Alex Massie is, of course, correct to say that it is a mistake to draw any inferences from something as freakishly unusual as the Philpott case. Why, then, has the mistake been so common? I suspect there are two biases at work.

One is the confirmation bias. If you think the welfare state incentivizes bad behaviour such as laziness or fecklessness, you'll see Philpott as an extreme confirmation of your prior.

We should, here, note an asymmetry. Those who see Philpott as confirmation of the "bad incentive" hypothesis don't regard HBOS's bad lending as evidence that low top taxes have perverse incentives, or Stephen Seddon's murder of his parents as evidence that low inheritance taxes have bad incentives. This asymmetry cannot be explained by sound statistical reasoning, unless your prior is that the welfare state has stronger bad incentives than the tax system.

The second bias is a form of the halo effect.Hostility towards benefit claimants is founded upon a moral instinct - the norm of reciprocity. People fear that claimants are getting something for nothing, that "hard-working tax-payers" are being ripped off. The halo effect - the tendency to believe that all bad things are connected - lead people to think that such scrounging is also a macroeconomic problem. Which it isn't. This bias is exacerbated by our natural tendency to exaggerate harms done to us.

Herein, though, lies a problem. I fear this reciprocity norm is an atavistic instinct with an evolutionary basis which is no longer relevant to post-industrial society.

In hunter-gatherer societies struggling for subsistence the man who does no work and expects others to provide for him is a threat to the very existence of the tribe. It's natural, therefore, for a norm of reciprocity to emerge so that shirkers are stigmatized and shunned.

But we no longer live in that world. We have the opposite problem. There's not enough work to go round. In this world, the shirker does not threaten our existence. If anything, he's a help, as his not looking for work increases the chances of others getting it. And, remember, on average the unemployed are significantly unhappier than others.

My point here is a depressing one. The tendency to stigmatize benefit claimants, and so to regard Philpott as sufficiently representative of them to justify inferences (or dog-whistle "questions") is not the sort of error that can be corrected by evidence. It arises from biases and beliefs which are hardwired into us.