Selz, Karen A.

The principles of action identification theory are used to form a foundation for a dynamic approach to social interaction that may capture both stability and change in social judgments. Attributions of blame in an ambiguous rape case were explored through the presentation of transcribed interviews with either an alleged rapist (Larry) or his alleged victim (Jane), under either an induced high or low level of action identification, followed by one of two courtroom "closing arguments," attributing primary responsibility for the incident to either Larry or Jane. Responsibility judgments and personality trait ratings were obtained from each subject on both Larry and Jane. Action identification questionnaires for the Larry and Jane perspectives were administered to each participant. As predicted, compared to the high level subjects, subjects in the low level identity condition were more malleable in their judgments of either target. Implications and future research directions are discussed.

Human subjects perform simple, relatively unconstrained, game-like computer tasks. "Meso-level" measures of behavioral complexity and time-dependencies (including entropies, grammatical complexity estimates and run statistics) are derived and computed. Individual behavioral differences in the resulting complexity measures are robust and, in a temporal-forcing paradigm, are statistically significantly related to the same individual's scores on a range of personality and demographic variables. Through an experimental manipulation and the statistical selection of maximally useful predictor sets personality and demographic variables are united in a "macro-level" temperament typology, based on "micro-level" behavioral tendencies. Further, I can compute a parameter value of a one dimensional dynamical system, the symmetric tent map, matched to the symbol sequence "meso-level" parities of the subject. When this parameter is used in the iterated map, it produces sequences that are of the same autocorrelation "category" and share much of the fine structure of the autocorrelograms of the subjects to which the map parameter had been matched.