I. Hayes and Dunning: Construal Processes and Trait Ambiguity: Implications for Self-Peer Agreement in Personality Judgments

• Understanding the Directional Bias:
  • the directional bias is the score averaged across all people for one trait
  • the directional bias should show response to desirability where agreement would show response to observation.
  • for example: an ambiguous trait that leaves more room for directional bias: 2/3 of drivers think that they are in the 80-90 percentile of good driving; an unambiguous trait that leaves room for high agreement and low directional bias: punctuality
  • less ambiguity = more agreement, less directional bias; more ambiguity = less agreement, more directional bias

• Agreements with friends versus non-friends differ in directional bias: friends show a slight directional bias across ambiguity; non-friends show that the more ambiguous the trait is, the more directional bias you get.

• Points to Ponder:
  Do people change their ratings in accordance to whether it is a friend or a non-friend? Does the self or the other change? We don't really know if effect is due to difference in self- changing or whether the peer shows more leniency in one rating than in the other.

• Moderation Effect:
  correlations between trait ambiguity and directional bias changes depending on level of friendship. There is a zero correlation between ambiguity and bias overall (table 1), but with friends there is a positive correlation and with non-friends there is negative correlation.

• Trait Definitions: Who decided how traits were defined?
  A list was provided and people put a check mark in the trait definitions that they agreed with. There was some effect of shared definitions but not a very big one.

• Some behaviors can be interpreted differently: For example, if Sam helps Sue on her take home exam, is he being dishonest or helpful? When later thinking about the person one tends to remember the trait one inferred from the behavior, not necessarily the behavior on which the inference was based.

• Different behaviors lead to the same personality judgments and the same behaviors lead to different personality judgments.

• Would length of a relationship decrease ambiguity? Probably -- yet strangers are more likely to give in than friends are with regards to definitions.

• False consensus and strangers: from the data, it does not look like false consensus creates the agreement between different judges of the same target because the judges are not similar to each other in the first place -- so even if they did use false consensus in judging the target, they would not agree with each other! False consensus is probably used, yet it probably doesn't make them any more accurate.
• Item by item scales should not be used bewcause reliability is too low.

• How is it possible that accuracy does not go up with length of relationship?
  • The results actually show very high accuracy correlations, so they may have reached the limit already
  • must remember that it is only a four scale measurement; there is a lot of "getting to know someone" that the article does not address. For example, partners become aware of certain behaviors and why these behaviors are done.
  • There may be too few "new couples" in the sample to statisitcally detect a difference between these couples and the longer-term couples. A systematic sampling of, say, 1-week, 5-week, 10-week, 50-week, 200-week couples would be better able to address the question of change of accuracy.
• Confidence and Accuracy are related on SLC (Self Esteem). This could be because this measure picks up more concrete levels of accuracy that you learn about as you get to know somebody (while also increasing your confidence in knowing the person).
• One problem of doing all these studies with college kids: They may change faster than their partern can catch up in accuracy, which puts a limit on within-couple accuracy.