JONES & HARRIS (1967): Early Evidence of Correspondence Bias

• Method:
  
  
  • Subjects: Thirty-six male and fifteen female Duke University students taking an introductory psychology course.
    
    
  • Procedure:
    
    
    • Subjects were given a short essay to read that was described as having been prepared by another undergraduate in a "creative writing course."
      
      
    • The subjects were told their task was to make "judgments of another's personality and attitudes on the basis of very limited information."
      
      
    • The subjects were assigned to one of four conditions defined by the crossing of two independent variables: essay choice (whether the position taken in the essay was freely chosen vs. assigned to the essayist by the course instructor) and behavior direction (whether the essay was pro-Castro or anti-Castro).
      
      
    • After reading an essay in one of the four conditions of the study, all subjects were asked to estimate the essayist's "true attitude toward Castro" using a 10-item scale. These ratings constituted the primary dependent measure in the study. Scores on this measure could vary from 10 (negative attitudes toward Castro) to 70 (positive attitudes toward Castro).
      
      
    • Please be certain that you are clear about the differences between independent and dependent variables in an experimental design.
      

    
    

• Results:
  
  
  • Choice Conditions:
    
    
    • Pro-Castro Essayist: Attributed attitude = 59.6
    • Anti-Castro Essayist: Attributed attitude = 17.4
      
      
  • No Choice Conditions:
    
    
    • Pro-Castro Essayist: Attributed attitude = 44.1
    • Anti-Castro Essayist: Attributed attitude = 22.9
      
      
• Discussion (Consider each of these in turn):
  
  
  1. What do these findings suggest about the existence of correspondent bias?
     
     
  2. To what real-life equivalents does the situation presented in the Jones and Harris study seem most directly similar? Click here for some suggestions.