Can we describe an individual's personality? agreement of stereotype versus individual attributes

Group members described other members of their group (known for at least a year) by four different procedures (free-descriptive phrases, free-descriptive adjectives, an adjective checklist, and a rating scale) at weekly intervals. As expected, the agreement between pairs of observers, as measured by overlap of content, increased from the free to the structured tasks. However, observers, agreement by chance and on group stereotypes, assessed by counting overlap between their descriptions of two separate persons, also increased. When chance and stereotype-associated aggreement were subtracted from overall agreement, the remaining, "person-specific" agreement averaged about 5%. Agreement measured separately by a matching procedure averaged 14%. No significant increase in person-specific agreement was foud between the free-descriptive and standard tasks. The findings indicate that agreement on individuating features of a person is very low and is not appreciably improved by providing observers with standard descriptive tasks.