Emergent processes as means of uncertainty reduction in group decision making
by Marie Ritter
Date of Examination:2022-10-19
Date of issue:2022-11-29
Advisor:Margarete Boos
Referee:Margarete Boos
Referee:Julia Fischer
Persistent Address:
http://dx.doi.org/10.53846/goediss-9589
Files in this item
Name:Rev_Dissertation_MarieRitter.pdf
Size:5.17Mb
Format:PDF
The following license files are associated with this item:
Abstract
English
This dissertation investigates how emergent processes can reduce uncertainty in group decision making.Emergent processes are processes that arise from individual characteristics and social interactions and result in higher level properties of a group. The following three research questions are addressed: Can groups reduce uncertainty through emergent processes? Which emergent processes can be identified? How can we analyze and visualize these emergent processes? An interdisciplinary theoretical framework was developed (Chapter 2), spanning theory and empirical evidence from biology, machine-learning, and psychology. Uncertainty is distinguished into informational uncertainty (i.e., uncertainty about the environment) and personal uncertainty (i.e., uncertainty about relationships and group characteristics). The effects of three emergent processes (collective cognition, group cohesion, and leader-/followership) on uncertainty reduction are investigated in groups. Chapter 3 presents two empirical papers in which groups were faced with informational uncertainty in an intellective problem-solving task within the HoneyComb paradigm (Boos et al., 2019). The HoneyComb paradigm is a virtual game platform in which participants, represented by an avatar on the playing field, can interact only through movement on the playing field while solving group tasks. Specifically, groups had to infer the best out of four options presented to them on the playing field by repeatedly choosing from the options (exploration). Groups in these studies reduced informational uncertainty by inferring the best option through collective cognition (i.e., pooling of individual information on the collective level) and exhibited exploration/exploitation patterns. The quality of information available to the groups impacted the group decision process: Information that was corrupted by group incentives (i.e., rewards for staying close to other group members on the playing field) led to sub-optimal decisions. A simulation study using the ε-greedy algorithm exhibited similar results. Behavioral leader-/followership patterns emerged spontaneously within most groups and were associated with self-reported leader-/followership. However, behavioral leadership was not associated with typical personality trait correlates (self-confidence, achievement maximization, decisiveness, and risk propensity). Behavioral group cohesion was associated with self-reported group entitativity and interactivity. Chapter 3 concludes that groups effectively reduce informational uncertainty using emergent processes (i.e., group cohesion, leader-/followership), although the differential contribution of single processes remains to be investigated. Chapter 4 presents empirical findings on the reduction of personal uncertainty through the emergence of collective trust. Collective trust is defined as a collective cognitive construct that emerges through repeated interactions of a group and reflects the shared level of trust a group holds for another individual, group, or organization. In Chapter 4, groups had to make investment decisions in the Collective Trust Game (CTG), a collective economic game, constituting a judgmental task. Findings suggest that collective trust, as a collective cognitive construct, emerged through interaction and reduced personal uncertainty in the CTG. This was indicated by an increase in consensus decisions and a decrease in the decision times necessary to reach a consensus on investment decisions. Chapter 4 concludes that collective trust, as an emergent cognitive construct, can reduce personal uncertainty within groups. Chapter 5 describes the methodological contributions of this dissertation. The HoneyComb paradigm (Boos et al., 2019) is presented as a tool to investigate emergent group processes using spatio-temporal data. Three levels of analysis are presented and combined with four analytical approaches based on network and visual analytical strategies. An example illustrates how the presented approaches can be used to investigate emergent processes in group decision making under uncertainty. This dissertation concludes: Groups can groups reduce informational and personal uncertainty through emergent processes. These processes include collective cognition, group cohesion, and leader-/followership. The identified processes, and possibly more, can be analyzed and visualized using network and visual analytic strategies as presented in this dissertation. The distinct contributions of each of those processes to uncertainty reduction should be the topic of future research.
Keywords: Emergent processes; Group decision making; Group cohesion; Collective cognition; Leadership; Exploration/Exploitation; Uncertainty