Thanks to all those who contacted me about air pollution, (mis)use of
statistics, or (mis)use of allstat. I hope I have made my peace with
everyone. One exchange that seems of sufficiently wide interest to justify
another allstat use was with Friends of the Earth, whose analysis prompted
the original media story. A poll of a few university staff early in the
week indicated that wherever they had seen or heard the story, the message
they had taken in was that air pollution was "the worst ever".
---------- Forwarded message ----------
Date: Tue, 18 Jan 2000 12:05:12 +0000 (GMT)
From: R. Allan Reese <[log in to unmask]>
To: Tony Bosworth <[log in to unmask]>
Subject: Re: Air pollution statistics
On Tue, 18 Jan 2000, Tony Bosworth wrote:
[RAR's message had been passed on, stating]
> > The media appear content to repeat the description by Friends of the Earth
> > of UK air quality as "the worst ever" or "the worst since records began."
[TB requested]
> Can I please clarify that Friends of the Earth did not describe UK air
> quality as the worst since records began. We said that 1999 showed the
> biggest deterioration since records began, an interpretation accepted by
> the Government.
[RAR responded]
See, for example,
http://news2.thls.bbc.co.uk/hi/english/sci/tech/newsid%5F604000/604162.stm
which has the headline I quoted, followed then by the wording that you
have now provided.
If you mean "the biggest year on year change in a negative direction",
that might well be a strictly factual description, but in terms of
statistical reporting is so perverse and emotively misleading that I
think it underlines my charge that quantitative data are being spun and
abused for political purposes. This is, I repeat, not a statement about
the measurement of pollution or any real trend, it is a disbelief in the
competence of the analysis. These measures should be reported and
interpreted in the same way that other quality control (QC) measures are
used, and those methods have been well described for over fifty years.
R. Allan Reese Email: [log in to unmask]
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|