Group Problem-Solving Processes: Social Interactions andIndividual Actions

Some groups work together successfully to solve problems while others do not. In particular, researchers have shown that group work can induce many beneficial outcomes in comparison to traditional forms of instruction (increased learning, decreased racial tension, more positive student attitudes toward school, etc.; see reviews by Good, Mulryan & McCaslin, 1992; Slavin, 1990). However, these results are not universal. Newman and Thompson (1987) showed that only 68% of 37 secondary school studies favored group work over other types of instruction. This variability suggests that some conditions prevent students from enjoying these theoretical benefits. A closer examination of the group’s interactions can help explain these disparities by identifying beneficial and harmful consequences of specific actions and sequences of actions. Armed with this knowledge, group members can change their behavior accordingly. Past discourse analysis research on social interactions has typically focused on either specific situations (ethnography of communication) or the role of specific interaction elements across different situations (interactive sociolinguistics, ethnomethodology, pragmatics, variation analyses, and speech act theory; see Schiffrin [1994] for an extensive comparsion of all six areas). Analyses of specific situations have shown how interactions reflect social and cultural beliefs (e.g., Goffman’s [1977] study of courtesy indicating expected gender roles and Gumperz [1983] analyses of inter-ethnic conversations). Meanwhile, analyses of specific elements identify their functions and application conditions (e.g., Sacks’s [1979] study of the word “hotrodder” and Jefferson’s [1986] study of overlaps in speaker’s turn-taking). As a result, these analyses have generated lists of important aspects of an interaction to consider (e.g. Hymes’ [1974] SPEAKING framework [Setting, Participants, Ends, Act sequence, Key, Instrumentalities, Norms, Genres] ). However, no one has introduced a systematic model suitable for statistical analyses of many interactions both in their entirety and in fine detail. Furthermore, group