Kelley’s Attribution Theory

Adapted from Mowday, R.T. Beliefs about the Causes of Behavior: The Motivational Implications of Attribution Processes. In R.M. Steers and R.T. Mowday (Eds.) Motivation and Work Behavior, 1987.

The attribution process involves reasoning backward from the observation of an event or behavior to a judgment about its cause. Unlike a prediction, the attribution process attempts to provide an explanation for an event that has already occurred. The best known work on attribution is Kelley’s theoretical approach that is based on two important contributions to our understanding of cognitive processes associated with forming causal beliefs. First, the principle of covariance states that a behavior will be attributed to a cause with which it covaries over time. Second, he identified three sources of information people can use in analyzing covariance and thus arriving at a causal judgment. More specifically, he suggested that individuals form causal beliefs by analyzing the consistency, consensus, and distinctiveness of a response or behavior.

Observations of a behavior or response across time provide information about consistency. For example, does an employee express dissatisfaction every time he or she is assigned to a specific task or only on some occasions? Observations of different people allow judgments to be made about consensus. Do all individuals assigned to a particular task express dissatisfaction or is it just one individual who indicates unhappiness? Finally, observations of entities or stimuli provide information about the distinctiveness or a response. Does an employee express dissatisfaction when assigned to many different tasks of just one particular task?

An important consideration in Kelley’s theory is how individuals process and combine these three sources of information for purposes of making a causal judgment. He suggested that cognitive process can be represented by a 2 x 2 x 2 analysis of variance framework. If we simplify the cognitive task by assuming that each type of information can take either a high or low value (e.g., high or low distinctiveness), this representation leads to eight unique combinations of information. Predictions cannot easily be made about the types of attributions associated with each of the eight information combinations. However, three information combinations lead to intuitively clear predictions. These predictions are illustrated in the table below:

 

Information Combination

Predicted Attribution

Example

High Consistency

Low Consensus

Low Distinctiveness

Personal Cause

Situation: Steve Jones expresses dissatisfaction every time he is assigned the task of proofreading. Other employees are not dissatisfied when asked to proofread. Steve expresses dissatisfaction on just about every task he is assigned.

Perceived cause: Most likely cause of Steve’s dissatisfaction is something to do with his personal characteristics (e.g., he is a chronic complainer.

High Consistency

High Consensus

High Distinctiveness

Environmental Cause

Situation: Steve Jones expresses dissatisfaction every time he is assigned the task of proofreading. Other employees also express dissatisfaction when asked to proofread. Steve is not dissatisfied with other tasks to which he is assigned.

Perceived cause: Most likely cause of Steve’s dissatisfaction is something to do with the task itself (e.g., proofreading is boring).

Low Consistency

Low Consensus

High Distinctiveness

Circumstantial Cause

Situation: Steve Jones expresses dissatisfaction when he is assigned the task of proofreading. Other employees are not dissatisfied when asked to proofread. Steve has not expressed dissatisfaction when assigned to proofread in the past.

Perceived cause: Most likely cause of Steve’s dissatisfaction is something related to the momentary circumstances surrounding his task assignment (e.g., he does not feel well, does not like the way he was asked to do the job by his supervisor, etc.)