Key points are not available for this paper at this time.
The intraindividual variability of self-ratings of 10 affects was studied in 69 older residents of a long-term care facility. Each person rated his or her affective state over 30 days. P-technique factor analysis was applied to each subject's ratings. In general, the expected two-factor structure showing one positive and one negative factor emerged from both traditional R-analysis and from "chain P-analysis" (all subjects' ratings pooled after standardizing within subjects). Because many subjects gave invariant ratings on one or more affect terms, only 28 P-analyses could be performed, of which 12 yielded one-factor solutions, 16 yielded two-factor solutions (10 the expected two factors), and 5 failed to converge. Possible reasons for the only partial intraindividual replication of the interindividual factors are discussed.
Kleban et al. (Fri,) studied this question.