K&F agree that most theorists reject a “PURE TRAIT” model, which states that there are personality influences that result in invariant behavior across situations. Even such trait theorists as Allport reject this notion. Invariant behavior means psychopathology. WHAT DO SUMMARIES OF RESEARCH STUDIES TELL US ABOUT PERSONALITY? This is expressed in a number of hypotheses that research has evaluated.

1. Is personality an illusion in the eye of the beholder only (Solipsism--that the self can know nothing but its own experiences and states. Vs. Consensus). Does our limited ability to understand all the information in the world force us to use a schema (such as personality) to organize our understanding--when continuity may not really exist?

Answer: No, when observers are well acquainted with the person they are evaluating (e.g., B&A), they show significant levels of agreement.

2. Is this agreement due to Semantic Generalization (Consensus without Discrimination). Are people making evaluations of others using stereotypes without discriminating the real person? Is it possible that anyone judged as “friendly”; is also judged as “empathic”, “altruistic”, and “sincere” because of Semantic Generalization (Stereotyping).

Answer: No, raters can agree on which traits apply most strongly to which person, showing discrimination across individuals.

3. Is agreement due to base rate accuracy (also Consensus without Discrimination). This is also known as the “Barnum” effect. Is agreement due to the use of descriptors that apply to most people?

Answer: No, Although it is well known that the Barnum effect exists, raters can still discriminate on which characteristics pertain to which individuals.

4. Is agreement due to shared Stereotypes that are based on Obvious (but Erroneous) cues (Discriminative Consensus without Behavioral Referents). This seems to be related to the Barnum effect, which is based on the idea that everyone is alike; but this hypothesis is quite different. #4 requires that an observer at least look at a target person, but then that observer assigns the person to a general category that is stereotypical.

Answer, No. Data has shown that inter-rater agreement improves as observers view the target person over time. A stereotype hypothesis would say that rating would stay the same because they are based on stereotypes, not real people. Also, this hypothesis does not account for correlations between judgments (e.g., friendliness) and independent behavioral measures that are objective (e.g., time to initiate a conversation with a stranger).

5. Is agreement due to discussion between observers and with targets (also Discriminative Consensus without Behavioral Referents). Is it possible that observers are influenced by target’s self reports, which are inconsistent with the target’s behaviors.

Answer: No, parent and peer ratings of raters who have never met do show consensus. Also, as observers view inconsistencies between targets’ self-reports and their behaviors, raters are more strongly influenced by the behaviors.

6. Is Agreement due to seeing others in the same setting (or a limited range of settings) and mistaking situational effects for trait effects (Differential Agreement about behaviors without internal traits as the source)?

Answer: Probably not. Peers (college roommates) and parents agree on trait ratings even though targets are seen by each source in different settings.

7. Compared with situational pressures, cross-situational consistencies in behavior are too weak to be important( also Differential Agreement about behaviors without internal traits as the source)?

Answer: K&F argue that personality coefficients of .3 are not too small to be meaningful. (Your teacher happens to disagree with this conclusion, and others share his opinion. In truth, different others share the opinion of K&F, although most everyone would hope that the personality coefficient would be higher.)

K&F argue that TRAIT EFFECTS are limited for a number of reasons.

1. Traits influence behavior only in RELEVANT SITUATIONS.

2. A person’s traits can CHANGE A SITUATION.

3. People with different traits will choose DIFFERENT SITUATIONS.

4. Traits can change with chronic exposure to SITUATIONS.

5. Traits are more easily expressed in SOME SITUATIONS THAN OTHERS.

PLEASE NOTE THAT EACH OF THE 5 POINTS LISTED ABOVE INVOLVE THE POSSIBILITY OF TRAIT/SITUATION INTERACTIONS!!!!

K&F also argue that predictive validity from trait ratings come only when:

1. Raters are very familiar with the person.

2. Multiple behavioral observations are used (i.e., aggregation).

3. Multiple observers form the basis of consensus.

4. Trait dimensions are publicly observable.

5. Behaviors are relevant to the trait dimension in question.

DO NOT EXPECT ACCURACY IN PREDICTING BEHAVIORS FROM TRAIT RATINGS:

1. When predicting behavior in “Strong” situations.

2. When predicting a single behavioral instance from another behavioral instance.