We all know statistics are misused to justify policy, but there is less awareness of the degree to which statistics themselves can be faulty.
The mix of faulty statistics and faulty political motives means the public needs to receive policy debate within a wide frame of probable error.
For example, a study in the UK found massive miscoding at the hospital ‘coalface’ had undermined the reliability of health service statistics [my italics].
in 2009-10, nearly 20,000 adults were coded as having attended paediatric outpatient services, and 3,000 patients under 19 were apparently treated in geriatric clinics. Even more striking, between 15,000 and 20,000 men have been admitted to obstetric wards each year since 2003, and almost 10,000 to gynaecology wards. Nearly 20,000 midwife “episodes” – NHS jargon for completed treatments – were with men.
This sort of corruption at the source is widespread [pdf]. Yet, when statistics are recieved by Government, Parliaments and government agenices, they are treated as absolutely correct.
It is not necessarily the job of politicians and their officials to second-guess statistics. Their focus is usually on applied ideology, public acceptability and policy decisions. Data is an input to help inform, and to justify, their decisions.
So it is down to the recipients of this process – voters – to recieve the data and the political policy decisions within a frame of human error.
The data is probably wrong within some sort of margin, and that result is probably being misconstrued by people with preferences for a particular outcome.