DfT Stats - 2009
wrote in message ...
In article
,
PhilO wrote:
Judith,
Are you deliberately trying to deceive?
No - just troll.
TSGB0107 shows pedal cycle Killed decreased from 24 in 2008 to 21 in
2009 - a 12.5% decrease
KSI has increased from 541 to 547 - a 1% increase.
Do you think the 1% change statistically significant?
Do you think that the 12.5% is, either?
Why did you fail to report the much larger decrease in cyclist Killed?
Numerically, it's only 3 versus 6. Statistically, neither are
significant at even the 20% level.
And Judith, how did you miss this statistic:
Killed per billion passenger km: Pedal cycle 21, Pedestrian 26.
Well, I could explain why it is deceptive - as are the raw numbers.
The KSI rate per trip might well show a statistically significant
increase, but none of the figures are reliable and meaningful
enough to draw definite conclusions from.
So, Judith, doesn't that mean you are more likely to die walking a
mile than cycling a mile?
No. Using population statistics (ESPECIALLY ones calculated in
the way that those have been) doesn't indicate individual risk at
all well.
I suggest that you ignore JMS's trolling - and don't post equally
bogus figures in return!
Regards,
Nick Maclaren.
================
Path: nntp-feed.chiark.greenend.org.uk
Says it all really. A server used by psycholists.
|