Millions of people used social media technologies to organize themselves to produce information and artifacts of value, such as Wikipedia, GNU/Linux, and PatientsLikeMe. However, these communities have significant problems: (1) They work so well to let people maintain existing relationships and find others who are similar to them that they can promote self-segregation and opinion polarization. (2) Knowledge production has been democratized, but at the price of devaluing expert participation. (3) They open people to fresh ideas from far-off places, but few sites incorporate the rich local knowledge of home communities. This project addresses these problems by creating algorithms and user interfaces that nudge individuals and communities into more effective patterns of social connections and interaction.
The intellectual merit includes: (1) Identification of principles people use to build their social networks, the range of potential contacts they consider, and the contexts in which they encounter potential contacts. (2) Empirical knowledge of the effectiveness of different social network structures for solving a range of realistic decision-making tasks. (3) Novel social recommendation algorithms that maximize network effectiveness and that recommend interaction opportunities that serve as contexts for communication and connection.
Broader impacts: This research will be performed in the context of two domains of societal interest, parenting and bicycling, enabling significant broader impacts. Providing people information that gives them confidence to bicycle more increases their personal health, decreases air pollution, and improves community cohesion. Research shows that parents with larger and more diverse social networks make better parenting decisions, and that parents also seek expert advice for specific issues. Therefore, a system that helps people form effective social networks and that provides access to expert educators offers direct benefits to parents. Better parents also tend to be more involved in their communities; thus, improved community health is a bonus effect. Finally, through the research team's connections with parent educators, the project will explicitly target parents from low-income families, who have the greatest needs for support and information.
