Road Safety News

RoSPA and PHE to host free evaluation and data workshop

Monday 4th September 2017

A free workshop taking place in London next month will cover the basics of road safety evaluation and a newly available data set which can be used as part of the evaluation process.

Organised by the Royal Society for the Prevention of Accidents (RoSPA) and Public Health England (PHE), the one-day workshop will take place at the TfL offices in Southwark on 25 October, and will comprise two sessions.

The opening session will be presented by Nick Lloyd and Becky Needham, RoSPA’s road safety manager and research and evaluation officer respectively.

This session will include an overview of the evaluation process, the basic steps to evaluating a road safety intervention and the support that RoSPA can offer via the Road Safety Evaluation website and the e-valu-it toolkit.

The second session will be hosted by Susan Koffler-Sluijter, PHE’s regional lead for South East England. This session will explore the newly available PHE ‘fingertips data’ which can be used in conjunction with STATS 19 data as part of the evaluation process, or to benchmark in comparison to other local authorities.

To find out more or register interest in attending the workshop, email Becky Needham confirming your name, organisation, job title and contact details.

Places on the free workshop are limited and will be allocated by date of registration of interest.

Category: Events.



Comment on this story
Report a reader comment

What's your view - comment on this story:

I confirm that I have read and accept the moderation policy and house rules relating to comments posted on this website.
Your comment:
Your name and location:
Your email:

It's great, Nick, that you say "keep an open mind and look at research" but the quality of research in road safety is shockingly poor, in my opinion.

Just take the effect of site-selection (often called RTM). Official reports have estimated that it might be huge. So large, in fact, that it might be larger that all other effects combined. Yet not a single official report has managed to exclude it's effect from their results.

If they cannot measure the largest effect, then they cannot measure the much smaller effect of the intervention. IOW, the effect of the intervention remains officially unknown and that remains true for all site-based interventions.

Official reports have proved the need for scientific trials and I have been calling for scientific trials for years. As yet, though, I appear to be a lone voice in an ocean of opinions, pressure groups and politics.
dave finney

Agree (3) | Disagree (3)

I have a little bit more to add though Nick - it is important after all. Why research when one can see it first-hand for oneself? 'Being in Heswall or Birkenhead, or Toxteth, or Chelsea or Hackney - trust me you will see the same road user behaviour - good and bad. That's 'research' if you like. As someone said earlier ".. huge sums of money are being spent... without proof of effectiveness.." and trying to read patterns and trends into stats which are not reliable in the first place is, I think, wasteful and ultimately non-productive.
Hugh Jones, Cheshire

Agree (1) | Disagree (11)

I have nothing more to add on this matter Hugh - except to say keep an open mind and look at research. There are several measures that are used when looking at deprivation and soci-demographic groupings and I am not sure that "being in Heswall" is one of them.
Nick, Lancashire

Agree (9) | Disagree (1)

Depends what you mean by being over-represented in the stats Nick - do you mean as innocent victims? e.g. belted-up passenger in a car; being in the wrong place at the wrong time e.g. standing at a bus stop when a drunk or speeder mounts the footway? or do you mean the individual who caused the collision and most at fault in each case i.e the perpetrator? To do that we would have to look at each collision, work out who was most at fault, who made mistakes, who was careless, who was reckless and determine whether those shortcomings were directly to their social background/demographic or not. How on earth would you do that?

Three other factors to bear in mind: we only have information on an estimated 20-25% if (that) of collisions so we can't say with any accuracy whether any particular group is over-represented in ALL collisions anyway; the information in STATS 19 is not necessarily accurate - it's fine for factual stuff like where and when and who, but not contributory factors and last but not least, chance and luck - the difference between a near miss and a KSI can be a second and inches in any area. Certain demographics may well be over-represented in the relatively few collisions we know about, but I don't believe it is anything more than an interesting statistic nor should anything necessarily be read into it.

I drove around affluent Heswall this afternoon and saw enough careless and reckless behaviour by all road user groups to keep the authorities and analysts going for weeks and no reason to think that would be any more or less than I would see in a deprived area - or anywhere in between - sociologically speaking.
Hugh Jones, Cheshire

Agree (3) | Disagree (7)

Try looking at this - very quick Google search
Nick, Lancashire

Agree (3) | Disagree (1)

Hugh please explain if you can why some demographic groups are over-represented in the stats.
Nick, Lancashire

Agree (4) | Disagree (1)

But the fallacy Nick, is in assuming that there MUST be some common factor amongst those from within those postcodes and demographics which somehow leads to involvement in collisions. Similarities in behaviour and attitudes yes, most definitely, but the behaviour which causes involvement in collisions is spread randomly throughout society. For example, a well-educated, middle class person could just as likely be speeding along a residential road in a poor area and hit a pedestrian, as could a poorly-educated, person from a deprived area doing the same thing in an affluent area. If you spend time in both types of areas, you will witness the same behaviour - both careful and careless.

Even better, observe road user behaviour to and from on the access roads and junctions serving a hospital - the one place where you will find the biggest mix of social class and backgrounds - and any careless or reckless behaviour will be spread evenly amongst individuals from all parts of society.

The motorway pile up example I gave is not one collision either - it is many, one after the other in close succession and each individual driver who did not avoid colliding could be from widely differing backgrounds, but who were possibly not paying attention or going to fast or too close or all three etc. - faults not peculiar to any one social group. The only patterns and trends are poor road user behaviour - equally spread throughout society.
Hugh Jones, Cheshire

Agree (1) | Disagree (7)

In overly simple terms Hugh, looking at one collision is not useful. Looking at many collisions over time can identify statistically significant patterns and trends in locations, demographies, behaviours, mechanical et al factors. These patterns and trends can then be used to design and target interventions. These should go in some way to reducing the likelihood of similar types of collisions happening in the future. It is not an exact science but methods are evolving in my opinion in the right way.
Nick, Lancashire

Agree (8) | Disagree (2)

Imagine a motorway pile up Nick, involving many, many vehicles, containing occupants from a diverse range of socio-demographic backgrounds and postcodes, some of whom were at fault, some of whom were innocent victims. Is that not random?

How could any analysis conclude anything meaningful from that, even if those circumstances could somehow be extracted accurately enough from Stats 19 anyway? What about the fact that no information is held at all about the majority of collisions and those individuals affected, that occur on UK roads?
Hugh Jones, Cheshire

Agree (2) | Disagree (8)

Hugh – I am not perhaps the best person to explain in full how we use the home postcodes but the connections between the indices of deprivation and risk of becoming a casualty are well known and there are areas of deprivation rather than randomly spread people. We can use MAST which is provided by Road Safety Analysis (sorry for the obvious plug) which allows us to find much about where casualties/non-injured participants in RTCs live and also their social-demographic characteristics. This enables us to design, target and deliver interventions much more efficiently than has been possible in the past.

Unfortunately I do not have the time to fully go into this but I can assure you Hugh that the chances of being involved in collisions (whether culprits or victims in your words) is not random as you say it is.
Nick, Lancashire

Agree (8) | Disagree (2)

I don't get Nick's reference to home postcodes of those involved in recorded collisions and the significance thereof. Is it the postcodes of the culprits or the victims? Either way, I can't see how the likelihood (or not) of being involved in a collision can relate to someone's home address. In my experience, 'collision likelihood' is scattered amongst the population randomly, geographically and demographically. I alluded elsewhere on another thread, to the tendency to be too reliant on data and stats and end up barking up the wrong tree.
Hugh Jones, Cheshire

Agree (1) | Disagree (11)

A couple of points to raise and questions relating to previous comments in this thread.
"Anon" makes a good point about the need to consider "resident risk". My Team is based within Public Health and we are very well aware that we need to consider the home postcodes of those involved in police recorded collisions to help target some of our education/publicity type interventions more effectively. This is a different approach to the engineering and enforcement interventions which are generally targeted towards where the collisions occur. There is/are lots of data out there and understanding the differences and choosing the appropriate datasets is vital. Initiatives such as the "Joining the Dots" type conferences can only help in spreading good practice. We have an in-house session on using and understanding public health data coming up at which I will be looking at alternative datasets which can be added to our use of Stats 19 data. Continual improvement is what it is all about. I am sure I am not alone in wanting to improve the way road safety is carried out?

Dave Finney – When planning road safety engineering type measures they are generally designed in my experience by qualified Safety/Highways Engineers and frequently go through Safety Audits/public consultations. This should remove dangerous combinations of physical measures when features proposed to improve safety are implemented. When an education/publicity type intervention is planned the move now is to consider behaviour change type models and techniques with clearly defined aims/objections for the projects. This leads to interventions being based on proven techniques which could be expected to produce positive results. I'm surmising therefore that you are referring to other types of interventions when you refer to road safety analysis making so many mistakes leading to many people potentially being killed or seriously injured. Perhaps you could point me towards examples of what you consider to have been at risk of creating more casualties. It may help me and any others understand better where you are coming from and how these trials could be included in our planning? Do we need to do a "scientific trial" of giving leaflets out to parents suggesting ways in which they may improve safety for their children when walking to school?
Nick, Lancashire

Agree (5) | Disagree (2)

RoSPA state they would have to question running any expensive ad campaign "without collecting scientific evidence of its effectiveness" and the only way to provide that "scientific evidence" is to run scientific trials.

Scientific trials use the same selection process and the same implementation as now, we just add a randomising process between the two.

Scientific trials are easy to run, they're cheap and they produce accurate results so here's the question: Why is every road safety authority, and all the pressure groups (both for and against), so desperate to ensure scientific trials are never run?
dave finney

Agree (2) | Disagree (9)

We would all welcome evaluation and research that will tell us which Road Safety education, training and publicity programmes work and why. The structure of Road Safety in England – through a small team or, nowadays, often one or two individuals within each Local Highway Authority means that it is almost impossible to develop and evaluate programmes to scientific or clinical standards and scales. There is no national provider of programmes beyond the (very welcome and much used) educational resources from the DfT.

These studies are absolutely not cheap and, for most small scale programmes, they would take up so large a proportion of the available budget and staff time that they are not affordable. This is why often they do not happen.

Where evaluation has been undertaken it has been rejected by the national scale studies as being either too small in scale or a single study not repeated elsewhere – neither of which is achievable under the present structure.

So local Road Safety teams are castigated for not providing evidence to a level and format that they are actually unable to achieve. We have many programmes – some of which are surely effective and others less so and we do evaluate them so far as we can. But we are in danger of losing them not because they don’t work but because they cannot be shown to work to a model that was designed for different purposes.

We do not demand that geography or history syllabus is assessed against a clinical or scientific double blind study for its future relevance, just that students pass the exams involved. Why should we demand a clinical or scientific level of evidence for an educational programme? Surely an educational model would be more relevant and achievable?
Honor Byford, North Yorkshire

Agree (12) | Disagree (2)

Unfortunatly there is a problem with the Public Health England indicators for Road Safety in that they compare the number of people injured in a county to the resident population, when many people injured in a county arn't residents. Hopefully this issue will be discussed in this conference and Public Health England will consider other indicators that are more accurate in future. Currently many areas cannot influence this indicator in any way unless they build a huge amount of new houses, without the collisions increasing, and yet are being ranked and judged on it.

Agree (12) | Disagree (1)

re-para 5: I take it they are aware that Stats 19 is neither accurate nor comprehensive enough to be a reliable source of causation factors, if that is what it is being used for.
Hugh Jones, Cheshire

Agree (5) | Disagree (4)

Road safety analysis is in crisis. So many mistakes are being made that huge sums of public money is being spent without proof of effectiveness and many people may be being killed or seriously injured as a result.

RoSPA and all other "stake-holders" need to insist upon scientific trials for all road safety interventions. Only when it can be demonstrated that a scientific trial is not possible should any intervention be permitted to go ahead without.

The solution is cheap, it is accurate and it's obvious. The solution is to run scientific trials.
dave finney

Agree (5) | Disagree (7)