We are very excited to announce a new project using the Future of Families data: The Fragile Families Challenge.
The Challenge is a scientific mass collaboration that combines predictive modeling, causal inference, and in-depth interviews in order to learn more about the lives of disadvantaged children. By having a community of social and data scientists working on this project together, we can accomplish things that none of us could do individually.
The Challenge follows a research model commonly used in machine learning called the “common task method.” Roughly we will release some of the year 15 data for six key outcomes, and then participants will use this data and the birth to year 9 data in order to build models that infer the six key outcomes for the rest of the year 15 data. Eventually we will combine all of these models into a single “community model” that we can all then use for our research. Your can learn more about the Challenge, which is supported by the Russell Sage Foundation's initiative on Computational Social Science, at the Challenge website.
Participants who make important contributions will be invited---but not required---to publish their work in scientific journals. Also, participants who make the most important contributions will be invited to Princeton for a workshop where people present their results and begin new collaborations.
Ways that you can get involved:
Participate: Use what you know about children and the FFCWS data to build models for the six key outcomes at age 15.
Encourage others to participate: If you have a research group or know others in your network who might be interested, encourage them to apply to participate!
Assign this in your class: Several statistics, machine learning, and social science classes are already planning to incorporate this as an assignment this semester. We think this could be a great way to get students working on a real scientific problem, and perhaps even give them a chance to publish their results.
If you have any questions, please e-mail [email protected].