Part D4. GILBERT, PELHAM, & KRULL (1988): Explaining Correspondence Bias

Overview: Why do people show a correspondence bias in social perception? Given the significance and pervasiveness of correspondence bias, that’s an important question to ask. What’s found may reveal larger features of how human cognitive processes operate and may possibly help perceivers control the tendency. This section of the module will not, however, present a definitive account of correspondence bias; it’s not clear that a single overarching determinant of correspondence bias exists (e.g., Gilbert, 1988; Jones, 1979). What follows offers one example of how research studies based on the Jones and Harris paradigm have attempted to articulate and evaluate potential causes. It’s inspired by a model advanced by Gilbert (1988, 1991), who suggests dispositional inference is an automatic response when making sense of the behavior of others and that a more effortful and deliberate correction process may then adjust an initial view in light of additional considerations (e.g., a weighing of situational constraints on behavior). The study summarized below, one of two reported in Gilbert, Pelham, and Krull (1988), attempts to demonstrate that reducing the cognitive resources available to participants to potentially “correct” their automatic inferences will magnify the degree of correspondence bias shown.

Method:

Participants:

Sixty three male and female University of Texas undergraduates.

Procedure:

• Participants were informed they were in a study about “extemporaneous public speaking.” They were told an earlier participant had been assigned to organize a pro-abortion or anti-abortion speech using arguments from two newspaper editorials.

• The participants were asked to listen to the speech and diagnose the speaker’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. Participants were told “You will have to use all of your skills and intuitions as a person perceiver to figure out what [the speaker] really believes” (p. 737).

• Half the participants 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 participants cognitively busy, that is, preoccupied with what they themselves were going to have to do.

• All participants then heard either the anti-abortion speech or the set pro-abortion speech in which the position taken was said to have been assigned to the speaker.
  
  
• The dependent variable consisted of the participants’ ratings of the true attitude of the speaker on a 1-13 scale, where higher values indicated more pro-abortion attitudes.

Results:

• Normally Attentive Participants:
  • Pro-Abortion Speech: Attributed attitude = 8.7
  • Anti-Abortion Speech: Attributed attitude = 5.4
    
• Cognitively Busy Participants:
  • Pro-Abortion Speech: Attributed attitude = 10.6
  • Anti-Abortion Speech: Attributed attitude = 4.2

__________________________________________________________________________

RESULTS REVIEW AND DISCUSSION QUESTIONS FOR STUDENT CONSIDERATION:

1. What does the comparison between the degree of correspondence bias shown by the normally attentive participants vs. the the cognitively busy ones suggest? How does this outcome fit Gilbert’s automatic inference and possible correction account of correspondence bias?
   
   
2. The Gilbert model assumes the automatic inference default is a correspondent or dispositional account of behavior, but why should that be? Try to identify possible reasons perceivers may be inclined to make correspondent inferences before engaging in possible adjustments or corrections. [Some discussion of this follows in Note 3 below.]
   
   
3. Under the normal research circumstances of the Jones and Harris paradigm, participants are not explicitly led to experience cognitive busyness and yet they still show correspondence bias. Why might this be so? Do participants lack the motivation or capability to adjust their automatic responses correctly? Are participants normally disinclined to engage in correction or are they otherwise cognitively busy? Do these challenges to corrective activity apply as well when perceivers try to make sense of human behavior outside the research situation?
   
   
4. What are some other possible explanations of correspondence bias? How might those be tested using variants of the Jones and Harris paradigm?
   
   

__________________________________________________________________________

NEXT STEP:

To return to the entry page for the four scientific applications of the Jones and Harris paradigm (the Part D overview), click here.

Alternatively, to end your exploration of the Module and move to the concluding Module assessment page, click here.

__________________________________________________________________________

Notes:

1. For an advanced review of possible causes of correspondence bias, see Gilbert and Malone (1995).

2. Gilbert’s automatic inference/correction account bears significant resemblance to the broader view of cognition summarized by Daniel Kahneman (2011) in Thinking Fast and Slow as well as, not surprisingly, to Gilbert’s own provocative (1991) perspective on belief.

3. Why might correspondent (dispositional) social judgments be basic? (a) Perceivers may attend to an outcome associated with a person (e.g., see a person acting) and this perceptual reality may be sufficient to produce an attribution. Our attention to persons makes sense in light of their intentionality and importance to our own immediate fates. Moreover, the influences associated with situations and settings may be less perceptually salient and concrete. (b) Perceivers may assume people believe what they say and act as they prefer. Whether this is generally accurate is another and important matter. It may be this assumption is validly grounded in everyday experience, or it may be a common misconception. Nonetheless, the base rate assumption of intentionality may be a powerful feature of interpersonal judgment. (c) Perceivers may have been socialized to hold others and themselves responsible for their behavior and lots in life. Correspondence inferences may reflect this cultural influence and potentially vary culturally as a result. See Part D3 of the Module on potential cultural variation in corrrespondence bias. (d) Alternatively, dispositional judgments may not be basic but the experimental situation itself may focus perceivers’ judgments by directing participants to diagnose the essayists’ true attitudes given their behavior.