evidence schmevidence?

Call it biting the hand that feeds it, but I've got a newly published article in Open Democracy, criticising the cult of evidence-based policy in Government. To be precise, the piece is not attacking the intentions of evidence-based policy, which are eminently sensible, but what I see as its political over-reach. Widespread recognitition of evidence in a vibrant and self-confident democracy would be unambiguously welcome; what is worrying is when it is used as a way of either ducking moral questions, or hiding one's answers to them. Here is the heart of the argument:

There is an important dimension to democracy that the evidence-based policy movement appears to miss. Democracy is not just about the desirability or otherwise of outcomes, but about the mechanisms used to select and achieve them. The government is bewildered that measured improvements in public services have not been noticed by the public, and that measured falls in crime have been accompanied by increased fears of crime, then tends to blame the media for excessive pessimism. But when policy is constructed and evaluated by anonymous statisticians, one cannot assume the public will share the government’s assessment of it. In politics, it is not enough for something simply to be the best option; people must reach agreement that it is the best option, a process which then becomes constitutive of that option’s value.

Comments

[Andrew Brown]

You're right that evidence that is not led by values leads to politics that free of beliefs and then we may as well give up and agree to live in a technocratic "utopia".

Perhaps you're right and the problem has been that those of us active in politics haven't been explicit enough in talking about the values that we bring to bear on the evidence and therefore the policies we espouse.

That said there seem to be plenty of politicians who seem to manage perfectly well without evidence in their policy formulation.

[Luke Smith]

A worthy target. In Tuesday's Guardian Polly Toynbee suggests that the reliance on statistical evidence is undermining trust in evidence itself:

Labour earnestly tried to create trust with an array of audited targets, monitoring and league tables. But it has had the perverse effect of raising suspicion of cheating and measuring the wrong things. It has subtly undermined trust in political leadership: targets have become coconut shies.

She argues that politicians need to create trust or 'belief' first, and then support it with evidence. Toynbee thus argues that politics is more of an art than a science. However, in science theories come first, they are then explored with evidence. Theories are informed significantly by values (as shown by the current 'intelligent' design debate). So even in science there is no frictionless agreement on evidence based rationality. Such consensus is even less realistic in politics so it would seem to accept the presence of differing values rather try smooth over them with 'facts'.

[Will Davies]

Andrew's point is well made. Someone pointed out to me yesterday that there are plenty of areas of policy where democracy (or at least, populism) entirely over-powers evidence: we know that putting bobbies on the beat is not the best way to tackle crime, but we do it anyway; we know that our current drugs policy isn't working, but public opinion prevents any room for maneouvre.

[jamie]

I agree with the general thrust of her column but could Polly Toynbee be missing something? EBP is embedded in that politicians (of all stripes) generally try and cite the evidence on which they have based their decisions. However, I don't think it would take 45 minutes to think of some areas where policy is determined and then evidence is sought.

Perhaps trust is most undermined under two conditions. First, when it is susbsequently revealed that evidence is sought to bolster pre-determined policies (this is by no means always the case). Second, as Will points out, when journalists are unable to determine between poor evidence and credible evidence in the first instance which is later revealed to be wrong (see MMR).