Speed Management Strategy Taskforce

Webmaster's note: the text below is a committee briefing document supplied by an official of Oxfordshire County Council. It provides an insight into the widespread introduction of 50 mph limits into that county. Chart E contains a display error - The lines for average accidents before and after are drawn to meet in the wrong month. The actual values for before and after are correct

a) National and international research into the relationship between speeds and accidents

The Transport Research Laboratory (TRL) has carried out extensive research in this area; two reports summarising findings relevant to the remit of the taskforce are:

TRL Report TRL511: The relationship between speed and accidents on rural single-carriageway roads (2002).

The findings of both reports confirm the close relationship between average speeds and accident frequency (although – as discussed at the last meeting of the taskforce - there is certainly evidence that vehicles travelling appreciably slower or faster than the average for a particular traffic stream have a higher than average accident involvement rate compared to those vehicles travelling close to the average speed, this in no way changes the overall relationship between average speed and accident frequency).

Monitoring of speed management measures (most of which have been in built up areas) in Oxfordshire appear to confirm the broad relationship that a 1mph reduction in average speed is associated with a 5% reduction in accident frequency.

The current lively debate on the application of a speed management strategy for Oxfordshire’s rural roads quite rightly raises the question of whether a similar relationship holds for rural roads, and whether speed limits are an effective way of reducing average speeds.

The following charts suggest that both the village speed limit project, and the 50mph speed limits introduced to date have indeed been effective in reducing accident frequencies.

This chart shows the monthly accident totals in the c150 new 30mph limits implemented in South Oxfordshire, mainly in the latter part of 1999. It also shows the average monthly accident frequency before and after implementation.

Chart B: Monthly accidents – control area (Cherwell District) for assessment of effect of village speed limits.

This chart shows the monthly accident totals in the identical period as covered in Chart A but for the Cherwell district (the last district to be covered by the village speed limit project, with implementation being carried out in the very latter part of the monitoring period); as can be seen although there was a very modest reduction in accident frequency in the after period, this was very small compared to that observed at the treated sites in South Oxfordshire.

It does appear that the ‘after’ accident frequency is markedly lower than the before period at the treated sites (even allowing for the inevitable substantial monthly variation in accident numbers).

This finding proved has proved something of an embarrassment in that even limits which have been applied well outside normal practice (and counter to both the advice of the Road Safety Engineering team) have proved to date surprisingly effective in reducing accidents, despite anecdotal accounts of such limits leading to driver frustration and poor behaviour.

A sample of before and after speed measurements indicated that the new limits typically reduced average speeds by around 3-5 mph (there was however considerable variation between sites).

There is no evidence for this from the accident data currently available that the new limits have led to any worsening of safety in the previously existing 30mph limits in the area.

Chart C: Oxfordshire 50mph limits with at least 30 months after data (13 limits)

Chart D: as above, but including only limits judged to have appreciable supporting measures such as enhanced signing, speed camera enforcement etc – (7 limits)

These charts suggest that the 50 limits have been effective in reducing accident frequency; this is particularly so for limits where more supporting measures have been provided. Although the reductions are less marked than those with shorter assessment periods, it may well be that other external factors (e.g. the markedly increased amount of enforcement now in place due to the greater use of speed cameras) has led to newer limits being more effective than older ones.

Reduction in accident frequency observed after 30 months:

‘Older’ limits: 20%

‘New’ limits: 48%

Although statistically speaking these results should be regarded with some caution, a measure of support for this being a real effect is given by the surprising (and to date maintained) accident records observed in the village speed limits in South Oxfordshire.

It should be noted that lengths of these roads where other significant safety engineering measures have been carried out in the assessment period have been excluded so as to try to ensure that the assessment is not biased by including non-relevant measures.

This chart provides control data (covering \ length of rural single carriageway of broadly comparable alignment to those included in Chart A for an equivalent length period) to help assess the significance of charts C and D. As can be seen, there in contrast to the 50mph limit. Again, although the same qualifications as listed above must be applied to this data, this control data does lend weight to the suggestion that the 50mph limits have been effective in reducing accident frequencies.

Unfortunately very limited before and after speed data is available for the rural 50mph limits and so it is impossible to relate changes in average speeds to the apparent reductions in accident frequency.

The relatively limited scope of the monitoring data available for 50mph limits on rural single carriageways inevitably means that there is a higher degree of uncertainty over their real effects on safety, especially given the inherent random element in the accident data. Having said this, it would be hard to avoid the conclusion that accidents are reduced – apparently in several cases substantially so - through the lower speed limit. This finding would be very much in accord with the national research into the relationship between average speeds and accident frequency.

The current 50mph limits are generally on intermediate / reasonable standard roads where the 85^th percentile speeds prior to the introduction of the limit can be expected to have been appreciably above 50mph. The fact that these limits seem to have a good level of acceptance (very few complaints have been received from members of the public on the existing limits) is perhaps a pointer to the fact that further limits on similar quality roads may not be as contentious as the very vocal (but possibly very unrepresentative) response to the current proposals may currently suggest.*

However, it should be noted that the current proposals include roads of a lower standard than that of most of the existing limits. Although therefore there must be greater uncertainty as to the effect of the limits on these roads, clearly it will be impossible to ever determine this without a reasonable scale trial.

*There may well be an opportunity to test public attitudes to further 50mph limits through a questionnaire survey on general attitudes to speed and speed enforcement to be carry out in mid 2004.