Icarus enters the cage, Bolo grunts, Icarus approaches, Bolo sex presents, Icarus mounts, Icarus then presents for groom, and Bolo grooms. How can we summarize this interaction between two adult male rhesus macaques, and ones similar to it? More importantly, how might we compare interactions recorded in several groups whose members differ in age/sex composition, personality, or some other characteristic? We might count the numbers of discrete behaviors (e.g., approaches, sex presents) recorded, and compare groups using analysis of variance. Or, we might employ a scheme that involves ratings of the interactions, the individuals, or the session using descriptive terms such as ''tense,'' ''affiliative'' or ''cohesive'' [Capitanio, Bond & Mason, 1997]. Finally, if the data were generated using an experimental design aimed at formally studying interaction, we might use a more static approach to social interaction analysis, focusing, for example, on the relative contributions of actors versus partners to the variance in behavior [Kenny, 1996].
In Observing Interaction, Bakeman and Gottman suggest a dynamic approach. Social interaction can be thought of as a stream of behavior that unfolds in time. The stream is patterned by the particular sequence in which the behavioral elements are performed by the individuals. Analysis of the sequences themselves can provide us with information about the organization of behavior during social interaction that can be very revealing about the quality of the relationship that exists between the interactants. For example, two groups of monkeys might display the same number of approaches and sex presents. In one group, however, approach is more commonly followed by a sex present, while in another group, a sex present often precedes approach, but the approach is more commonly followed by withdraw. This finding could suggest important differences between the two groups. Sequential analyses have been more common in the human psychological literature (most of the examples in this volume derive from the longstanding human-oriented research programs of Bakeman, Gottman, and colleagues), but such analyses have been performed in primatology by Altmann [1965] and Sackett, whose edited volume in 1978 (which contained chapters by Bakeman and Gottman) described lag sequential methods for social interaction analysis and upon whose work the present authors build.
The volume is described in the Preface to the first edition as a primer that aims to 'provide a clear and straightforward introduction to scientific methods of observing social behavior' (p. xi). In the Preface to the second edition, the authors indicate that the book has been updated principally to reflect a new approach to sequential analysis: the development of log-linear techniques, covered in a fair amount of detail in this edition, provides those interested in sequential analysis with a broader statistical framework within which to work. Several of the chapters in the second edition are extensively reworked based upon this change.
At a basic level, there is some overlap in topics between this volume and other introductory guides to behavioral measurement (e.g., Martin and Bateson