A valid procedure for obtaining self-report of affect in marital interaction.

Thirty married couples interacted in a low-conflict situation and a high-conflict situation during which continuous physiological measures were obtained. Each spouse returned separately for a second session in which they watched the videotape of the interaction and provided a continuous self-report rating of their own affect while the same physiological measures were again obtained. Observers coded the spouses' affect during each speech unit. The self-reports of affect (a) discriminated the high-conflict interaction from the low-conflict interaction, (b) correlated significantly with marital satisfaction, (c) were coherent between husband and wife, and (d) were significantly related to the observers' coding of the couples' affect. Physiological data obtained during the interaction session were significantly related (using time-series analyses) to physiological data obtained during the recall session. In all comprehensive discussions of emotion the subject's own experience of affect has been considered an important channel of information (see Strongman, 1978). This channel has always been troublesome from an empirical standpoint in terms of meeting psychometric criteria of reliability and validity. One important application of a valid selfreport procedure would be in the study of emotion during social interaction. Ekman, Friesen, and Ellsworth (1972) noted that the study of emotion in the context of social interaction can make a number of unique contributions. One of these contributions is the study of interaction as it unfolds in time, which permits the analysis of the sequential nature of the interaction using time-series analysis. Thus, a method for procuring the self-report of affect that could provide a continuous record over an interaction session would be extremely useful in the study of emotion in social interaction.