Dissertation Title: Designing for a new "ZIP code destiny"

Abstract:

We live immersed in “cocoons”: tight-knit, segregated psychosocial units that shape who we encounter, the media we consume, what we believe, and ultimately, the opportunities we are able to access in order to positively shape ourselves, families, and communities.  These cocoons are fractal in nature, manifesting geographically as schools and neighborhoods segregated by race and income; as social networks that shape which media, mentors and role models we are exposed to; and even in our minds as cloistered concepts that spur biases and make us less welcoming of difference.  They are one of the reasons that America is the land of “ZIP code destinies”: the geographic and social contexts in which a child grows up often dramatically affect the opportunities they are able to capitalize on.

Advances in social media and communications platforms were supposed to create new connective tissues between cocoons to enable a freer flow of knowledge and opportunity between disparate groups.  In some ways, this has happened, and many of these advances have also spun cocoons where marginalized groups can build solidarity and offer mutual support.  Yet these advances have also produced social media ecosystems that are highly fragmented, amplifying a priori preferences for which information to consume, and from whom.  They have also enabled those with various privileges to more easily access and act on information to obtain education, healthcare, jobs, and other critical resources.

This dissertation explores how the analysis of data from digitally and physically-mediated social environments might help inform the design of new technologies to mitigate cocoons across two domains: politics and education.  We start with an analysis of political fragmentation on Twitter in the wake of the 2016 US Presidential Election.  This analysis motivates the design of a web application, Social Mirror, to probe how prompting social media users to reflect on their own “echo chambers” might help mitigate such fragmentation—which has become a crippling feature of US society, often impeding policies that could positively shape children’s futures.  

Social fragmentation, of course, is not only rampant in our social media ecosystems: the neighborhoods in which many children grow up are fragmented by race and income, creating cocoons that impede access to quality role models, schools, and other educational opportunities.  First, we investigate existing neighborhood-level datasets detailing the importance of exposure to role models to inform the design of INSPIRE, a new video-based social network for middle schoolers to enhance exposure to role models as they start to think about their future aspirations.  Next, given the importance of schools and the role parents play in school choice—turning both to personal networks and online resources to inform their choices—we use recent advances in natural language processing to analyze parents’ reviews of schools posted online.  Our analyses, however, reveal that affluent parents are more likely to post reviews, and that reviews recapitulate well-documented racial and income disparities in education.  We use these insights to inform the design of EdMirror, a “community-sourcing” platform that seeks to surface less biased, more actionable insights from Boston Public Schools parents to other parents and school leaders in ways that might help spark sustainable, positive changes in schools.

A central theme across these efforts is the role of prompted reflection and introspection as a potential mechanism for mitigating the biases and other psychological barriers that perpetuate cocoons.  The dissertation concludes by exploring how the analysis and design of communications platforms can inform new tools for reflection—i.e., “mirrors”—and how these mirrors might combine with other structural interventions (like policy change) to fuel designs for a new ZIP code destiny.