The engine room has been collecting and summarizing research on ‘what works’ in projects that use technology to solve complex problems. We’re sharing research that we find useful on our blog, in the hope that a wider range of people can use it to inform their work.

Jonathan Mellon, Tiago Peixoto and Fredrik Sjoburg at the World Bank’s Digital Engagement Evaluation Team (DEET) have assessed four online civic engagement initiatives in Brazil, the UK, Uganda and the US, finding that citizens who engaged online were “systematically more privileged than the population or offline participants.”

However, the authors also found that these initiatives didn’t necessarily produce demands that reflected these inequalities. Neither did the outcomes uniformly reflect the priorities of these more privileged groups. In other words, the fact that these initiatives engaged more privileged citizens might not be a problem in itself – the structure of an initiative and the government’s response to it could play much more significant roles. The authors suggest that this means “a new methodological approach that looks beyond the profile of users and also to the institutional design and the government’s response” is needed.

The presentation is here (slides marked ‘preliminary’ represent initial analysis where more work will be done later). Here’s a quick summary of each of the case studies:

• Participatory budgeting in Rio Grande do Sul (Brazil): “Online voters are from traditionally privileged groups” but there was “no significant difference in voting behaviour between online and offline communities”. The authors suggested that this might be because the participatory process vets any proposal before it is issued.  This was the same when it came to results: government implementation was designed to systematically favour poorer areas.
• Fix My Street (a platform for reporting street problems to local authorities in the UK): Preliminary findings suggest that “requests for help are coming from more privileged areas”, but that the demands that come from those areas are “less unequal than the participants’ profiles”. However, they also suggested that “government response replicates unequal demands.”
• U-Report (an SMS platform run by UNICEF in Uganda): The typical participant profile is “highly unequal”, but preliminary findings suggest that the effect this has on responses varies depending on the question being asked. The authors are still assessing outcomes related to U-Report.
• Change.org (a global online petition platform): “Women are systematically underrepresented in petition creation across nearly all countries,” though preliminary findings indicated that men and women tend to sign petitions in roughly equal numbers. Government response “seems to generally follow the categories that get most signatures.”