This research used data from 53 high-technology firms to test three alternative models of the effects of the top management team’s demography and process on organizational performance: (1) a demography model, in which team demography accounts entirely for performance outcomes, and process has no impact; (2) a process model, in which process contributes incrementally and directly to performance outcomes, over and above the team’s demography; and (3) an intervening model, in which the effects of the top management team on performance outcomes are due entirely to the effects of its demography on process. The study found the top management team’s demography indirectly related to performance through process and process directly related to performance, although direct effects of team demography on performance were also found. These results suggest a fourth, more complex model of top management team behavior. Although some researchers have argued that leaders and top management teams have little impact on organizational outcomes (Lieberson and O’Connor, 1972; Aldrich, 1979; Astley and Van de Ven, 1983), the emerging view from more recent research suggests otherwise (Romanelli and Tushman, 1986; Finkelstein and Hambrick, 1990). Finkelstein and Hambrick (1990:500) found that in high-discretion industries, such as computers, for example, managers seem to “matter greatly.” This recent stream of research has been facilitated by Hambrick and Mason’s (1984) upper-echelons theory, which was inspired by Cyert and March’s (1963) concept of the dominant coalition. According to Hambrick and Mason’s upper-echelon theory, upper-level managers have an important impact on organizational outcomes because of the decisions they are empowered to make for the organization. Since these managers make decisions consistent with their cognitive base, which is in part a function of their personal values and experiences, their personal experiences and values can be linked to organizational outcomes. Based on this upper-echelon logic, scholars have linked top management teams to organizational innovation (Bantel and Jackson, 1989; O’Reilly and Flatt, 1989), strategy (Finkelstein and Hambrick, 1990; Michel and Hambrick, 1992), strategic change (Grimm and Smith, 1991; Wiersema and Bantel, 1992) and performance (O’Reilly and Flatt, 1989; Michel and Hambrick, 1992; Hambrick and D’Aveni, 1992). The three main clusters of concepts that are of interest in upper-echelons research are the team’s demography and process and organizational performance. Demography refers to the aggregated external characteristics of the team, such as heterogeneity, tenure, and size, while process concerns the team’s actions and behaviors, such as communication, and psychological dimensions, such as social integration. Pfeffer (1983:348) argued that demography is an important, causal variable that affects a number of intervening variables and processes and, through them, a number of organizational outcomes. Hambrick and Mason (1984) contended that a manager’s personal experiences and values can be inferred from observable demographic characteristics, such as years of experience, and that studying these observable characteristics overcomes the difficult problem of gaining access to executives to measure psychological or group dynamic variables, which may be the more direct underlying process characteristics linking the top management team’s attributes to organizational outcomes. Following Hambrick and Mason (1984), scholars have empirically linked the top management team’s demography to organizational performance (Murray, 1989; Eisenhardt and Schoonhoven, 1990; Michel and Hambrick, 1992), but no specific effort has been made to investigate the more fundamental intervening processes. If upper-echelons theory is to become useful in improving our understanding of top management teams, we need to elaborate and fully understand how team demography influences the organization. While no empirical studies have directly investigated the process through which the top management team’s demography influences organizational outcomes, several social-psychological explanations for the linkages have been proposed. Michel and Hambrick (1992) used the concept of social integration to explain links between average team tenure and diversification strategy and performance. They proposed that the length of team tenure is a proxy for the level of team cohesion and that cohesion in turn affects performance. Similarly, Murray (1989) used social integration and communication patterns to predict the form of the relationship between team heterogeneity and organizational performance. He argued that team heterogeneity may lower performance in stable environments because the team would be less cohesive and require more formal communication. Eisenhardt and Schoonhoven (1990), Keck (1991), and Hambrick and D’Aveni (1992) have all attributed findings of links between team demography and organizational performance to unmeasured social psychological concepts. The logic in these studies has been that team demography influences team processes, such as social integration and communication, and these processes in turn affect organizational outcomes. Although the study of the underlying processes through which team demography affects organizational performance would seem a critical research task, Pfeffer (1983:350) maintained that such research is unnecessary, because as soon as one says that it is necessary to understand the intervening constructs or processes one inevitably embarks on an infinite regress of reductionism from which there is no logical escape. Moreover, he argued that (1) most organizational theories are premised on a number of hypothetical constructs that are not directly observable or measurable; (2) underlying process variables are neither concrete nor unambiguous in their meaning and interpretation and often violate rules of parsimony, producing two-, three-, and even four-way interactions as explanations of behavior; and (3) the amount of variance explained by process measures is generally quite small (Pfeffer, 1983:302). He thus questioned whether intervening process variables could account for any incremental variation in dependent variables beyond that explained by demographic measures alone. In the present research we attempt to reconcile these alternative views and investigate how team demography affects performance. The primary questions are empirical: To what extent does team demography predict variation in intervening team process variables and/or directly predict variation in organizational outcomes? Second, to what extent does the addition of team process as an explanatory factor account for variation left unexplained by the team’s demography? We test three alternative models of team demography and process and organizational performance. The direct demography model hypothesizes that team demography accounts entirely for performance outcomes and that team process will have no impact on performance. In contrast, the process model predicts that demography and team process each contribute separately to performance outcomes and that team process contributes incrementally over demography. Finally, the intervening model predicts that team demography affects team processes which, in turn, influence performance. This third model posits no direct links between demography and performance and predicts that all of the effects of demography will work through process. Prior research has shown that top management teams are likely to have greatest impact on organizational outcomes in high-velocity environments (Eisenhardt, 1990), where managers have more decision discretion (Hambrick and Finkelstein, 1987; Finkelstein and Hambrick, 1990). Therefore, the models were tested with a sample of technology-intensive companies from a high-discretion, high-velocity environment.