Search for: All records

Targeting the right group of workers for crowdsourcing often achieves better quality results. One unique example of targeted crowdsourcing is seeking community-situated workers whose familiarity with the background and the norms of a particular group can help produce better outcome or accuracy. These community-situated crowd workers can be recruited in different ways from generic online crowdsourcing platforms or from online recovery communities. We evaluate three different approaches to recruit generic and community-situated crowd in terms of the time and the cost of recruitment, and the accuracy of task completion. We consider the context of Alcoholics Anonymous (AA), the largest peer support group for recovering alcoholics, and the task of identifying and validating AA meeting information. We discuss the benefits and trade-offs of recruiting paid vs. unpaid community-situated workers and provide implications for future research in the recovery context and relevant domains of HCI, and for the design of crowdsourcing ICT systems. 

Alcoholics Anonymous (AA) is the largest grassroots peer support group for any health condition. While AA meeting attendance is particularly important for people who are newly sober, newcomers often have trouble finding meetings because of a lack of global up-to-date meeting list due to preference for regional autonomy in AA's organizational structure. Detection of regional webpages containing meetings and extraction of day, time, and address of meetings from those pages are essential steps in making the information available and up-to-date in a global meeting list. However, varied structure of the webpages and the meetings pose challenges in achieving the goal with traditional information retrieval methods. In this paper we propose HAIR: a semi-automated human-aided information retrieval technique and explore its potential to solve this problem. We describe future directions in developing this critical tool and discuss major implications of our work in pointing to the importance of context-specific rather than context-agnostic semi-automated in-formation retrieval techniques by conceptualizing the proposed methods and results in a broader context.