This presentation focuses on methodological problems in the analysis of individual differences in face-to-face interaction. Work in this area is complicated by the characteristics of interaction that render the standard psychological approaches to individual differences problematic.

Early work on individual differences in interaction applied the approach of psychological testing to participants’ actions. The rate at which a certain action, such as gazing at the partner or smiling, was observed was taken as an indicator of some transient state or more enduring personal characteristic of the participant. These rates were calculated by counting the number or timing the duration of the action, and dividing by some broader number, such as the total time of the interaction or the total time speaking. It was argued [2] that these simple-rate variables could not be interpreted as representing exclusively the observed participant. Rather, because of the deep sequential interconnectedness of actions in interaction, the partner may be contributing substantially to a simple-rate variable intended to represent the participant.

An alternative approach to the analysis of individual differences is proposed based on the notion of interaction as a convention-based, or rule-governed, phenomenon (e.g. [1]). From this perspective it becomes possible to speak of grammars or structures for observed interactions. Sequential analyses can yield hypotheses concerning the actions, rules, and other elements of an observed interaction. These analyses can be significantly facilitated by Theme [3]. In many cases the rules for an interaction provide for various options, just as do the rules for chess or other game.

An hypothesis of an interaction structure can be represented as a flow chart. All possible paths through the flowchart can be identified. Each path represents one of the possible ways that participants can jointly construct an interaction within the constraints of the structure. Based on these paths or portions of them, it becomes possible to generate data from the observed interactions on a variety of phenomena, such as the number of times each path is taken, the number of times a particular option is exercised, and the number of times action sequences of particular interest occur.

It is proposed that data of the sort just described can be used for analyses of individual differences. That is, individual differences in interaction can be analyzed in terms of the jointly constructed interaction, thus avoiding the problem of analyzing a participant’s action apart from the sequential interaction context in which it occurs.

A variety of techniques involving computational modeling of flowchart phenomena will be illustrated, ranging from simple independent observations, through stationary Markov chains, to dynamic nonlinear models with feedback loops--models of the type whose output may display the type of sensitivity to initial conditions typical of mathematic chaos.

References

1. Duncan, S.D.; Fiske, D.W. (1977). Face-to-Face Interaction: Research, Methods, and Theory. Hillsdale, NJ: Erlbaum.
2. Duncan, S.D.; Kanki, B.G.; Mokros, H.B.; Fiske, D.W. (1984). Pseudounilaterality, simple-rate variables, and other ills to which interaction research is heir. Journal of Personality and Social Psychology, 46, 1335-1348.
3. Magnusson, M.S. (1998). Recursive time patterns in behavior. Submitted for publication.