What makes a good scientific policy paper, defined as a paper likely to influence and improve policy decisions based on science?
The starting point for any piece of communication, and a scientific paper is a form of communication, is: who is the audience?
It should be the easiest to fix if, as academics and scientists, we are serious about trying to improve policy.
I was asked to write it having just stopped being Chief Scientific Adviser at the UK Department for International Development (DFID).
Policymakers discussed in this paper, however, are those having to operate at a larger, generally national or multinational scale.
Whilst there are many variations on this, there are some common features which are often forgotten by those who think they are writing for policymakers.This is the responsibility and skill-set of academia, although policy-makers can help prioritize key questions.The second is a demand-side problem, with policymakers unwilling or unable to take account of good existing evidence.Models should, wherever possible, allow policymakers to vary assumptions.Objective, rigorous, original studies from multiple disciplines relevant to a policy question need to be synthesized before being incorporated into policy.Barriers include the evidence not being there; lack of demand by policymakers; academics not producing rigorous, relevant papers within the timeframe of the policy cycle. Academics underestimate the speed of the policy process, and publish excellent papers after a policy decision rather than good ones before it.To be useful in policy, papers must be at least as rigorous about reporting their methods as for other academic uses.Papers which are as simple as possible (but no simpler) are most likely to be taken up in policy.Most policy questions have many scientific questions, from different disciplines, within them.The accurate synthesis of existing information is the most important single offering by academics to the policy process.Since policymakers are making economic decisions, economic analysis is central, as are the qualitative social sciences.