It's not just me

I was encouraged to see the Royal Statistical Society questioning the lack of detail in the government swine flu statistics - though I fear this is one press release that is going to pass most journalists by. President David Hand had written to Sir Liam Donaldson with a few questions. You will find his letter online but the key points are:
The dangers of producing weekly reports on swine flu instead of daily ones

• Some vagueness on statistical standards
• Lack of demography in the data
• Lack of other data relating to sex and age of patients

And interestingly a concern that those who die from H1N1 away from hospital might be missed out of the statistics. This line from David Hand's letter particularly interested me:

"Readers of the UK’s weekly H1N1v updates need to know numerators and denominators, not just percentages, and the empirical basis for multipliers used in inferring from GP-consultations for influenza-like-illness (ILI) to headline figures on the estimated number of H1N1v cases. Moreover, the provenance for intervals of uncertainty should be properly explained."

I think that's what I've been saying - but the RSS puts it in their own jargon. It seems to be saying there's a basic lack of statistical data missing. If so, journalists should be highlighting the government's failure to understand the necessity of such information.

But I fear not much is going to come of this. Journalists and maths/statistics just do not mix. Very, very few journalists have the understanding to look at a graph or table of stats and ask pertinent questions. And there's another problem. The Royal Statistical Society can't talk in plain English. They need a press officer who can explain in layman's term precisely what their concerns are with H1N1 statistics. It's no good hiding behind "oh it's all very complicated and can't be simplified". It's the job of press officers and journalists to make sure the public can understand such information.

One bright light on the horizon: The RSS does hand out awards on statistical excellence in journalism which hopefully encourages the hacks to improve their act.