Every statistic may be manipulated. Two versions of the same data: - Bad outcome X rose by 55% over last year (under Party Y, those scoundrels) - Bad outcome

Every statistic may be manipulated. Two versions of the same data: - Bad outcome X rose by 55% over last year (under Party Y, those scoundrels) - Bad outcome X affected 17 people this year, up from 11 last year, out of a population of 35 million