• Judgment of others' personality (Catrambone & Markus)
• Prediction what others will do (Ross et al, Dawes)
• Ascribing mental states (Gordon)
• Explaining action (Gordon)

Research Traditions

• Freud: "Projection" (unacceptable impulses or thoughts are denied as being part of the self and are instead assumed to exist in the other person)
• Personality judgments
  1. Standing on trait vs. use of trait.
     There is little to no evidence that people's own standing on a trait influences their judgments of other people's standing on that trait. However, there is good evidence that many of the traits people use to describe others are actually traits that they find important in the self.
  2. Confidence vs. accuracy
     Catramabone & Markus show that people are more confident in judging others on traits that are important ot the self than on traits that are unimportant to the self. However, there are no comparable accuracy effects.
  3. Availability and Priming
     The major mechanism that drives the above effects is probably "availability" -- that is, the fact that traits about which you think a lot are "available to the mind" and are more likely to be used in descriptions of others. One reason why you may think about a trait a lot is that it is important to your self-concept; hence the effect of self-relevance of a trait on its use in describing others. Priming is a variation of availability. For example, if you have just thought about yourself, the traits on which you describe yourself have been "primed." No wonder, then, that you would use those traits in describing other people as well.
  4. Conditions: Perceived similarity and salience of self
     One condition under which self is used as a reference in judgments of personality is that the other person is perceived as similar. So you would use self-relevant traits to describe another college student but less likely so to describe an older person or a child. A second condition under which self is used as a reference in judgments of personality is that the self is made salient, i.e. "activated." Such activation can happen via a newspaper article, a request to describe yourself, an important personal experience or a mirror/camera (which increase one's self-awareness).

• "False consensus bias" in population prediction. First discussed in the literature by Ross, Greene, & House (1977). This "bias" refers to the fact that when one has to predict other people's behavior (e.g., whether or not they would walk around with a sign that says "repent!"), one uses one's own inclination to act as a reference. So people who would walk around with the sign predict that many others would too, whereas people who wouldn't walk around with the sign predict that few others would.

  That this tendency is perfectly defensible (and not a bias) was argued by Dawes (1990). His arguments are as follows:

  1. Symmetry of inference (many to 1, 1 to many). Just as it is defensible (and recommended) to use the population "base rate" (the rate of people in a population who would perform a certain behavior) to predict an individual's behavior, so is it defensible to use an individual's behavior to predict the population. The second predict is necessarily less reliable, but it will on average be correct, especially if the population rate is far from 50%.
  2. Diagnosticity of single response. Dawes argues in several ways that a single response is "diagnostic" (i.e., useful and decently representative) for prediction of the whole population. (it's not hugely diagnostic, but it's not non-diagnostic either.) One argument is that "Baysian reasoning" (i.e., reasoning according to the laws of probability, especially conditional probability) would demand that you take a single response into account. If you have a prior belief that 40% of people like newly introduced ice-cream flavors and then learn that one person liked the new "lilac" ice-cream, you may adjust you belief (now called "posterior" belief) to, say, 43%.
  3. Dawes then goes on to say that if any one individual's response is diagnostic (and should be taken into account), then why not your own? After all, you are a member of the population just as everyone else is.
  4. For base rates that differ from 50%, the population base rate correlates with single responses. Therefore, any single response is a better-than-chance predictor of the population rate. Imagine a 85% base rate. That is, 85% of people would perform behavior B. Now if those 85% of people would all base their population estimate on their own inclination to perform B, they would often be accurate; only the minority of 15% (who will base their population estimate on their non-inclination) would often be inaccurate.

• Simulation theory vs. "Theory" theory
  "Theory" theory claims that children (and adults) figure out other minds by applying a general theory to the problem -- a theory that contains abstrract rules (e.g., that actions are based on beliefs and desires, that crying stands for sadness, etc.). Simulation theory claims that people figure out other minds by (a) projecting their own view of the situation onto others (total projection) and (b) simulating what it would be like to be in the other person's situation (perspecitve taking, which includes "adjustments" for assumed differences between self and other).