralizing and decentralizing decision-making structures in teams. Results from 93 four-person teams working on a command and control simulation generally supported SAT, documenting that it was more difficult for teams to adapt to a centralized decision-making structure after formerly working within a decentralized structure, than it was to adapt in the alternative direction. The negative effects of centralized shifts were mediated by efficiency and adaptability, in the sense that former decentralized teams experienced the negative aspects of centralization (lack of adaptability), but not the positive aspects (efficiency). The dangers of employing structural reconfiguration to solve certain problems in teams are discussed, especially if these changes are based upon expectations generalized from cross-sectional research that did not directly observe teams that experienced true structural change.

Teams have been defined as small groups of interdependent individuals who share responsibility for specific outcomes (Sundstrom, De Meuse, & Futrell, 1990, p. 120). Team-based structures have played an increasingly important role in contemporary organizations, and longitudinal surveys of Fortune 1000 firms have shown a steady increase in the use of team-based structures moving from less than 20% in 1980, to roughly 50% in 1990, to over 80% in 2000 (Garvey, 2002). This has prompted a great deal of research on teams, much of which has focused on either the impact of various dimensions of team structure or team processes and outcomes (Ilgen, Hollenbeck, Johnson, & Jundt, 2005). One of the most critical aspects of structure that has to be determined within any work group is where the locus of formal authority for decision-making is going to reside. Members of the group may have different ideas about what they each should be doing based upon variability in their experiences, preferences, knowledge, or information held. In centralized structures, authority is concentrated at the top of the team and a formal team leader has responsibility for making decisions. In contrast, in decentralized structures, authority and decision-making responsibility are dispersed downward and outward through the hierarchy, and individual team members are empowered to make their own decisions (Pugh, Hickson, Hinings, & Turner, 1968). A great deal of research has been conducted on the virtues and liabilities of alternative team decision-making structures (Bonaccio and Dalal, 2006 and Ilgen et al., 2005), and several formal theories have been developed that describe why, when, where, and with whom various structures work well or work poorly (Burns and Stalker, 1961, House, 1971, Pennings, 1992 and Vroom and Yetton, 1973). All of these theories emphasize that there is “no one best way” to structure decision-making in teams, and they all describe contingencies associated with why one structure or another is best depending upon the team’s goals and strategy. These theories have been supported by empirical research and are routinely incorporated into textbook treatments of this topic, as well as popular press accounts of management teams. For example, if one were to peruse the popular press, it is clear that when it comes to centralization and decentralization, different organizations are moving in diametrically opposite directions for managing this aspect of structure. Within the United States government, the Department of Homeland Security represented a major centralization of formerly autonomous units (Peters, 2004) whereas the Veterans Administration’s new system of publicly supported, but regionally autonomous health centers reflected a move toward decentralization (Rogers, 2005). In the technology sector, IBM has moved in the direction of centralizing operations (i.e., creating a small number of super-regional centers that replaced a larger number of smaller and locally managed units; Flynn, 2000), whereas Microsoft has moved toward decentralizing operations (i.e., creating separate business units, such as Home Entertainment, Internet Services, Operating Systems, etc. that generate their own profit and loss statements; Greene, Hamm, & Kerstetter, 2002). In retail, Home Depot centralized purchasing operations that used to take place at nine different regional centers into one single location in Atlanta, while its primary competitor, Lowes, decentralized the same purchasing decisions to regional directors (Foust, 2003). On the one hand, even though these organizations are moving in opposite directions, all this movement from one structural alternative to another could be rational in the sense that both formal theories and empirical evidence suggest there is “no one best structure” when it comes to centralization versus decentralization (e.g., Burns and Stalker, 1961 and Pennings, 1992). At both the organizational level and the team level, one can find conceptual justification and empirical data from cross-sectional studies that suggest that each alternative has its own set of virtues and liabilities (e.g., Drazin and Van de Ven, 1985 and Hollenbeck et al., 2002). If it is true that one structure is associated with one desirable outcome, whereas the other structure is associated with a different desirable outcome, then it might be tempting for an organization or work group to change their structure if there was a corresponding change in goals or strategies. On the other hand, the cross-sectional and static nature of the current set of theories and data on this topic does not logically support the dynamic generalization that is often inferred in both the popular press and the academic literature. That is, one can accept the static conclusion that one structure is best for one outcome whereas the alternative structure is best for a different outcome, and still challenge the dynamic conclusion that if a group or organization has a change in goals or strategies they should change their structure. This dynamic generalization implies that “history does not matter” and that the experience of having worked within one structure has no impact on the ability of one to adapt to the alternative structure. However, emerging evidence suggests that in most group contexts, history does matter and past experience has a pronounced impact on future processes and outcomes (McGrath, Arrow, & Berdahl, 2000). More specifically, when it comes to changes in team structure, research based on Structural Adaptation Theory (SAT) has shown directly that it is more difficult for teams to change their structures in some directions than in others. For example, the major dimensions of team structure work to create teams where the members are either highly tightly coupled or loosely coupled (Orton and Weick, 1990 and Weick, 1976), and one of the major propositions of SAT is that is easier for teams to shift into more loosely coupled structures relative to more tightly coupled structures. For example, research on SAT has found that it is more difficult to shift a team’s task allocation structure from a loosely coupled divisional scheme to a tightly coupled functional scheme than it is to change their structure in the alternative direction (Moon et al., 2004). Similarly, with respect to reward structures, the evidence suggests that it is more difficult for teams to change from a loosely coupled individually-based system to a tightly coupled group-based system than it is to make the shift in the opposite direction ( Johnson et al., 2006 and Beersma et al., 2009). The purpose of this study is to extend this line of research to the dimension of decision-making structure, and we present conceptual arguments and empirical data that challenge the notion of symmetry in structural movement, documenting that it is more difficult for teams to shift from a decentralized to centralized structure than it is to shift in the opposite direction. This has theoretical implications for extending the breadth of Structural Adaptation Theory, but also has practical implications for organizations that might be contemplating changing their structures, hoping to accrue benefits promised by current contingency theories that may never materialize.

There is no “one best way” to design a group’s or organizations decision-making structure. There are virtues and liabilities associated with both centralized and decentralized systems, and several different contingency theories that spell these out in some detail. For this reason, as in the examples that we cited at the beginning of this article, it is natural for organizations to consider changing their structure from one system to another if they change their goals or strategies. However, our results suggest that, in line with Structural Adaptation Theory, some changes are more natural relative to others. Groups that move from a more highly ordered state to a less highly ordered state (centralized to decentralized) seem to derive the benefits of the new structure but not its traditional liabilities. In contrast, groups that move from a less ordered state to a more highly ordered state are attempting to reverse entropy, and struggle with this specific type of adaptation. As a result, the static predictions made by many contingency theories such as the Vroom–Yetton Model, Path Goal Theory, and Structural Contingency Theory, fail to hold up as teams dynamically shift from one structure to another.