Methodical Snark critical reflections on how we measure and assess civic tech

research links w 37 -17


I’m going to start prioritizing brevity, leaving out some of the absurdity and academic opps, let me know if you miss anything.


How to improve the quality of crowdsourced citizen science data? Technical measures help, but only when accompanied by instructions, according to an empirical study of four cases. Meanwhile, open data on public safety and transportation are the most popular datasets in US cities, according to research from @SunFoundation.

On engagement and impact, a study of Indonesian open data users reveals a laundry list of things government should do to make data #opendata portals more trustworthy and user friendly, while new research from @guygrossman on citizen mobile reporting in Uganda, finding a significant uptake and enthusiasm but no impact. Findings suggest the content of reporting matters.

New research supported by @allvoicescount explores the role of norms in digital citizen engagement. The final report applies structuration theory to eight case studies, and in the course of   13 figures, 24 findings, and 68 dizzying pages, concludes that norms are important. Still working to wrap my head around whether anything is being said here. More concretely, a field experiment in Nigeria documents the importance of social norms to facilitate corruption reporting. Cool design.

Social media is still not being as cool as it should be. For example, tagging fake news on social media doesn’t work, and exacerbates the problem for some groups of readers (for example Trump supporters under the age of 26) according to a recent US study. The number of African countries requesting user info or info takedowns from Facebook rose from 5 to 18 in the last three years, according to a new research report which describes similar trends across the continent.

Community and Resources

This blogpost from @bbcmediaaction argues for the importance of qualitative research to capture the nuance and stories behind global data when measuring progress in development. Meanwhile, @alieholder reminds us why the gender data problem isn’t just about supply and disaggregation, it’s fundamentally also a problem about social institutions and trust. Trust in governance, by the way, is the focus of 23 essays and opinion pieces recently published by  @opengovpart.

DataThief III is a program to extract (reverse engineer) data points from a graph” (yay.), and @bellingcat has put together a list of open source verification and investigation tools and methods.

Most importantly, there’s research showing that only really hot guys win at Tinder. So there’s that.


Add Comment

Methodical Snark critical reflections on how we measure and assess civic tech