Almost nobody is against evidence-informed policy and practice. Like motherhood and apple pie, there's not a lot there to object to. After all, who wants to take to a stage and argue in favour of evidence-free policy?

Yet, despite this near universal support, the use of evidence has been far from inexorably rising. Zombie policies – those that we believe had been killed off by evidence – rise again and again, while the truly evidence based struggle to gain traction.

Why is this?

First, in our experience, it is because our definition of evidence is too general. Research methodologies that are ill-suited to answer questions about the impact of a particular intervention are too often used to prove that the intervention is ‘evidence based’. Far too little evidence and research is concerned with a question of central fixation to policy makers and practitioners: ‘If I do X, what will happen to Y?’ An analysis by the National Audit Office between 2006 and 2012 found only half of the departments looked at had ‘evaluations [that] were of a sufficient standard to give confidence in the effects attributed to policy because they had a robust counterfactual’ (National Audit Office, 2013, p 7).

Second, in many, many fields of social policy, saying you want to base your decisions on evidence is like saying you want to commute by flying car – the evidence simply does not exist to base your policy on it. For example, out of the UK government's 108 most complex and strategically significant projects – the socalled Government Major Projects Portfolio, costing £432 billion of taxpayers’ money – only nine are evaluated robustly (National Audit Office, 2021, p 6). The majority – 77 major projects – have no evaluation arrangements at all. And these are major projects. There is no data kept on how many of the ‘business as usual’ activities in the rest of government are evaluated.

Third, even when there is evidence, it is not explained clearly, or in the right way, to the right people, at the right time. Too much evidence exists behind academic paywalls, or, where it is not, behind academic language, spoken between academics and researchers.