Official Statistics

Road safety factors: initial analysis

Published 30 May 2024

Introduction

The department’s published road casualty statistics are based on data reported by police forces using the STATS19 system, which is reviewed around every 5 to 10 years.  The most recent review recommendations are currently being rolled out, including replacing the currently collected contributory factors (CFs) with a new system of road safety factors (RSFs).

In time, this change will have implications for the published statistics and this report presents details of the change, an initial analysis of the potential impacts on reporting of factors involved in injury collisions and sets out the department’s planned approach to the transition.

Any comments from users of the statistics are welcome, via our short user survey or by email. 

Background

Contributory factors

CFs have been collected as part of the STATS19 system from 2005. While they do not direct identify causation, they provide details of up to 6 factors which, in the opinion of the reporting police officer, may have contributed to the collision occurring.  Further details can be found in our contributory factors guide and the published figures in data tables RAS0701 to RAS0706. 

STATS19 review recommendation

The most recent STATS19 review started in 2018, with a final review report published in 2021.  This included a recommendation that CFs be replaced with a new system of road safety factors.  These were designed to focus on recording factors related to areas where action can be taken to improve road safety, and a reduction in the list of potential factors from 79 to 36 factors, with new codes.  The recommendation was strongly supported by stakeholders.   

The review recommendations are currently being implemented.  Police forces using the Collision Reporting and Sharing (CRASH) system began using the new specification from November 2023. Other forces expected to adopt it by end 2024. 

Collection of data on RSFs

Some, but not all, of the police forces using the CRASH system are now collecting data on RSFs.  In other cases, data are still being recorded as CFs with a conversion applied in the system. 

This means that a small amount of data (provisionally, around 2% of the total) for 2023 collisions has been recorded with RSFs.  However this proportion will be higher for 2024, and in future years. 

As a summary of the anticipated transition from CF to RSF reporting:

• up to 2022, only have CFs
• in 2023, mostly CFs with a few RSFs
• in 2024, still many CFs but more RSFs
• in 2025, expecting almost entirely RSFs

Therefore consideration of the reporting of the new RSFs will be required, if not for 2023 data then certainly for 2024.

Overview of road safety factors

The current CFs are grouped into 9 distinct categories, covering road environment, vehicle defects, driver or rider injudicious action, driver or rider error or reaction, driver or rider impairment or distraction, driver or rider behaviour or inexperience, driver or rider vision affected, pedestrian factors and special codes.

RSFs have been designed to align to the safe system pillars, and have been grouped into the following 6 categories (with the ‘safer road users’ component of the safe system split across the first 3 – B, D and P).

Table 1: Road safety factors categories

RSF category	RSF category name
B	Behaviour or inexperience
D	Distraction or impairment
P	Non-motorised road users or PPT (pedestrian, cyclist, horse rider or powered personal transporter)
R	Road
S	Speed related
V	Vehicles

Within each of these categories are a number of individual factors, which are outlined in a spreadsheet data table. 

In some cases, these are a direct equivalent of an existing CF (for example, R2 ‘road surface slippery due to weather’ is a direct equivalent of CF 103).    

In other cases, several CFs have been combined into one RSF, for example where there is little analytical benefit in having separate factors to ease the burden on reporting police officers.  For example, CFs 405 (driver or rider failed to look properly) and 406 (driver or rider failed to judge other person’s path or speed) have been grouped together in B4 (ineffective observation).  

In the CF framework, there were often different factors for equivalent behaviours for drivers (or riders) and pedestrians, for example CF 802 (pedestrian failed to look properly) is the equivalent of 405 (driver or rider failed to look properly).  With RSFs, these have been combined so that B4 (ineffective observation) applies to all road user types, and the non-motorised road user category includes only those factors specific to pedestrians, cyclists horse riders or users of powered personal transporters.

Mapping from contributory factors

There is a many to one mapping between CFs and RSFs.  This means that it is possible to recode CFs as RSFs (with some limitations), but not always possible to recode from data captured as RSFs back to CFs.

Therefore, data covering past and future years can only sensibly by expressed in terms of RSFs, by mapping from CFs where data was recorded as CFs and combining with data recorded directly as RSFs (which will recorded in future).

To map CFs to RSFs requires a well-defined mapping, as shown in our mapping table.  However, even in this case, there are a number of caveats to note.  For example:

• there are 16 current CFs which are not mapped to any RSF (such as defective traffic signals)
• mapping factors coded as CFs to RSFs may not reflect what police officers may assign directly as RSFs

Nonetheless, as only RSFs will be recorded in the long term, and are considered to represent an improvement over the current CFs, we intend to begin reporting using the RSFs framework alongside CFs for 2023 data, and then instead of CFs for 2024 data onwards.  

Illustrative analysis

To illustrate the effect of mapping between CFs and RSFs, we have produced data table RAS0701 converted to RSFs.  The following highlights some of the changes, based on fatal collisions (where CF data is most complete).

Tables showing a time series of the last 10 years data can be found in a spreadsheet data table.

Mapping of factors for fatal collisions

The diagram below illustrates the mapping from CFs to RSFs for fatal collisions in 2022.  The size of the flows is proportional to the number of fatal collisions.

This shows, based on this one year of data for fatal collisions:

• in the CF framework, the driver error or reaction category is largest – this largely maps into the ‘behaviour or inexperience’ RSF category, but some factors (such as swerving or loss of control) map into speed
• the majority of the CFs currently recorded as ‘behaviour or inexperience’ also map to speed RSFs (these include driver careless, reckless or in a hurry, and aggressive driving)
• when mapped to RSFs, speed becomes the single biggest category, as it gains factors previously included elsewhere
• some of the categories map fairly directly across, for example impairment and distraction (though the RSFs now include equivalent pedestrian factors)
• road and vehicle factors remain relatively rarely assigned

Figure 1: Mapping from CFs to RSFs, fatal collisions, Great Britain 2022

It should be noted that there are some collisions where RSFs cannot be mapped (as shown in the mapping table).  Where possible, these are at least mapped to the category level.  For example, CF 109 (animal or object in carriageway) cannot be mapped to an individual RSF, but can sensibly be considered to be within the road RSF category.          

Table 2: CFs which cannot be mapped to individual RSFs, fatal collisions, Great Britain 2022

Contributory factor code	Contributory factor label	Road safety factor section	Number
104	Inadequate or masked signs or road markings	Road	4
105	Defective traffic signals	Road	1
106	Traffic calming (eg. road humps, chicane)	Road	1
107	Temporary road layout (eg. contraflow)	Road	3
108	Road layout (eg. bend, hill, narrow road)	Road	30
109	Animal or object in carriageway	Road	10
205	Defective or missing mirrors	Vehicles	0
309	Vehicle travelling along pavement	Behaviour or inexperience	8
402	Junction restart (moving off at junction)	Behaviour or inexperience	6
404	Failed to signal or misleading signal	Behaviour or inexperience	3
709	Visor or windscreen dirty, scratched or frosted etc.	Vehicles	2
801	Crossing road masked by stationary or parked vehicle	Non-motorised road users	10
901	Stolen vehicle	Speed	9
903	Emergency vehicle on a call	Distraction or impairment	3
999	Other	Not coded	82

Proportion of fatal collisions with factors in different categories

The charts below show the proportion of fatal collisions with at least one factor recorded, for fatal collisions in 2022, based on the CFs recorded, and the equivalent RSFs when mapped across from CFs. 

This again illustrates that, when grouped into the RSF categories, speed becomes the most commonly assigned category in fatal collisions.  This reflects the inclusion of factors such as ‘loss of control’ and ‘aggressive driving’ within this category.

Chart 1: Proportion of fatal collisions with RSFs assigned, by RSF section, Great Britain 2022

Road safety factor section	Percentage with factor
S Speed	56
B Behaviour or inexperience	45
D Distraction or impairment	34
R Road	11
P Non-motorised road users	8
X Not coded	6
V Vehicles	4

Chart 2: Proportion of fatal collisions with CFs assigned, by CF section, Great Britain 2022

Contributory factor section	Percentage with factor
400 Driver/Rider error or reaction	56
300 Injudicious action	29
500 Impairment or distraction	29
600 Behaviour or inexperience	28
800 Pedestrian only (casualty or uninjured)	15
100 Road environment contributed	7
900 Special Codes	7
700 Vision affected by external factors	6
200 Vehicle defects	2

Most common individual factors

When looking at the most commonly assigned individual factors, the factors are broadly similar, though some are categorised differently.  As RSFs have fewer factors, and in many cases group several CFs, the top few factors are assigned more often.  However, this may in part be an artefact of the mapping process, and could change when RSFs are recorded directly.  

Chart 3: Top 10 most commonly assigned RSFs for fatal collisions, Great Britain 2022

Road safety factor	Percentage with factor
B4 Ineffective observation by either the driver, rider or pedestrian	32
S2 Driver/rider travelling too fast for conditions (including loss of control or swerving)	30
S4 Driver/rider being aggressive, dangerous or reckless	26
S1 Driver/rider exceeding speed limit	19
D1 Affected by alcohol	13
B3 Driver / rider overshot junction or poor turn / manoeuvre	12
D5 Illness or disability	10
D2 Affected by drugs	7
D7 Distraction to driver/rider from inside/outside or on vehicle	6
B1 Driver / rider illegal turn / direction of travel or failed to comply with traffic sign / signal	4

Chart 4: Top 10 most commonly assigned CFs for fatal collisions, Great Britain 2022

Contributory factor	Percentage with factor
410 Loss of control	24
405 Driver/Rider failed to look properly	23
602 Driver/Rider careless, reckless or in a hurry	21
306 Exceeding speed limit	19
406 Driver/Rider failed to judge other person`s path or speed	11
403 Poor turn or manoeuvre	11
501 Driver/Rider impaired by alcohol	10
307 Travelling too fast for conditions	8
505 Driver/Rider illness or disability, mental or physical	8
601 Aggressive driving	8

Directly recorded RSF data

As is noted above, in the data for 2023, only a very small proportion of collisions have RSFs reported directly, for some (but not all) of the forces using the CRASH system, during November and December.  Therefore, it remains too early to draw any firm conclusions about how the statistics may differ when based on RSFs recorded directly, rather than based on mapping from CFs.

An initial analysis suggests that there are around 2,300 collisions in 2023 where the new specification has been used in full.  On average, 2.6 RSFs have been reported per collision, which compares with around 2 CFs per collision, for comparable collisions using the CF system.

The STATS19 review recommended that each collision have 3 to 6 factors assigned, and it appears that around two-thirds of collisions with RSFs have 3 or more factors.

At this point, limited data exists to explore the impact of recording of RSFs, and we intend to produce a more comprehensive analysis when more data is available.  

Summary and questions for consideration

Summary

This initial analysis, mapping data for CFs to the new RSFs , shows:

• at national level, the broad patterns shown by CFs and RSFs are unsurprising fairly similar, with most factors related to road user behaviour (rather than roads or vehicles)
• however, RSFs align better to the safe system pillars – and are clearer and more actionable
• in particular, RSFs highlight the impact of speed as a factor in (fatal) collisions more clearly than existing CFs
• while some detail is lost in the mapping, this is typically related to factors that are either rarely recorded, or not useful for analysis

However, currently a small proportion (around 2%) of the 2023 data has been recorded directly as RSFs.  Further exploration of this is needed, particularly of the impacts of forces moving to directly recording RSFs (rather than mapping from CFs), and this will require more data to analyse.

Questions for further consideration

In addition to exploring the impact on the data of RSFs which are directly recorded, which may impact on trends over time, there are a number of questions which require further consideration, and where feedback from users of the statistics is welcome.

Reporting of speed-related factors.  Currently, the CFs 306 (exceeding the speed limit) and 307 (travelling too fast for conditions) are considered as ‘speed related’ factors – as presented in table RAS0704. As is illustrated above, in moving to the RSF approach, the speed category is wider, including factors which were previously classified elsewhere.  This leaves a question as whether the narrower interpretation of ‘speed factors’ has value, or whether the RSF category for ‘speed’ effectively replaces this.  This will have implications for the reporting related to contribution of speed in fatal collisions.       

Mapping between CFs and RSFs.  Alongside this analysis, we have published a mapping table which shows a proposed mapping between CFs and RSFs.  This is not definitive, and other mappings may be possible.  Any comments or proposed changes relating to this mapping would be welcome, with a view to agreeing on a standard approach for wider use if possible.  

Assignment of factors not mapped to individual RSFs to RSF categories.  As noted above, and illustrated in the mapping table, there are 16 current CFs which not mapped across to new RSFs.  In these cases, we intend to allocate to the wider RSF section, and flag as a CF which cannot be directly mapped in any published tables.

Incorporation of information recorded in the collision circumstances.  In developing RSFs, an effort was made to reduce the duplication of information collected within the STATS19 system.  For example, carriageway hazards were captured as part of the collision record, as well as having a CF.  In this initial mapping, we have not considered whether or how information from elsewhere in STATS19 might be combined with RSFs. 

Proposed next steps

For the 2023 statistics (scheduled for publication in September 2024) we intend to: 

• produce data tables showing RSFs (mapping from CFs) alongside those for CFs (which will be slightly incomplete for 2023)
• provide an update on progress in collection of RSFs data

For the 2024 statistics (September 2025) and beyond:

• we will update RSFs with more (but not all) data recorded directly by forces
• CF data is unlikely to have comprehensive coverage beyond 2023, and may not be updated (but existing historic tables retained)

Any feedback on this proposed approach, comments or concerns are welcome and we would be very happy to discuss our approach further with anyone interested.

Acknowledgements

The department is grateful to the members of the Standing Committee for Road Injury Collision Statistics (SCRICS) for their support in developing the new road safety factors, in particular Bruce Walton from Agilysis for suggestions regarding coding of unmapped factors.

Instructions for printing and saving

Depending on which browser you use and the type of device you use (such as a mobile or laptop) these instructions may vary.

Tablets and mobile devices normally have the option to “find in text” and “print or save” in their sharing or quick options menu of their browser, but this will vary by device model.

How to search

Select Ctrl and F on a Windows laptop or Command and F on a Mac

This will open a search box in the top right-hand corner of the page. Type the word you are looking for in the search bar and press enter.

Your browser will highlight the word, usually in yellow, wherever it appears on the page. Press enter to move to the next place it appears.

Contact details

Road safety statistics

Email roadacc.stats@dft.gov.uk