Correlation versus causation

  • by

Correlation versus causation – I run into this issue most days when analysing stats but I quite liked this one as it is a good example – when comparing the levels of pay amongst individuals working within the engineering industries it appears that if women are seeking equal pay within one of the associated sub-sectors (set out by Engineering UK) , the next best thing to changing gender is actually ‘just’ to get married. Does getting married really improve your salary – unlikely of course (more likely age is the key factor here at least) but many would simply look at this and take as read. Possibly more interesting though is why ONS have a variable splitting out females marital status and not males? – on this one I can’t hazard a guess.