Wednesday, 3 November 2010

Correlation and Causation

 There was an item on the Today programme yesterday morning about the research Frank Field is doing into poverty, parenting and education. It also referred to a Sutton Trust report that said children from the poorest backgrounds are twice as likely to have behavioural problems at age 5 as more affluent peers. The presenter then made the classic logical error of assuming that because there is a correlation, there must also be causation, and that the causation must run one way, in this case the poverty causing the behavioural problems.  This is utterly utterly absurd. To see how absurd it is, let us consider what would happen if the families of these kids were to win the lottery. Do we really think that those kids would overnight become little angels? Ridiculous. If anything, they'd use their greater spending power to be even more of a nuisance. Joey Barton and Jermaine Pennant didn't have their behavioural problems suddenly cured by a pay packet of several thousand a week.  And yet the principle that giving poor people more money will miraculously improve all of their life outcomes has been the principle that underpinned most of the social policy of the last decade. It completely ignores the fact that culture and values play an enormous part in behaviour and educational success. That's why certain ethnic groups - Chinese girls, for instance - do incredibly well at school despite poverty.  Culture and values don't cost money.

No comments:

Post a Comment