Covariation Model
From PsychWiki - A Collaborative Psychology Wiki
Every day we are constantly making choices, impressions, and forming perceptions regarding situations and people. Although there are different ways to formulate perceptions, a specialized theory states that these judgments are not solely based on one factor. The covariation model by Harold Kelley, states that an individual will examine more than one occurrence of behavior at different times, and in different situations to answer a question. This is a causal theory that uses questions like “why” to answer the question in mind (Kelley, 1973). Kelley (1973) also states that covariation is what exists between what is being observed and the possible causes. Typically the behaviors are associated to a particular stimulus (Orvis, Cunningham, & Kelley, 1975). The causes can be a person, an entity, or different circumstances (Forsterling, 1989). The model is essentially based on three factors: consensus, distinctiveness, and consistency (Aronson, Wilson, & Akert, 2007, p.106). By being able to identify why people acted the way they did, one will be able to come to a conclusion as to why the behavior occurred.
The first factor that helps answer the question in mind is consensus information. This factor answers how other people behave toward the same stimulus. For example, imagine a teacher yelling at Mary. Using the consensus information one asks if Mary’s other teachers yell at her as well. Distinctiveness information answers how the person whose behavior is in question responds to other stimuli. Does the teacher yell at other students in Mary’s class, and students in her other classes? Lastly, the consistency information deals with the consistency of the behavior between the person in question and the same stimulus. Does the teacher always yell at Mary? The possibility of high or low of these the three factors, consensus, distinctiveness, and consistency, can combine into eight different possibilities (Forsterling, 1989).
Kelley’s theory states that if there is congruence of the three factors into two distinct patterns, an internal attribution can be clearly made (Aronson, Wilson, & Akert, 2007, p.107). Although the covariation model is to make causal attributions, people do not use the consensus information as readily as much as they use the consistency and distinctiveness (Aronson, Wilson, & Akert, 2007, p.107).
Example - Research
In an experiment conducted by Eastman (1994) in order to test what the true intention of employees was when performing well and in an extrarole manner. In this experiment, extrarole meant that employees went above and beyond their normal work responsibilities. The hypothesis of this study was that employees would be labeled as good citizens when their extrarole had high consistency, low distinctiveness, and high consensus. In other words, the good performance of the employee would be considered to be an attribution of the employee rather than their interest in promotions or other external factors if the other employees had the same work ethics, if all the different type tasks the employee did was of the same merit, and lastly if their excellent work was frequent across time.
The participants in the study were of a mean age of 40, all enrolled in a Masters program for Business or part of a business management association. The study consisted of a different subgroup, 55% male and 45% female, which made the study valid. The operalization of consensus was whether other employees acted the same way. Distinctiveness was whether the employee had acted that same way toward other employees and someone other than the boss. Consistency was the frequency in which the employee behaved the way they did, if the employee’s actions were stable and frequent across time. All three factors were all manipulated through behavioral logs. This experiment was a two two-by-two-by-two experimental design that consisted of eight possible scenarios that involved: high consensus, high distinctness, and high consistency, to low consensus, low distinctiveness, and low consistency.
Each of the respondents was assigned the role of supervisor of five customer service representatives which were all designated the letters A thru E. The respondents assigned the role of the supervisor were then handed five behavioral logs that would account for one year. The supervisors had to rate and evaluate the work of the employees.
The respondents were given envelopes with 95 statements consisting of job-related behaviors like “Told my boss I liked the way the company was run”. The logs maintained consistency, distinctiveness, and consensus manipulation by changing neutral statements with extrarole behaviors. Logs that represented high distinctiveness contained extrarole behaviors that were only directed at a supervisor, and low distinctiveness logs contained extrarole behaviors directed at the supervisor and others. High consistency and distinctiveness were behaviors that were evenly distributed throughout 52 weeks. The respondents used a 1 to 5 rating scale; meaning 1 was poor and 5 outstanding. The respondents had $6,200 to distribute among their 5 employees that displayed the most outstanding work.
The results of the study showed that consensus, distinctiveness and consistency did not all together influence the supervisors. The factor that influenced the supervisor the most was consensus. As stated in the experiment, although others may have acted the same way over time they were still able to arise different causal attributions.
The criticism of this experiment is that even though the results showed that employees were viewed differently, an employer may have some sort of bias towards some. The implications of this research study are that extrarole behaviors can be perceived differently by the supervisors.
Example - Real-life
Imagine the example given by Orvis, Cunningham, and Kelley (1975) where Dr. Stanton complimented Barry’s work. Although Dr. Stanton complimented his work, everyone compliments Barry’s work making this high consensus. We know that Dr. Stanton does not compliment everyone’s work, making his action of complimenting Barry highly distinctive. We also know that in the past Dr. Stanton has always complimented Barry’s work, which results in high consistency. According to Kelley’s model, as a result of all three components being high, it can be assessed that it was an external attribution (Aronson, Wilson, & Akert, 2007, p.107). By making an external attribution, we believe that it was something that Barry did to make everyone compliment his work, or that his work alone was praise worthy.
An internal attribution can be made if the contrast is found about Barry. There is low consensus if hardly anyone compliments Barry. There is low distinctiveness if Dr. Stanton compliments everyone’s work. Lastly, there is high consistency if in the past Dr. Stanton has always complimented Barry’s work. There can be an attribution made about Dr. Stanton rather than Barry because Dr. Stanton compliments everyone.
This example displays the theory of covariation using external and internal attributions. Although this model can be used to make attributions about situation and people, sometimes the attributions can be wrong (Aronson, Wilson, & Akert, 2007, p.108). At times there may be circumstances, which are not foreseeable, or have known reasons for the behavior.
The research which has been done on covariation model and attribution reports have helped in the conquests of answering questions regarding internal and external motives which underlie in situations. This model is also a step further in the relation of different stimulus, actors and the frequency of time. With Social Psychology on the move, and models like these are becoming significantly more abundant, society is moving towards the integration of discovering more about itself in order to facilitate learning about questions and problems at hand.
References
Aronson, E., Wilson, T.D, & Akert, R. M. (2007). Social psychology. Upper Saddle River, NJ: Pearson Education Inc.
Eastman, K. (1994). In the eyes of the beholder: An attributional approach to ingratiation and organizational citizenship behavior. Academy of Management Journal, 37, 1379-1391.
Försterling, F. (1989). Models of covariation and attribution: How do they relate to the analogy of analysis of variance? Journal of Personality and Social Psychology, 57, 615-625.
Kelley, H. (1973). The processes of causal attribution. American Psychologist, 28, 107-128.
Orvis, B., Cunningham, J., & Kelley, H. (1975). A closer examination of causal inference: The
roles of consensus, distinctiveness, and consistency information. Journal of Personality and Social Psychology, 32, 605-616.
◄ Back to Fall 2009 PSY 322 Assignment page
