Personality differences in spontaneous personality inferences at encoding.

In this article, a method is described for studying personality differences in the spontaneous encoding of complex information about others, and is illustrated in a study of authoritarianism (F). Two sets of sentences were developed in pretesting. In the first set, events suggested a consensual disposition for each sentence actor to either high F (HF) or low F (LF) subjects. In the second set, events suggested a different consensual disposition to HF and LF subjects. Then 77 HF and LF subjects read both sets of sentences for a memory study. Consensual dispositions and semantic associates to the actors were used to cue recall. For the first set, there was a significant interaction between subject type, sentence type, and cue type. LF dispositions were more effective retrieval cues for LF than for HF subjects. Subjects had very little accurate awareness of having made trait inferences. No significant effects were found for HF cues alone or for the second sentence set. Results indicated that HF and LF subjects differ in their spontaneous social inferences about others and have little awareness of making these inferences. Implications for integrating trait and cognitive approaches to personality are discussed. The interactionist approach to personality emerging from recent debates over personality traits (Magnusson & Endler, 1977) has taken several forms. Most research has been at a macrolevel, and demonstrates that the expression of traits is conditional upon situational variables, producing statistical interactions between traits and situations. Crowne and Marlowe's (1964) work on the need for social approval, and Christie and Geis' (1970) work on Machiavellianism provide numerous examples. More recent macrolevel research indicates that traits predict the individual's active selection and even creation of different situations (dynamic interactions). Snyder's (1979) work on selfmonitoring, Swann's (1983) research on self-esteem and selfverification, and research on self-consciousness (Turner, Scheier, Carver, & Ickes, 1978) describe some of the macrolevel processes involved. At this macrolevel, the smallest unit of analysis is the individual's response or action. More microlevel analyses break the production of behavior itself into multistage processes (e.g.. attention, encoding, storage, retrieval), and examine effects of individual differences at particular stages. These may be measured with personality inventories more typically used to measure traits at the macrolevel, such as anxiety (Sarason, 1980), locus of control (Lef