Five minutes with… Dr Lasana Harris

The Nuffield Foundation, in collaboration with the British Academy, recently awarded six research teams £1.1 million for research into the ways communities can improve social wellbeing in the UK. Professor of Social Neuroscience Lasana Harris (UCL) is leading one of those teams. Ella Rhodes asked him about his research using administrative data to understand community wellbeing.

What inspired this project?

A personal frustration of mine is that social variables – things about people and their interactions with others – are not treated in the same manner as other forms of human behaviour. In economics, there exist indicators of economic stability and well-being; why can't we have an agreed upon set of indices for social and community well-being? Instead, we rely on surveys and focus groups primarily for insight about social functioning and community cohesion.

As a behavioural scientist, I have spent my career turning abstract social concepts into behaviour that we can measure. Behaviour provides additional insight because it is often independent of conscious, self-reports common in all the approaches to understanding community well-being. Such behavioural data currently sits within local government agencies. Couldn't we harness data science approaches, along with operationalised community well-being concepts, to gain insight into community well-being? Certainly, corporations have been using behavioural data for years… why not local government?

What will the project involve?

The project involves firstly synthesising the existing community well-being literature and revealing commonalities across measures. We will use this review to guide the development of a community well-being model or framework, hopefully condensing many such frameworks into one. The second part of the project involves considering the ethical implications of this approach. Data ethics is still in its infancy, and there exist very little legal coverage for data and AI, save for privacy laws around sharing private information. Therefore, local government must tread lightly if they wish to engage in such analyses, particularly since communities often do not trust government. Theoretically, any kind of data is useful for being folded into the AI when calculating community well-being indices. But some of this data (e.g. CCTV footage, tax payment history) makes people more anxious when used by government than others (e.g. parking violations, library fines). So, the second part of the project will develop guidelines and best practices for engaging with all these types of administrative data. The third part requires building the AI, but since the ethical concerns are not resolved, we will simulate data, and evaluate any additional insight derived from such an approach in the final part of the project. Specifically, we may be able to better target where services need boosting, or where policy interventions have proved useful. We can predict areas at risk for social upheaval, and the impact of global challenges like the cost-of-living crisis or the pandemic on communities.

Can you tell me more about those ethical concerns?

As I mentioned before, the ethical concerns are many, and involve privacy and consent primarily. Privacy involves having a say in when your data gets shared and by whom to whom. Consent involves prior approval of your data for experimental, marketing, or other purposes than those stated at the time of collection. Lots of administrative data is not collected with prior consent for research and insight, so thinking through what needs to change to provide such consent is key. Finally, community voice is needed when considering these issues, so we will involve community organisations and stakeholders in these discussions, ensuring that any suggestions are sensitive to community needs. These may lead to innovative approaches, such as encouraging ownership of the data by the community, providing more say in who uses it and how, as well as serving as a resource for the community.

Do you know of other studies that have used administrative data in this way before? Do you think this work will inspire future projects, and if so what might these be? 

There are a few local governments across the UK that have started using this approach to administrative data. However, ethical concerns restrict further involvement, as well as a lack of a data infrastructure to facilitate data sharing within and across local governments, and a lack of data scientists working in local government give budget restrictions. However, current efforts are still well underway, such as the Civic Strength Index project in the Greater London Authority, which still relies on self-report data largely, but also focuses somewhat on behaviour. Future research will go beyond behaviour and integrate information about the environment, such as pollution level, as well as economic indices to better understand communities and the impact of policy on them.