![Sample our Communication Studies journals, sign in here to start your access, 2013 & 2014 volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5927/TOC-Subject-Sample-2021-Information-Science2.jpg"})
Abstract
Functional communication in small groups is generally treated as a medium (i.e., reflecting and explained by individual proclivities) or as constitutive (i.e., interaction has independent effects on outcomes). Rather than approach the problem as either/or, we assume that functional communication consists of both processes. To explore this assumption, we applied two models that offer contrasting views on this debate—Hewes's (1986, 1996, 2009) socio-egocentric model and Gonzalez and Griffin's (2002) latent group model—to four published group interaction datasets to examine the extent to which communication has individual- and group-level characteristics. Analyzed individually, the data exhibited a large proportion of egocentricity but very little group-level connections. An analysis of the combined and recoded data, however, revealed that discussion consists of both egocentric and group speech, but that the patterns differ. We then discuss the theoretical implications for observed patterns at different levels of analysis in the data.
Acknowledgments
The authors thank Charles Pavitt and Laurie Weingart for sharing their data.
Notes
1. And this is true even if both types of processes lead to the same outcomes (Sanders & Bonito, Citation2010).
2. We find the “process loss” argument particularly vexing because, in the absence of telepathy, interaction (including face-to-face and mediated contexts) is the only vehicle through which even intellective tasks (i.e., ones with correct answers) might be solved.
3. These data sets do not, obviously, constitute a random sample of studies in the group literature nor are they meant to. Rather, we identified a small set of studies that would fit with our theoretical and analytical requirements and then contacted the researchers to inquire about using their datasets in this study. At base, the data had to (a) have some type of functional coding scheme that could be adapted to current purposes, (b) be segmented comparably, and (c) track (or have unique identification for) each individual's contribution within each group. In fact, not many data sets, especially ones used in examination of functional communication, meet this last criterion. After exploring several potential data sets, we felt these four satisfied our requirements.
4. One of the tasks, described by Meyers and Brashers, involved medical options for a middle-aged, sedentary accountant with a heart problem. One choice, an operation, was very risky but would allow the patient to maintain his current lifestyle if it was successful. If the patient chose to be less risky and bypass the operation, he would have to drastically change his habits. The other two tasks provided similar sets of closed-ended choices.
5. Kivlighan (Citation2007) applied the LGM to dyadic data with distinguishable partners using structural equation modeling. But in the case of group data with exchangeable members, structural equation models quickly become complicated and unwieldy.
6. Certainly categories from the various data sets did not always fit perfectly within these seven grouping options. To achieve some sense of reliability, both authors discussed the recoding until agreement was reached. For example, “substantiation” as used by Weingart et al. (2007) includes both defending arguments and making factual statements, whereas Hirokawa's (1982b) definition includes attempts to provide proof, evidence, or support. We reasoned that participants use deeper level reasoning when defending arguments, hence the inclusion of Weingart et al.'s notion of substantiation within the evaluation category.