I've been gathering data for 2003 fatalities from official web sites and newspaper reports. With 37% of the total fatality figures available we have an overall 6.2% rise. The same basket of 19 counties showed an insignificant change between 2001 and 2002.
If the changes in the 19 counties are typical, and if they don't manage to fiddle the figures we WILL see national fatalities of 3,625 for 2003.
You can download the latest spreadsheet from:
http://www.safespeed.org.uk/2003.xls
There are potential reasons for thinking the figures may not be typical:
1) They do not include any large urban areas.
2) There might be a bias against releasing bad figures earlier than absolutely necessary.
And reasons for assuming the figures may well be typical:
1) 2001 to 2002 changes for the same basket of counties were well aligned with national figures (i.e. no change)
2) We have no solid basis for thinking that they are NOT typical, so it's best to assume they are.
So I voted for 3,600+.