Crowdsourcing User-Centered Teams

Abstract

Crowdsourcing has become an increasingly important tool for team formation and collaboration. This thesis investigates how User-Centered Design, an iterative process that prioritizes users and their needs, can be applied to improve the efficiency and effectiveness of crowdsourcing systems for teamwork and team formation. To achieve this, we conducted a series of studies to explore the role of various factors in shaping crowd workers' behaviour and preferences in collaborative contexts. The main findings of our research are as follows. In online team formation settings, crowd workers prefer disclosing overt traits (e.g., age, gender, topical interests) and avoid sharing sensitive information (e.g., ethnicity, depression). However, they are willing to share information regarding their personality and values, typically considered deep-level sensitive traits. Well-defined digital nudging interventions, such as a diversity progress bar, can promote diverse team formation. In contrast, subtler forms of nudging may inadvertently trigger biases working against the intended objectives. Ad-hoc crowd teams working under pressure can benefit from systems that account for differences in personality traits, as these can influence collaboration outcomes and perceptions. Designing crowdsourcing systems for emergency response requires modelling communication tools that aid, assist, and monitor the shared load, considering the strictly cooperative roles and task- and user-dependent communication styles between collaborators. When forming teams, crowd workers tend to balance attributes between and within groups, with a preference for Openness to Experience among the Big-5 personality traits. Based on these findings, we recommend applying a User-Centered approach to design collaborative crowdsourcing systems, considering user needs, behaviour, intents, and perceptions of digital environments. Future research should continue to explore and evaluate innovative strategies for promoting effective collaboration and team formation in crowdsourcing contexts.