by Yating Chang & Leslie Claridge

Problem: The communication hypothesis and the overlap hypothesis seem to contradict each other (p. 661):

Overlap hypothesis: Judges agree with each other only because they observe the target in the same context.

Results: It is true that judges agree better if they observe the target in the same context (e.g., at school) than in different contexts (e.g., at school vs. at home); but they agreed with each other even when they did not observe the target in the same context.

Communication hypothesis: Judges agree because they communicate with each other about the target and therefore develop a shared representation of the target.

Results: Those who had never met and those who had met showed about the same level of agreement, disconfirming the communication hypothesis

Thus, the two hypotheses refer to different comparisons (same vs. different context; met before vs. not met before).

Puzzle: Average correlation for the met and never met conditions under communication hypo. is low (only about .20). Why?

Observation: Generally, the self-other agreement correlations are higher for college friends and parents than for hometown friends. This can be caused by:

• a recency effect exists among college friends: self & college friends have seen each other more often in recent times
• parents and self enjoy a longer history than hometown friends
• college friendships exist in the same context, hometown friendships do not anymore
• hometown friendships may not be kept up as consistently as parent-child relations

Criticism: The motivation of participants filling out the questionnaire varied (e.g., hometown vs. college friends), thus affecting the results.

Rebuttal: If this argument is correcy, then the parent variable should have the highest correlation, the college friends variable the next highest correlation, followed by the hometown friends. Yet, this is not so.

Question: Initial similarity fostering friendship vs. friendship fostering growing similarity. Which comes first?

What types of similarity increase over time? Literature suggests that facial similarity of married couples increases over time. (As emotion and feelings are shared among couples, the emotions are amplified and then imprinted on the faces, e.g., both laugh more, therefore, both have more wrinkles on their faces.) Studies demsontrating this effect, however, use cross-sectional methods (comparing marriages of different length). It does not rule out that those who stay longer together are the ones that were more similar in the first place (and the ones who were not so similar may last for a while but are more likely to get divorced and therefore do not show up in the sample of long-lasting marriages). Due to changes of society and culture over time, other confound may exist with cross-sectional studies. Only longitudinal studies can account for the increased similarities over time.

Hypothesis of Similarity+ projection: More similar people tend associate with one another, so they will judge the target similarly to how the target judges him- or herself simply because the judges use a projection of their own personality as the basis of judging the other. Therefore, friends (drawn to each other because of their similarity) will on average show greater self-other agreement than strangers. This process may explain the acquaintanceship effect. For example:

In both conditions, judges may use projection. The difference is that in the stranger condition, judges’ projections are on average false (because strangers are not expected to be similar to self), which is a form of false consensus, whereas in the friend condition, judges’ projections are on average accurate (because friends are on average similar to each other). So to test whether this hypothesis is true, Funder examines whether friends are really more similar to each other than strangers are to each other.

In fact, the actual similarities between strangers and between friends is similar – around 0.

This is surprising (other research shows that friends are indeed quite similar to each other). In fact, Funder may be wrong:

• the questionnaire used may not fully sample relevant personality dimensions
• the study compares the average of the correlation without taking into consideration that a given pari of friends may be very similar on some dimensions (those that are important to their friendship) and not similar on other dimensions (those not so important to their friendship. Without analyzing

Furthermore, perhaps people don’t use projection all the time (e.g., not for strangers). They may project only for those domains that they and their friend are similar in (that would justify projection), not on other ones.

Question for thought: If it's not similarity that causes the acquaintanceship effect, what is it?

Important function of studies on agreement/accuracy:

• People can realize that they are not as accurate as they think they are when judging people.
• This may encourage people to respect others’ individuality and to keep learning about them more.
• Other functions?

Swann used better test than Funder: longitudinal design.

The results showed quite good self-other agreement, but Swann does not focus on accuracy. He wants to undermine the belief that with increasing length of relationships, accuracy increases (not only confidence in one’s accuracy).

People often rely on their "intuition" to make personality judgments. That is, they use confidence, not really knowing about their accuracy.

If we are aware that confidence and intuition mislead us in judgment, why then do we still use them in judging others in relationships?

• Perhaps people want to preserve the "added surprise" they can obtain from making wrong judgments.
• or maybe, they just don't care.