Group Information Sharing: True Advocacy and the Importance and Distribution of Information

This dissertation examined the structure of information sharing at the individual and group levels using a hidden profile methodology. The study was designed to examine the relative merits of shared information sampling explanations of information sharing (Stasser & Titus, 1985; 1987; 2003), importance-based preference-consistency explanations of information sharing (e.g., Mojzisch & Schulz-Hardt, 2006; Wittenbaum et al., 2004), and true advocacy explanations of information sharing (Stasser & Titus, 1985; 1987). This was done by examining the relative predictive power of information with a shared distribution, information that supports a participant’s preferred decision, and the perceived importance of information in a laboratory experiment in which participants discussed a national security scenario involving weapons of mass destruction. To determine whether separate processes occur at different levels of analysis, separate analyses were conducted at the individual level (3432 observations of 264 participants) and at two different conceptions of the group level: low observation collapse (2200 observations of 88 three-person groups) and high observation collapse (172 observations of 88 three-person groups). At all levels of analysis, support for true advocacy explanations was found. In other words, participants primarily mentioned information that supported his or her preferred decision. Perceived information importance was only an important predictor of mentioning information if information supported his or her preferred decision. Only at the group level was a shared information distribution an important predictor of mentioning information. At the individual level, the ability of shared information to predict mentioning was caused by the tendency of shared information to support the participant’s preferred decision. Results suggest that the field's focus on shared versus unshared information distributions may have been misplaced. The results also suggest that decreasing advocacy for a preferred decision may play a larger role in decision quality than increasing the amount of unshared information that is mentioned.