Emergence of organizational attributions: the role of a shared cognitive schema

Journal of Management, Spring, 1993 by Farzad Moussavi, Dorla A. Evans

Daft and Weick (1984) sketched the outlines of the process through which individual managers' interpretations converge into an interpretation which could be ascribed to the organization. In what we will call the convergence argument, Daft & Weick suggested that the individual-level interpretations of top strategic managers could be expected to converge into an organizational interpretation because managers use identical cognitive schemata when making their personal interpretations. When any differences in personal interpretations occur, the managers discuss the differences until a single, organizational interpretation is reached.

The convergence argument has been used as a foundation for many studies to justify selection of a collective outcome originating from individual-level cognitive processes (e.g., Dutton & Ottensmeyer, 1987; Ford, 1985; Milliken, 1987). For example, Dutton and Duncan (1987) hypothesized that the momentum to change strategic policy is influenced by managers' collective assessment of feasibility. They invoked Daft and Weick's (1984) convergence argument to fill the gap between individual managers' feasibility assessments and a collective, organizational assessment.

Rather than provoking an exploration of whether and how individual managers' interpretations converge into an organizational interpretation, the Daft and Weick (1984) argument has been used to bypass completely a description of the convergence process. Moreover, despite the significance of the convergence argument in the organizational interpretation literature, its plausibility has not been questioned. The purpose of our paper is to conceptually examine the plausibility of the convergence argument. We address a number of questions in this task. As Daft and Weick do not identify operable shared cognitive schemata, we ask whether all cognitive schemata always lead to convergence of managers' interpretations. If a cognitive schema known to describe individual-level interpretation were to be identified, how would convergence of interpretations occur? Are there conditions under which the convergence process will not yield a single organizational interpretation despite a shared schema?

As currently developed, the convergence argument cannot be readily examined because no schemata have been offered. In order to examine the argument, we needed to first elaborate it using a single, identifiable schema.

We will first explain our selection of the interpersonal attribution schema and describe it. Then, we will adapt the interpersonal schema to an organizational context, calling the adaptation the organizational attribution schema. The organizational schema will then be used to elaborate the convergence argument and to derive testable propositions. Finally, we will derive testable propositions on the conditions under which the schema will fail to bring about convergence of individual interpretations.

Interpersonal Attribution Schema--The Foundation

To gain a concrete understanding of whether and how convergence of interpretations occurs, we need a better foundation than a multitude of unidentified cognitive schemata. We need a single, fully developed schema that would guide the presumed convergence as well as our exploration.

For the following reasons we found Kelley's (1967, 1973) causal attribution schema to be appropriate. First, it is fairly well-accepted in social psychology as a cognitive schema through which individuals actually process information to make causal determinations (Weiner, 1985). To the extent that people use Kelley's schema when asking why questions, we can safely assume that it would be a schema used by managers as well. In fact, the schema has been successfully applied to explain managers' attributions of their subordinates' behaviors (Mitchell, Green, & Wood, 1981). Second, there is already research on what could be characterized as the outcome of an organizational attribution process (Ford, 1985; Staw, McKechnie, & Puffer, 1983). Choosing the attribution schema therefore will allow us to examine the plausibility of the convergence argument within a process that has already been presumed to yield a singular outcome.

Attribution theory, developed by Heider (1958), has been widely applied in micro organizational behavior (Brown, 1984; Green & Liden, 1980; Mitchell, Green, & Wood, 1981; Pearce & DeNisi, 1983; Phillips & Lord, 1980; Staw, 1975).

A basic model of the attribution theory concerns the process by which an observer diagnoses the cause of an individual's behavior (Kelley, 1967, 1973). According to this model, the observer collects information about an actor's behavior and then forms a causal attribution. The cause of the behavior may be attributed (1) to the actor, (2) to a causative entity with which the actor interacts, or (3) to the environment surrounding the actor and the behavior. In trying to find which factor caused the behavior, the observer uses a schema based on three primary informational dimensions: distinctiveness, consistency, and consensus. A behavior (1) is distinct if it uniquely occurs when a particular entity is present and is absent in the presence of other entities; (2) is consistent if the actor's reaction to the entity is the same every time and under every mode of interaction; and (3) is consensual if the behavior of other actors in reaction to the entity is the same as that of the actor.