Using sequential collaboration to aggregate judgments into accurate estimates
Workgroup | Knowledge Construction Lab |
Duration | 08/2023-open |
Funding | IWM budget resources |
Project description
With the spread of the Internet and the emerging of collaborative online projects such as Wikipedia and OpenStreetMap, collaboration in groups also changed radically. Instead of in-person groups sharing information and making decisions together, contributors in sequential collaboration form a sequential chain in which the first contributor creates an entry independently which can be edited and improved or maintained by following contributors. This project focuses whether contributors generate accurate estimates in sequential collaboration and which conditions foster or hinder this process.
First studies examining sequential collaboration show that numeric and geographic judgments indeed become increasingly more accurate over the course of a sequential chain consisting of four to six contributors. Moreover, the sequential estimate at the end of a sequential chain can even outperform independent judgments aggregated with an unweighted average which are already found to be highly accurate. Improved judgments in sequential collaboration can be attributed to contributors adjusting, and thereby improving, or maintain judgments according to their own expertise. There results demonstrate that sequential collaboration is a successful method for aggregation individual judgments while considering contributors expertise.
In the following subprojects, we will investigate the role of opting out of providing a judgment in sequential collaboration and how providing additional information about other contributors or the sequential process influences judgments. Thereby, we focus on how to modify the sequential process to foster accurate judgments and reduce worsening of already highly accurate judgments.
Cooperations
Prof. Dr. Daniel W. Heck, University of Marburg