Photo: Caerphilly from the hills from Jeremy Segrott Flickr photo stream

Government departments and their frontline staff are constantly experimenting with new policies and different methods of delivering public services to try to discover better ways of doing things. For example, it’s thought that some health boards in Wales are now running more than 50 different parenting programmes without any evidence or systematic evaluation to inform commissioning. The families engaging with these services are effectively taking part in a huge experiment but there’s no effort made to measure the outcomes scientifically. Some of these programmes could be doing more harm than good. We also don’t know which ones work best, or why, so it’s not possible to scale up the most effective programmes. This means we’re living in a public services ‘lab’ but there’s no resources to do, or coordinate, the scientific lab work of measuring, observing and theorising change.

A solution is for new public services innovation ‘labs’ to generate better evidence more routinely and more responsively, much like clinical lab teams do. One example of this is Y Lab in Wales, led by Cardiff University and Nesta in collaboration with Welsh Government, which I’m involved in setting up. The Cabinet Office’s Behavioural Insights Team – aka ‘Nudge Unit’ – has already pioneered responsive, policy trials in Whitehall. As governments begin to answer questions about public services more rigorously, the next challenge is to create a more scientific culture across the whole of society to address more complex social problems. This means building on the success of the Nudge Unit, and going even further, with more sociologists in our public services labs, more social policy trials, and more frontline staff working routinely with scientists to generate new ideas and evidence through everyday trial and error in the ‘living lab’.

Traditionally, principles of good science such as acknowledging uncertainty and experimentation have been rejected by politicians and frontline public services staff as a sign of weakness. However, David Halpern’s inside story of the ‘Nudge Unit’ documents growing support from Cabinet Ministers and senior Whitehall officials for experimental methods. His team have shown that behavioural science insights will help government departments operate more efficiently but they have also had a much wider scientific impact on government too. The experiment with experiments was a success and they’ve now done over 150 trials and worked across all major policy areas. There are changes on the frontline too. Hundreds of schools are now getting involved in trials for the first time thanks to the Education Endowment Fund (EEF) and Ben Goldacre’s work with the Department of Education.

The question is no longer whether we should apply the ‘geek manifesto’ principles of scepticism and scientific testing to public services – but how should we do this? We shouldn’t restrict ourselves to just nudging or confine ourselves to specific settings such as schools. If we’re going to solve more complex problems, such as massive inequalities in health, there are three major challenges – and opportunities – for making public services science labs work.

First, sociologists need to do scientific ‘lab’ work and not just critique. We know that, in isolation, nudging people towards the salad bar won’t reduce obesity and health inequalities. This sociological critique of behavioural insights is relatively easy. The practical challenge is harder: we now need to apply a sociological imagination, as well as behavioural insights, to design and evaluate new public service solutions. The success of the behavioural insights team in fostering a culture of experimental government allowed them to rapidly test and improve many operational features of government (such as new tax letters to generate more revenue) and evaluate health improvement nudges (like new NHS text messages to reduce missed appointments). The challenge now is to do as many responsive, low-cost trials of more complex social interventions.

Second, there is a no stopping rule or simple ‘right’ and ‘wrong’ answers when it comes to supporting poor families or delivering health services. We’ll never find any ‘iron laws’ of public services, nor can we develop ‘social vaccines’ that will work perennially. Public services are unique to each country too. But this complexity isn’t an argument against controlled trials – it’s actually an argument for them. It does mean we need lots more trials to generate sufficient evidence, which means doing trials cheaper. To trial more public service innovations at scale and at low-cost we need to train and support more staff on the frontline, such as teachers and the police, to do research, linking them with teams of scientists and routine data to help them. This is beginning to happen in schools already via the EEF evaluation toolkit and the Wales school health research network.

Third, lab work is about discovering new ways of doing things. More marketisation, choice and competition isn’t the only option for public services reform. We need science to drive innovation though. This also means moving towards a wider model of experimentalism and frontline, citizen science. Lots of ideas need to be tested on a small scale first before being potentially being scaled up as part of larger policy trials. It’s only through repeated failure that we learn and adapt. This is how new discoveries are made. Advances in medicine don’t start with costly, mega-trials. First, there are “bench” scientists trying – often failing – to discover new ways of meeting the needs of health staff and their patients.

Being scientific is something we all do naturally, and we are more data driven than ever before, but we’re still typically more comfortable being scientific sceptics on our own rather than collectively as a society. A scientific society is one that is engaged in a constant cycle of curiosity, critical thinking and experimentation. This is how we made the advances in medicine, engineering and computing science that improved people’s lives in the twentieth century. There is no society in the world, yet, that is anywhere near as good at applying these same scientific principles to its public services. Experimental government is coming of age in the UK. But it’s an unfinished, long-term project. Sociologists and the rest of the UK need to catch up with the behavioural psychologists leading the way in London, and we need to help frontline staff ask and answer questions directly. This means accepting and learning from failure too.