Density, Transitivity, and Diffuse Status in Task-Oriented Groups

• Published in: Social psychology quarterly 1995-12, Vol.58 (4), p.241-254
• Main Authors: Robinson, Dawn T., Balkwell, James W.
• Format: Article
• Language: English
• Subjects: Aggregation; Behavior; Biological and medical sciences; Characteristics; Cooperative learning; Cues; Data; Density; Discussion groups; Dyadic relations; Fundamental and applied biological sciences. Psychology; Gender; Gender differences; Grammatical gender; Group Dynamics; Group Structure; Heterogeneous Grouping; Homogeneous Grouping; Individual Characteristics; Laboratory Experiments; Matrix; Modeling; Monte Carlo methods; Participation; Psychology. Psychoanalysis. Psychiatry; Psychology. Psychophysiology; R&D; Race; Racial Factors; Relational status; Research & development; Research methodology; Researchers; Simulation; Simulations; Small Groups; Social Change; Social interaction; Social interactions. Communication. Group processes; Social psychology; Social Status; Social structures; Studies; Task Oriented Groups; Theory; Transitivity; Triads; Valence; Working groups
• ISSN: 0190-2725; 1939-8999
• DOI: 10.2307/2787126

Small-group researchers recently have been concerned with how diffuse status characteristics such as race and gender combine with verbal and nonverbal cues to structure relations in task groups. In this paper we investigate the structure of precedence relations in such groups, including whether these relations are dense, whether they are transitive, and whether they are structured in any obvious way by diffuse status characteristics. Using transcript data on six-person mixed-gender discussion groups, we found relational density in about 55 percent of the dyads, transitivity in about 75 percent of the completed triads, and no clear pattern by gender. Each of these findings is contrary to assumptions contained in a formal participation model by Fisek, Berger, and Norman for which these authors reported strong empirical support. To investigate this apparent contradiction, we designed and carried out Monte Carlo simulations, creating simulated data with a known underlying structure, and applying Fisek et al.'s research methodology to our simulated evidence. Although the simulated data, by design, systematically violate Fisek et al.'s density and diffuse status assumptions, their model actually fits these data better than it fits the empirical Smith-Lovin data. We discuss both theoretical and methodological implications of our empirical and Monte Carlo results.