That tweet by Robert Murdoch has been widely condemned. But nobody has pointed out that Murdoch's error is not confined to either him or to Islam's critics.

Murdoch said:

Maybe most Moslems [are] peaceful, but until they recognize and destroy their growing jihadist cancer they must be held responsible.

The inconsistency here is obvious. Murdoch did not, AFAIK, hold right-wingers responsible for the cancerous beliefs that led Timothy McVeigh or Anders Breivik to commit mass murder. Nor did he consider sexually frustrated men responsible for Elliot Rodger. So why hold Muslims responsible for jihadism?

The difference is that, in Murdoch's mind, Breivik, McVeigh and Rodger - to name but three - are atypical of white right wing men and are isolated instances. But, he thinks, Muslims and jihadists have more in common. So, whereas right-wing men are not to blame for the sins of others, Muslims are.

There's a name for this. It's the outgroup homogeneity effect, or the "they are all alike; we are diverse" bias.

And here's the thing. Murdoch is by no means alone in this error. Like all cognitive biases, it's a widespread one. We see it, for example, when: some feminists claim that all men are rapists; when critics of orthodox economics fail to see how diverse the subject is; when lefties attack "greedy bankers"; when people think they are experts on a country because they have visted it a few times; when right-wingers think that because I'm a Marxist I somehow endorse Stalinism; or when torture is claimed to be "inconsistent with our values" when our side does it but typical of "them". And so on.

You might object that there is a strand of violence within Islam. Maybe, but so what? There is also a tendency towards violence in maleness or in several political ideologies - including Mr Murdoch's and (yes) mine. But there is a massive diffence between those who reject that tendency and those who embrace it. Failing to see this difference - and failing to see that Muslims were also victims and heroes last week - is just stupid.

There is, though, a way to guard against the outgroup homogeneity effect. Whenever we see something, we should ask: from what sample is that drawn? Is it typical, or an outlier? Merely asking this question reminds us that human beings are a very diverse bunch.