Crowdwork, crisis and convergence: How the connected crowd organizes information during mass disruption events

Social media have experienced widespread adoption in recent years. Though designed and appropriated for a range of purposes, users are consistently turning to these platforms during times of crisis and mass disruption —a term used here to characterize events, including mass emergencies, natural disasters and political protests, that cause significant disruption to normal routines. Social media are playing host to new, digital forms of the social convergence behavior long known to occur in the wake of crisis events. This activity, which includes participation from local citizens, emergency responders, and global onlookers alike, produces huge volumes of data, some with potential value to affected people and responders. It also creates new challenges. Noise, misinformation, lost context and the unstructured nature of social media updates all contribute to an emerging information processing problem, with information seekers forced to "drink from the firehose" to identify the data they need. Noting the difficulties of completely solving this problem with purely computational solutions, I address the challenge of processing social media updates into usable information from a perspective that positions the participating crowd as an asset in the effort. At the center of this inquiry is the discovery of an emerging role for remote participants during mass disruption events—that of the digital volunteer. This dissertation consists of four separate studies of digital volunteerism and other forms of remote participation, examining several ways members of the remote crowd help to organize information during mass disruption events. Across the different studies, I employ a mixture of methods, including qualitative and quantitative analysis of large volumes of Twitter data, interviews with digital volunteers, and participant observation within a virtual volunteer organization. Integrating the findings from these separate studies, I introduce a new term, crowdwork, to describe the productive activity of remote participants during mass disruption events. Throughout, this dissertation works to unpack the popular crowdsourcing term, by identifying salient features of crowdwork in this context and comparing those with current understandings of crowdsourcing. Examining the larger ecosystem of digital volunteerism during mass disruption events, I describe crowdwork in this context as a multilevel filtration system, explaining how information is processed through a variety of different activities at different layers within a complex information space that includes crowdworkers, virtual organizations, and social media sites that host both the information and the information processing. This model identifies several potential "sites" of innovation where computational algorithms could both support and leverage crowdwork. Finally, from another perspective, I examine crowdwork through the movement and transformation of information. Using the theory of distributed cognition in combination with this information-centered approach, this dissertation concludes with a holistic view of crowdwork on social media platforms as collective intelligence manifested within a global cognitive system