Skip to main content

University of Michigan School of Information


Studying how crowds collaborate in times of shock

Thursday, 07/28/2016

Large-scale collaborative crowdsourcing is a recent phenomenon, allowing millions of people to participate in the same project, whether it be contributing information to Wikipedia or helping to develop software on GitHub. Understanding crowds, therefore, is extremely important, yet research in this area has largely focused only on internal crowd dynamics during times of normalcy.

UMSI assistant professor Daniel Romero and his fellow investigators, UMSI assistant professors Ceren Budak and Lionel Robert, propose to extend the present body of research to examine crowd dynamics during times of irregularity, or shock.

The National Science Foundation has awarded Romero a $499,463 grant over three years for his project “Surviving the Storm: Large-Scale Examination of the Impact of Shocks on Crowd Attributes and Performance.”

The project involves analyzing data sets, primarily from Wikipedia, as well as GitHub, to determine how crowds react to specific types of shock such as losing an important contributor, a celebrity death, or environmental disruptions.

Their research will follow three major directions:  Characterizing the changes in crowd attributes during times of shocks; identifying the changes in crowd attributes during shocks that optimize the future performance of the crowd, and the attributes that make crowds most resilient to crowds; and identifying properties of a crowd that can predict crowd-dependent shocks.

The results of this study should provide insights into collaborative crowd behavior, which can help crowds optimize future performance and facilitate self-organization. The study will also provide a set of basic recommendations for desirable attributes that enable crowds to be more efficient and resilient to environmental changes. This will be essential for the growing trend of crowds generating solutions to increasingly complex objectives.

- UMSI News Service