You know what it's like. You're sitting there quietly minding you own business when two men of a certain age start holding forth about something they obviously know nothing about. You try to ignore it: it's none of your business. But the voices get louder and louder as they become more passionate in their uninformed conviction. The first lines of 'Oliver's Army' are repeating themselves in your head:

'Don't start me talking/I could talk all night/ My mind goes sleepwalking/While I'm putting the world to right'.

You want it to stop, you want to explain to them that their premises are false, that the facts they are citing are nonsense, that they are making a spectacle of their own ignorance. But you can't, because you know what will happen.To open your mouth would involve entering into a dialogue from which there can be no escape. Better to leave them to it.

For some reason, I am reminded of this all-too-common experience by this dialogue from yesterday's Andrew Marr show.

About 15 per cent of all of the families have more than two children - the average is about 1.8 in the UK - and they cluster hugely down in the bottom couple of deciles and right at the top of the income distribution. Across the broad swathe of the middle …
So the very rich and the very poor have lots of children?
Is it right …
And most predominantly in the very, very bottom areas; and a large, large proportion, the majority, are mostly out of work.

I know I'm going to regret this, but let's get it over with....

The only statement in this extract which is not grossly inaccurate is 'about 15% of families have more than two children'. Families with children of all sizes, not just those with more than two children, are 'clustered' towards the lower end of the income distribution because for the purposes of monitoring living standards household incomes are adjusted to reflect family size and composition, a process known as 'equivalisation'. 25% of all children are in the bottom two deciles of the equivalised income distribution and 32% of children in families with more than two kids (on the After Housing Costs measure). The difference in 'clustering' between family sizes is neither 'huge' nor unexpected to anyone familiar with how incomes are measured. (It would be considerably greater if the incomes of families with children were not boosted through child benefit and child tax credits.) There is /no/ clustering of families with more than two children at the top of the income distribution: on the contrary, only 8% of children in these families are in the top 20%. The claim that the majority of these families are out of work, if that is what Mr Duncan Smith is saying, is simply false: 90% of those with three children, 80% of those with four and 67% of those with five or more have at least one parent in employment, as shown in the in-depth research on large families commissioned by Mr Duncan Smith's own department . (There could be some prevarication going on here: what Mr Duncan Smith says is that the majority are 'mostly out of work' which at a pinch could be taken as including families where /one/ of the parents isn't working, although this is obviously not what his statement would convey to most people.)

Having got that off my chest, I'm getting out of here before the pub bores have a chance to respond.