How can academia and government work together better to make evidence-based policy?

---

By: Natalie Gold

Senior Director and Head of Trials, Behavioural Practice, Verian UK

Behavioural public policy is the use of behavioural science to assist the design and implementation of public policy. As a part of this movement, both academics and civil servants have an interest in effective collaboration and “knowledge exchange”, including the sharing of ideas, evidence and expertise. However, the two sectors have different ways of operating and different – sometimes competing – goals. If they can understand each other better, then they can get the most out of working together.

In order to achieve their different ends, the sectors need different types of activities. Academics conduct research: gathering or collating data (which can be qualitative or quantitative), conducting secondary analysis of data, or modelling. This creates knowledge. Government and the private sector also use consultancy, or the giving of expert advice (which may come from academics), to help decide what to do. However, even when government is looking for research, their requirements may differ from academia.

Research in academia vs research in government

Academia and government are aiming to do fundamentally different things. Universities are in the business of knowledge creation, and there is a corresponding emphasis by academics on theory and hypothesis testing. Even peer-reviewed journals that publish applied research often require a contribution to theory from the papers that they publish. At the most extreme end of theory testing, some behavioural experiments (especially laboratory or online experiments) focus on teasing out different causal mechanisms that could underpin a phenomenon. This can lead to a lot of knowledge about what mechanisms exist in principle, without much knowledge about how effective they are in practice, outside of the lab. So we shouldn’t be surprised that some findings obtained in the lab do not transfer well to the field. For instance, theoretical evidence on the effect of gain vs loss framing in laboratory experiments is very robust, but evidence of effects in the field is at best mixed (for instance, in vaccination and disease prevention behaviours).

In contrast, government attaches more importance to practical and actionable knowledge, which can be used to guide policy. The emphasis is on what will be effective, not why it is effective, and on estimates of how effective it will be (or the “effect size”). Related to this is a slightly different emphasis on generalisability from that found in academia. Everyone is concerned to find “true” and reliable results. However, academic research is implicitly looking to make general lawlike claims, whereas government is more focussed on evidence about this instance or this policy. Which turns us full circle back to theory testing: academics are invested in specific theories, whereas policy makers may be less concerned with what they see as “arcane theory”.

The sectors also have different audiences. Academics’ primary audience is other academics. Of course, they are also interested in disseminating their knowledge to policy makers and the public; but they often don’t know how to reach policy makers effectively. In contrast, government research is for the ears of policy makers; and the government is ultimately answerable to the public. This connects to publishing: academics are strongly incentivised to publish in peer-reviewed journals; while government may publish some of its work in reports for reasons of democratic accountability, but that is ancillary.

This difference in ultimate aims and audience also has implications for sharing of data. Many academics believe in open data and open science. However, in government, creating open data is not a part of the core business. Even if civil servants are in principle in favour of making data accessible, creating and saving anonymous datasets is extra work that is not a part of the day job; and in trying to publish data may come up against organisational policies about how long they are allowed to hold data and worries about potentially costly breaches of GDPR.

The relationship between research design and budget also runs in different directions in different sectors. For academic grants, a research design is proposed; money is requested to fulfil the needs of the design, as are timelines, which are relatively long. In contrast, government research is much more strongly driven by deadlines and budgets (which often run with the financial year), with research plans being proposed to meet those needs. 

Channels for knowledge exchange between academia and government

If academic and government research are so different, one might wonder how they academics and policymakers can be brought closer together?

A first thing to realise is that academics are not especially needed to do government research. There is a whole ecology of researchers, including inside government in the Government Social Research profession; and outside government in research agencies, which exist to conduct research that fulfils client needs, including some that specifically serve social or government needs. If a policy maker has a research question and a budget, and needs a rapid policy-focussed answer, then they have places they can go to for that.

Instead, a better place to start is to identify what added-value academics have to offer policymakers; and then to construct channels that will allow them to share it. Some starters for ten:

Best practice is often defined in academia and filters outwards, so academics could:

• define pragmatic or “rapid” best practice – to help those producing work to timelines, e.g., guidelines for rapid evidence reviews, and set achievable standards for such work
• establish norms of good practice on data storage and establish places to store data or signpost to repositories – to make data sharing easier and support open science outside of academia.

Academics are more likely than other researchers to be at the cutting edge of theory development. For instance, it’s widely accepted in academic circles that “priming” is a victim of the replication crisis, but how many non-academic users know that the “P” in MINDSPACE is discredited? So academics could:

• provide continuing professional development for those outside academia, or train practitioners
• keep practitioners’ tools up-to-date, giving guidance to help them apply theory to practice.

Finally, the learning from knowledge exchange need not all be in one direction. Academics have research programmes which can be relevant to policymakers and which they may hope will have impact. In order for that to happen, they need places where they can get guidance on what makes research useful to policy makers and on what makes policy makers pay attention to research. This is why forums such as the LSE KEI Behavioural Public Policy Exchange Group, which foster a deeper connection between academics and professionals are so valuable.