IWF statistics

[Richard Clayton]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

In article <BII9AdmpFPRJFA2C@perry.co.uk>, Roland Perry <lists@internetp
olicyagency.com> writes

>It was clear enough to me, indeed I simply can't see what the confusion
>is.
>        "When child abuse images are hosted in the UK, the IWF can
>        identify and notify the host and the host removes them. Its work
>        has cut dramatically the volume of illegal images hosted in the
>        UK.

Changing topic slightly (and hence the subject line), this is a much
repeated claim....  which I can trace back to the IWF 2003 Annual
Report, which gave the following figures for illegal content hosted in
the UK:

        1997    18%
        1998    12%
        1999     5%
        2000     3%
        2001     3%
        2002     2%
        2003     1%

Later reports show that the UK figure has remained under 1%.

There's a problem with this : and that's that these are percentage
figures of a rapidly growing number of reports (more and more of which
are unique).

- From the same document, the actual _numbers_ of reports (consistently
about two thirds are not of illegal material) are:

        1997     1291
        1998     1991
        1999     4297
        2000     8942
        2001    11357
        2002    17868
        2003    19553

So if we take the percentages of these, and divide by 3 we get an
indication of the number of UK sites:

        1997     77
        1998     80
        1999     72
        2000     89
        2001    114
        2002    119
        2003     65

hmm... doesn't look so compelling now, does it ?

However, I believe that the UK figures include Usenet servers; viz in
the 1990s a lot of the IWF's business centred around child sexual abuse
images posted to Usenet. Although in some cases it has been clear
whether or not the poster is in the UK, it generally isn't possible to
tell at the start of an investigation.  Viz: most of the reports that
they were processing in the 1990s they had no idea if it was UK linked
or not -- but the Usenet servers were in the UK.  They seldom publish
methodology :( so it's completely unclear to me whether the 77 in the
calculation above is 77 UK-based websites, or 77 reports made to ISPs
with Usenet services asking them to delete specific articles.

If the IWF are including Usenet in these "UK" figures, then there's no
real comparison between what was happening in 1997 and what is happening
in 2008 -- claims of reductions in activity would be very misleading.

- -- 
richard                                              Richard Clayton

They that can give up essential liberty to obtain a little temporary
safety deserve neither liberty nor safety.         Benjamin Franklin

-----BEGIN PGP SIGNATURE-----
Version: PGPsdk version 1.7.1

iQA/AwUBSUU0OZoAxkTY1oPiEQJSNwCgmM0r/o2EMhCJPDzS+uorLK5Os/oAoL2Z
BZmDOeCau2bKR5B8w9Xrwk2a
=BE2x
-----END PGP SIGNATURE-----