1. Aren’t there a lot of cognitive biases here? Camilla Cavendish says a mix of groupthink and Bayesian conservatism may have led social workers to under-estimating the danger Baby P was in. But aren’t we also at risk of committing the hindsight bias? With hindsight, things look as if they were inevitable when they were in fact only vague possibilities?
And of course, there’s also the confirmation bias. Every pub bore who thought social workers were Guardian-reading bed-wetters thinks this shows their ineptitude.
2. Aren’t we holding social workers to impossibly high standards? I mean, we don’t expect the police to catch all or even most criminals, we don’t expect hospitals to avoid all unnecessary deaths, and we certainly don’t expect politicians to make perfectly rational judgments. So why expect perfection from social workers? The fact is that they have to take huge decisions under immense uncertainty. Errors are inevitable. The question is: what sort of errors should they make? Should they err on the side of intervening in which case they might prevent abuse but at the expense of breaking up some tolerable families? Or should they regard taking children into care as a last resort, and so risk children’s lives? In conditions of necessarily limited knowledge, one of these mistakes will be made?
3. Have we drawn the line in the right place here? The legitimate* reason for hating seeing children put into care is that they do so awfully badly there; four-fifths end up with no GCSEs.
But is this because state care is worse than even poor parenting? Or might it instead be due to selection effects? If kids from only the very worst parents go into care, you’d expect them to do badly at school simply because they have a poor genetic inheritance and/or a bad start (pdf) which perhaps even good care can do little to ameliorate (except at a huge price?) If this is the case, then our aversion to putting kids into care rests at least in part upon a simple fallacy - post hoc ergo propter hoc - and we are endangering lives unnecessarily.
4. The DCSF says “proper procedures” were followed after it got a complaint about Haringey social services in 2007. This has naturally caused outrage; how can “proper procedure” lead to the death of a child?
So how should we interpret it? One possibility is that this is mere procedural fetishism - the belief (not confined to the public sector) that as long as the boxes are ticked and the papers in order, then outcomes don’t matter.
But there’s a more benign possibility. Under uncertainty, errors will be made. In this case, we cannot judge processes by single outcomes, as even the most effective process will lead to bad outcomes sometimes. Instead, the best we can do is to have in place a process that, on average, minimizes error. And in this case, following “proper process” will be the right thing to do not because it will always achieve the right outcome - nothing can do that - but because it minimizes error. If so, the DCSF is right.
But how can we tell which process minimizes error? We can’t run market-style experiments in which many differing processes compete alongside each other.
Which leaves us vulnerable to the possibility that the claim that “proper procedures” are in place is no more than a mix of unscientific wishful thinking and special pleading.
I don’t pretend to have any answers here. All I want to do is raise the possibility that perhaps a tiny bit of progress can be made by at least avoiding the most egregious intellectual errors.
* Is keeping families intact really an end in itself? I'm not sure.