John

I had a look for the report you described, but could only come up with:
NSQR Series Report No.49 - Review of Road Traffic and Road Length Statistics
Is that the one you are referring to?

A year or two back the DfT ran a working group on cycle monitoring, for
which Sustrans did some consultancy work on a couple of case studies, but
the officials I talked to described the outcomes of the WP as
"inconclusive".

I had a detailed look at the Sustrans report, particularly as one of the
case studies was in Surrey, and I have to say that the methodology used can
best described as "limited".  A quick look at a number of the counter
locations revealed systematic counting errors of up to 30%, arising from,
all to often, poor positioning of the counters.

As regards automatic counters, when I last looked at the DfT information
"How the National Road Traffic Estimates are made", in 2006, there are only
190 automatic counters run by the DfT.  The bulk of the counting is still
done manually and there are 5100 manual counting sites on major roads, plus
some sites on minor roads, used on a less frequent basis.

The basic problem is not particularly that the estimating method is designed
specifically for motor vehicles (manual counting does, to an extent,
mitigate against the insensitivity of automatic counters to cycles) but that
the surveying methodology is optimized to satisfy the National Traffic
Model, so survey points are located on primary and high usage secondary
links on the model.  Furthermore, counting locations are chosen to avoid
congested locations, i.e. high density urban sites where cycle usage might
be higher, so would hardly be representative of a national cycle "network of
use" (in the same way as the Sustrans National Cycle Network).

To ensure a "true" estimate of overall cycle usage, locations have to be
chosen to be statistically "representative" and, ideally, "statistically
consistent" with the general traffic estimating methodology (otherwise it
undermines the objectivity of overall traffic comparisons).

The other area of confusion is that cycling figures seem to be derived both
from the traffic counts and the National Traffic Survey.  The recent quality
review acknowledges there has been a creeping decline on the robustness of
this survey, and recommends the number of participant logs be increased to
15,000 from the present 9,500 completed travel diaries.

I'm currently working on a paper describing an approach that might provide
more objective and consistent cycle counting within the frame of both
national and local traffic counts.  The paper I presented at the symposium
represented my attempts at estimating how large the possible error in the
current methodology might be...as I realized it was a bit presumptuous to
just say that the current usage estimate are plain wrong!  The stunned
silence at the end somewhat surprised me...I'd thought that a stochastic
approach would be more controversial than it appears it was (or was it
something else).

I would interested if a discussion group were got together to discuss the
problems of cycle monitoring, as all attempts to resolve the issues seem to
have failed.

Regards
John meudell




-----Original Message-----
From: Cycling and Society Research Group discussion list
[mailto:[log in to unmask]] On Behalf Of Parkin, John
Sent: 04 October 2007 14:38
To: [log in to unmask]
Subject: Re: Retracted DfT cycling data?

Paul,
 
You might be referring to the national road traffic estimates (NRTE). These
are published quarterly, with a yearly summary. Cycle data as part of these
estimates is no longer published on a quarterly basis and this is because
the level of variation in the data from one quarter to the next is such that
it has called into question their veracity. Pooling the data for a year,
however, continues.
 
I have been in discussion with the people at DfT about this along with Andy
Cope of Sustrans. The idea is to try to encourage the DfT to add to the
(rather low) number of 50 counters they have for tracking bicycles using the
300 or so counters used by Local authorities and reported to Sustrans on a
regular basis as part of National Cycle Network Monitoring. (Note: there are
thousands of counters for motor traffic, sorry exact figures not to hand).
 
They have recently done a data quality review of NRTE, and this report is on
the DfT web pages.
 
 
Dr John Parkin
Reader in Transport Engineering and Planning
Department of the Built Environment
The University of Bolton
Deane Road, Bolton, BL3 5AB, UK
Tel 01204 903027 Fax 01204 399074 mob 07903 523 017
www.bolton.ac.uk/staff/jp10

________________________________

From: Cycling and Society Research Group discussion list on behalf of Paul
Rosen
Sent: Thu 04/10/2007 1:26 PM
To: [log in to unmask]
Subject: Retracted DfT cycling data?



I have an idea that some DfT cycling figures were retracted recently.
Anybody have an idea what these were and when it happened ... web links???

Thanks,

Paul