نقش، عملکرد و سهم نظریه اسناد برای رهبری: بررسی

The purpose of this article is to review literature that has focused on the role of attributions in leadership processes and to explore and explain how the study of attributions does, and can, contribute to our understanding of the dynamics of leadership. The historical roots of attribution research are discussed, along with early attributional research in the leadership area. Two streams of attributional criticisms are addressed and recent attributional research relevant to leadership is reviewed. We argue and demonstrate that attributions account for significant proportions of the variance in leadership behaviors. We conclude with suggestions for including attributional perspectives in comprehensive models of leader behavior.

The purpose of this article is to review literature that has focused on the role of attributions in leadership processes and to explore and explain how the study of attributions does, and can, contribute to our understanding of leadership. There are at least three reasons why such a review and discussion is important to the field of leadership at this time. First, there have been more than 50 published studies concerned with the role of attributions in leadership processes since the late 1970s. Although there have been some general reviews of the role of attribution theory in the organizational sciences (Martinko, 1995, Martinko, 2004 and Martinko et al., 2006), we are unaware of any comprehensive review that focuses on the role of attributions in leadership processes. More importantly, there appears to have been little systematic effort to synthesize and understand this relatively large and significant body of research within the context of leadership. Second, as will be discussed below, there appears to be some confusion regarding the definition and role of attribution processes. In some cases it appears that the term attribution has been viewed as synonymous with perceptions, while in others attribution theory appears to be constrained within the domain of the theory originally suggested by Heider (1958), Weiner, Frieze, Kukla, Reed, Rest, & Rosenbaum (1971) and Kelley (1967). Clearing up these definitional problems and domain issues is important because there have been criticisms of attribution theory in the context of leadership processes (e.g., Cronshaw & Lord, 1987) that appear to be directed more toward perceptual processes in general rather than toward the domain that is typically considered to be encompassed by attribution theory. Misunderstandings of these criticisms need to be cleared up because they may lead to the rejection of attributional perspectives in situations where they can make valuable contributions. Finally, as Weiner (2004) has noted, the extension of attribution theory into areas such as leadership has “stretched” attribution theory and has resulted in extensions of the theory that contribute to new research contexts, and also enhance and enrich the more general domain of attribution theory. We believe that it is important to point out these contributions. We begin with a brief discussion of the historical roots of attribution research to clarify what we consider to be the domain of attribution theory. We then begin our review of the published literature within a historical context, starting with the work in the early 1980s conducted by Mitchell and his colleagues. Following the review of that body of literature, we examine the criticisms that have been directed toward attribution research, particularly those of Mitchell (1982), who was critical of his own work, and those of Lord and his colleagues (e.g., Cronshaw & Lord, 1987), which appear to hold attribution theory accountable for processes that are beyond its domain. Although there appeared to be a decline in attribution research in the leadership domain immediately after those criticisms, a steady stream of research has developed and continued through recent years. In the latter part of the manuscript we review that research, integrate it with the prior work, and describe how and where it contributes and can add to the domain of leadership research. Before beginning our review, it is appropriate to delineate the criteria we used to select the articles for review. First, we wanted to make sure that we kept the scope of the review manageable and focused on leadership processes. Therefore, the articles we included had to either discuss or investigate leadership and attribution processes. Articles that discussed more general motivational processes such as goal setting, impression management, and employee selection were not included unless they were explicitly concerned with leadership processes within an attributional context. In addition, we did not consider articles that discussed leadership as a causal explanation for organizational performance (e.g., Lord, 1985, Staw et al., 1983 and Staw, 1975) in detail in this review because they are not concerned with leadership processes. Second, we searched the literature published in the major journals from 1975 to the present. Included among these journals were the Academy of Management Journal, The Leadership Quarterly, Academy of Management Review, Journal of Applied Psychology, Organizational Behavior and Human Performance, Organizational Behavior and Human Decision Processes, Journal of Organizational Behavior, Journal of Management, Administrative Science Quarterly, and Personnel Psychology. Finally, we searched major databases including Google Scholar, Expanded Academic ASAP, JSTOR, ABI/Inform, and APA PsychArticles.

In this section we summarize the major findings from our review, address the limitations and boundary conditions of attribution theory within the context of leadership, and point to future directions that can enhance the contributions of attribution theory to the explication of leadership processes. As we described, much of the early research relating leadership and attribution processes focused on the Green and Mitchell model. The early work on this model conducted by Mitchell and his colleagues provided general confirmation of the model but also demonstrated that many factors besides attributions affected leadership behaviors directed at members. Examples of these factors included the severity of the consequences of poor performance, task interdependence, and the experience of the leader. The effects of these other factors had been anticipated in the explication of the Green & Mitchell model (1979). However, despite the supportive findings regarding the attributional component of the model, Mitchell (1982) expressed reservations about the potential of attribution theory to predict practically significant portions of the variance in leader behaviors. The volume of attributional work by Mitchell and his colleagues declined markedly after this critique. After the work of Mitchell, Ashkanasy engaged in a series of studies that clearly demonstrated the relations between Kelley's dimensions of information, Weiner's attributional dimensions and explanations, and leader behaviors. In addition to the work of Ashkanasy, about a half dozen other studies had similar supportive results. Ashkanasy's conclusions regarding his work are similar to our conclusions. The overall results of the work on the Green and Mitchell model are robust and support the major tenets of the model across a variety of different samples and methods. In virtually every study, there were significant relations between and among informational dimensions, attributional dimensions, attributional explanations, and leader behaviors. In expanding our review beyond the Green and Mitchell model we found consistent support for the notion that both leaders and members display the self-serving and actor–observer biases. There was also significant support for the notions that age, gender, culture, and individual differences including locus of control, negative affectivity, and self-efficacy are related to attributions. The work on attribution styles demonstrated that, when the attribution styles of leaders and members are incompatible, members report poor leader–member relations, suggesting important practical implications for differences in attributions between leaders and members. We also reviewed a series of empirical studies and conceptual works that examined and considered the role of attributions in forming members' impressions of charisma, transformational leadership, organizational cynicism, and dues paying. In contrast to much of the prior work on leadership, this body of work suggests that, as opposed to the effortful and rational process described by the Green and Mitchell model, the process by which attributions are generated in these contexts is more appropriately characterized as automatic and cybernetic. The discussion which followed suggested that these types of social perceptions are probably best explained by theories of categorization and template matching (i.e., automatic processing) as opposed to the more cognitively laborious processes described by the Kelley and Weiner models. The emerging research on the relationships between emotions and attributions in the context of leadership was also considered. This research demonstrates how attributions are related to emotions such as sympathy and anger which generate helping behavior (Weiner, 1995) and punishment (Tjosvold, 1985) respectively, but also shows how leaders' displays of emotions affect member attributions (Dasborough & Ashkanasy, 2002). We also looked at research that took a more interactive approach looking at the relationships between the attributions of leaders and members. Again, this research supported the notion that attributions accounted for significant portions of variance but also suggests that more recent works are beginning to recognize and focus on an interactive and dyadic level of analysis as opposed to much of the earlier work that was leader-centric. We expect this trend to continue and enable a more complete understanding of the causes and consequences of differences between the attributions of leaders and members. Finally, we looked at conceptual advances in two general categories. First we observed that there had been considerable work extending the Green and Mitchell model by further articulating the relationships between and among Kelley's dimensions of information, Weiner's attributional dimensions and explanations, and both leader and member behavior. These extensions appear helpful in increasing the precision of the predictions of the model, but also in developing applications designed to enhance leader and member relations. The second group of conceptual works focused on alternative explanations for attributional processes focusing on cybernetic “mindless” processing. As we discussed these perspectives it became more evident that they appear particularly appropriate for the work concerned with social impressions such as those associated with charisma and transformational leadership. When we view this body of research as a whole, three impressions are most salient. First, there is strong support for the relationships between and among Kelley's dimensions of information, Weiner's attributional dimensions and explanations, and leaders' behaviors as suggested by the Green and Mitchell model, particularly when it is applied to leaders' assessments of the causes of members' poor performance. Second, when the above perspective is stretched to other contexts it does not appear to provide a valid description of either leaders' or members' cognitive processes. Thus, in routine situations that are not particularly meaningful or related to a specific performance incident, and when we consider social impressions such as those associated with charismatic and transformational leadership, the less cognitively effortful processes appear to provide better explanations for experimental results. Finally, we believe that there is some definitional and domain uncertainty with respect to attributions and attributional processes. More specifically, it appears that when researchers describe alternatives to rational information processing models they also tend to discount more rational processes. Thus, even though Lord (1995, p. 348) indicates that rational models may be most appropriate for assessing causality, he ends his discussion by indicating that “to the extent that these alternative models describe causal analysis and social perceptions in organizations, the perceiver-as-primitive-processor-analogy may be more appropriate.” In the next section we will address this criticism as well as other criticisms of attribution theory, discuss boundary conditions for a rational model of attributional processes, and suggest directions for future research.