Charities and government have similar aims. We all want people to be happier, healthier, do better at school, get jobs, stay out of prison and so on. Hence, across Britain, and at all times, there is a huge amount of charitable and government activity taking place towards these aims.
Unfortunately, very little of this work is evaluated properly – so we don’t learn much from it. Government departments do commission and publish evaluations of flagship projects, but these are expensive to do so are much less common among charities. This means that compared to, say, medical science, those of us working in social policy evaluation still feel we are starting the journey towards understanding what kinds of support and social programmes work for different people in different circumstances.
One of the most positive developments over the last decade has been the “data labs” model for accessing and using government data for research. The starting point for the data labs idea is that the government routinely collects data that can help charities understand what difference they have made.
For example, if you are a small charity providing mentoring to people leaving prison, you will really want to know whether the people you supported stayed out of trouble. To do this, the charity can either try to collect the data itself, which is difficult and costly to do at any scale, or they could try to get this information out of the criminal justice system. The relevant data exists: the Police National Computer is an up-to-date record of all convictions against people in England and Wales. But, of course, the PNC is very sensitive data. Legally, it is almost impossible for charities to access its information.
The Justice Data Lab, launched in 2013, is an elegant solution. It allows charities and other organisations working with offenders to get reoffending data from cohorts of their service users, thereby ensuring no individual’s criminal records are revealed. Cleverly, it statistically derives a control group of similar people who did not use the service so results can be compared. If the difference in reoffending between service users and the control group is big enough, then we have a good basis for thinking the programme has made a difference. As a real example, this means the Prison Education Trust can now show that their work to provide prisoners with educational opportunities is associated with a meaningful reduction in reoffending.
For us at New Philanthropy Capital – where we focus on improving the impact of the charity sector – data labs are a game changer. Not only do they enable robust analysis of charities’ services, beyond that which has been possible before, they also help society to accumulate knowledge through multiple analyses of different programmes, thereby improving our understanding of what works in criminal justice and a wide range of other policy areas.
This is why we are especially pleased to see the launch of a new data lab from the Department for Work and Pensions last month with an analysis of the Spear Programme, a six-week course for young people not in education or employment and delivered by the charity Resurgo.
“Other departments have data that could achieve enormous social good through a data-lab approach. Most notably the Department for Education and the Department of Health and Social Care – both of whom currently invest billions in contracts and services but miss a key opportunity to learn about their value”
The DWP Data Lab uses a similar method to the JDL, and this first analysis shows that the Spear Programme is associated with a five-to-12 percentage point reduction in the number of young people classified as NEET after one year, and seven to 13 more weeks in employment in the two years following the programme. So, a successful programme.
The DWP Data Lab has been a long time in the making. The department has committed resources to it and has had to overcome issues like making sure the right data protections are in place and on accessing, cleaning, and analysing the data. If the story of the JDL is anything to go by, the method and outputs will become more sophisticated over time as more charities use it.
Both departments deserve huge credit for their commitment to making data available for public good – and for the dedication and creativity of the officials involved. The Royal Statistical Society awarded the JDL its 2014 prize for official statistics. Internationally, several countries are interested in applying the model themselves, and the United Nations invited the JDL to speak at its 2021 conference on crime prevention.
There is still a long way to go. Other departments have data that could achieve enormous social good through a data-lab approach. Most notably the Department for Education and the Department of Health and Social Care – both of whom currently invest billions in contracts and services but miss a key opportunity to learn about their value. So, I’d like to encourage all analysts in government to think about what, if anything, they could do to make their own data available, because here is a real opportunity to make a breakthrough in our shared understanding of what works to help people.
James Noble is associate director for data and learning at the New Philanthropy Capital think tank