GILBERT, PELHAM, & KRULL (1988): Is cognitive attention
required to correct correspondence bias?

• Method:
  
  
  • Subjects: The subjects were 63 male and female University of Texas undergraduates.
    
    
  • Procedure:
    
    
    • Subjects were told they were participating in a study about "extemporaneous public speaking." They were told an earlier subject had been assigned to organize a pro-abortion or anti-abortion speech using the arguments in two newspaper editorials for assistance.
      
      
    • The subjects were told their task was to listen to the speech and diagnose the presenter's true attitude toward abortion. The experimenter stressed that this was a difficult task, given that the speaker had been randomly assigned a position to defend. Subjects were told "You will have to use all of your skills and intuitions as a person perceiver to figure out what he really believes."
      
      
    • Half of the subjects then received an additional instruction. They were told that after listening to the speech, they were going to be given 20 minutes to construct a speech on an assigned topic. This manipulation was designed to make these subjects cognitively busy, that is, preoccupied with what they were going to have to do.
      
      
    • All subjects then heard either a set anti-abortion speech or a set pro-abortion speech.
      
      
    • In sum, there were two independent variables in this study. What were they?
      
      
    • The dependent variable consisted of the subjects' ratings on a 1-13 scale (where higher values = more pro-abortion attitudes).
      
      
• Results:
  
  
  • Normally Attentive Subjects:
    
    
    • Pro-Abortion Speech: Attributed attitude = 8.7
    • Anti-Abortion Speech: Attributed attitude = 5.4
      
      
  • Cognitively Busy Subjects:
    
    
    • Pro-Abortion Speech: Attributed attitude = 10.6
    • Anti-Abortion Speech: Attributed attitude = 4.2
      
      
• Discussion:
  
  
  • What do these findings suggest about why correspondence bias occurs?
    
    
  • What do these findings suggest about what is required to reduce correspondence bias?