After the Institute for Fiscal Studies released figures on changing levels of income in Britian, it was predictable that different newspapers would manipulate the figures to present them in a way that suits their political line.
In short, the data shows an “average” 0.2 per cent drop in household incomes between 2002/3 and 2003/4. This interpretation suits right-leaning newspapers. But, as any foo’ (who has taken freshman-stats-for-trees) know, taking the mean skews the data because it is disproportionatly affected by changes in income among high earners. Using the median or splitting the data into income quintile classes provides a much more interesting picture. Jim of Our Word is Our Weapon provides an elementary statistics primer explaining all this and showing the real story with his trademark charts. He also provides an income distribution chart showing the difference between mean, median, and mode. (Want to know where you belong on this graph? Go here.)
It really is a good example of how journalists can lie (or at least demonstrate incompetence) with statistics. British Spin goes so far as to charge that Evening Standard’s Finance Editor Anthony Hilton, “either cannot read a press release or is terminally dishonest” after he wrote that the IFS data show that “most people” are worse off.