Conversational Argumentation in Decision Making: Chinese and U.S. Participants in Face-to-Face and Instant-Messaging Interactions

Abstract

This study investigates cultural and communication medium effects on conversational argumentation in a decision-making context. Chinese and U.S. participants worked in pairs on two decision-making tasks via face-to-face (FtF) and instant messaging (IM). The analyses showed that Chinese participants tended to engage in potentially more complex argumentation, whereas U.S. participants tended to utilize proportionally more statements of claims and statements of convergence (agreements, acknowledgments, and concessions). Argumentation in IM tended to be more direct than in FtF interactions. There were no interaction effects between culture and communication medium on argumentation behavior. In addition, statements of convergence were found to be negatively related to measures of persuasion, indicating that such statements do not necessarily indicate true agreements or shifts in opinion. The results are discussed in terms of structuration theory and the socioegocentric model of communication.

Notes

¹One half of the transcripts were divided into thought units by Leslie D. Setlock, the other by one of the trained coders. Both used the following guidelines to parse the transcripts, based on Hatfield and Weider-Hatfield (1978): (a) Look for beginnings and ends of statements or ideas. If the syntax suggests that there is a single idea being expressed—for example, together the turns or lines express a single clause, or an independent clause and then a closely related dependent clause, then these should be combined as a single thought unit; (b) look for clear breaks in idea—things that indicate a different idea is being expressed. This might be a shift in topic, or indication of a shift from making a claim to giving a reason (e.g., because, since, etc.), or other indications of a shift (e.g., but, however, etc.).

²The data for one pair was recovered after all of the other transcripts had been coded. Thus, both the face-to-face and instant-messaging transcripts for this pair were coded by only one of the trained coders.

*p < .05.

**p < .01.

***p < .001.

³These correlations are with individual participants as the unit of analysis. With dyads as the unit of analysis, the correlations are essentially the same:r = .339 (p < .07) for agreement and r = −.581 (p < .01) for change (N = 30).

*p < .05.

**p < .01.

*p < .05.

**p < .01.