Current Projects

Postcolonial Computing

I am investigating new frameworks for thinking about issues in cross-cultural issues in design methods, drawing from postcolonial theory, feminist studies, and science technology studies. While the fields of HCI and STS have recognized the way artifacts and technologies are made meaningful and useful in relation to local cultures, knowledges, and infrastructures, there has been less recognition of how design methods as a form of knowledge production and cultural performance are similarly entangled in local contexts. HCI methods in practice almost always exist in the context of the colonial histories, unequal economic relations, and unequal power relations. Many ICT4D studies have noted ways in which design methods fail in contexts very different from those of their origins.

I will be undertaking fieldwork at a design firm in India that works both with local clients and with multinational corporations. Through participant observation and interviews, I seek to answer questions that include:

• What do design methods do in practice?
• What tacit assumptions do HCI and UCD design methods make epistemologically, infrastructurally, and culturally?
• What do concepts such as user, designer, and innovation do in design practice?

See workshop position paper on Portability of Design Methods, accepted at CHI 2008 Workshop on ICT4D or Postcolonial Interculturality from IWIC 2009.

I undertake this project with Paul Dourish (Informatics) and Kavita Philip (Women's Studies).

Turkopticon: Political Human-Machine Configurations

Amazon Mechanical Turk (AMT) is a meeting place for requesters who want large volumes of cognitive piece work done and people who want to do those tasks. People often make $1-2 dollars an hour and most are from the US. Who are Turk workers? While some Turk for fun, others Turk because in tough economic times, it helps make ends meet even though the pay is very low. We (unscientifically) surveyed workers, asking them to draft a what they would like as a worker's bill of rights. A concrete and frequently cited platform inequity was that workers could be negatively reviewed, but employers went unreviewed. Employers who scammed, stole work, or did not respond to questions were unaccountable. Based on these learnings, Six and I built Turkopticon, a Firefox Extention that allows Mechanical Turk workers to review requesters and see reviews in context as they browse for AMT tasks. Turkopticon has been installed by 900 Turk workers and is currently operating, with reviews on almost 300 requesters.

• Check out the Turkopticon site
• Slideshare: "Agency and Exploitation in Mechanical Turk" (talk given Nov. 13, 2009 at Internet as Playground and Factory at The New School, NY)
• Ross, Zaldivar, Irani, Tomlinson (2009). "Who are the Turkers? Worker Demographics in Amazon Mechanical Turk." Technical Report SocialCode-2009-01.
• Essay on my concerns about crowdsourcing

Through this work, I question how we might talk about the politics of technologies that have diverse meanings in diverse cultural contexts. Turkopticon is also a study in designing as a relationship, rather than a process leading to an artifact.

I undertake this project with Six Silberman, Joel Ross, and Andrew Zaldivar. Dolores Labs also supported us in these efforts and gave us some press. Beatriz da Costa (Arts, Computation, and Engineering) advised us.

Collective privacy strategies in and around virtual worlds

As web applications collect increasing amounts of people's personal and collaborative content, we need to understand people's online collective information practices, which include issues of privacy, trust, intimacy, and sharing. I am beginning a study of online activist groups that use a combination of virtual environments, social networking services, email, and blogs to stay connected. By studying individuals who have a heightened need for intimacy, coordination, and information sharing, I hope to render visible tensions, needs, and strategies that may also exist more subtly in broader populations. Work from this project has been published at CSCW: Situated Practices of Looking: Visual Practice in an Online World. I am collaborating with Gillian Hayes and Paul Dourish.

Past research

Investigation attrition in undergraduate computer science

I spent a year and a half designing and conducting a longintudinal mixed-methods study of students in Stanford's introductory computer science sequence. By following a cohort of 30 students with quarterly interviews, we found aspects of course experience that affected confidence, motivation, and interest in the major. Through surveys of a broader population and analysis of course performance by gender, we set qualitative findings against measures of confidence and motivation in the broader course population. The research was awarded Stanford's Firestone Medal for Undergraduate Research (top 10% of theses).

• Thesis - A Different Voice: Women Exploring Stanford Computer Science (pdf)
• ACM SIGCSE: Understanding gender and confidence in CS course culture

Past design work

This is only a sample of projects I worked on while at Google. I am only providing publicly available screenshots here.