Road Safety Analysis has announced the inclusion of ‘contributory factors’ (CFs) in MAST Online, in a new dataset called ‘MAST Professional’ which is only available to those working in the public sector.
MAST Online is a web based data analysis tool which is designed “to revolutionise information use in road safety”. It provides national crash data for user-customised in-depth analysis and also offers unique insight into people involved in crashes with the help of Mosaic Public Sector, the market leading socio-demographic database from Experian.
Road Safety Analysis has been working closely with the DfT to gain permission to include CFs in MAST.
Access to MAST Professional will be restricted to MAST users in the public sector who can normally access CFs locally. In order to make the data as robust as possible it is pre-analysed to only include crashes attended by a police officer, where at least one contributory factor was recorded.
Tanya Fosdick, senior analyst at Road Safety Analysis, said: “The inclusion of CFs means that MAST users will now be able to work directly with national CF data for the first time, using the large number of records for trend analysis.
“The creation of CF dimensions in relation to crashes, vehicles and casualties provides users with greater flexibility and insight, coupled with the power to undertake socio-demographic profiling using Mosaic Public Sector.
“MAST users will now be able to use the system to fully explore the STATS19 dataset to answer the ‘who’, ‘what’, when’, ‘where’ and now ‘how’ questions about why crashes occurred and who was involved.”
For more information contact Tanya Fosdick on 07795 385770.
Being data-led is from what I understand the way to direct finite resources to where they are needed most and can be of most benefit. That may be towards distinct geographic locations such as individual junctions or towards groups of population that are shown to be at more risk than others.
On top of that, when communicating messages to the specific groups which have been identified it is more efficient to target the messages through relevant media streams.
The latter two instances are where MAST adds to what we have in-house. I am sure that Local Authority staff have the skills to analyse the Stas 19 data but what we are often short of are the time and resources to do so. It seems more efficient to me that the work is carried out once and then the cost of that analysis is shared between subscribers. Attending the MAST User Groups I have seen examples of where MAST has been used to produce campaigns/initiatives where previously it would not have been practicable to have done so.
I was taught “many years ago” that collisions are “rare, random, multi-factor events” but the more we investigate population risk and see clusters at sites/routes the more I am led to think they are not perhaps as “random” as the lotto numbers are, especially in regards to demogrpahic features.
MAST is not the solution to data analysis but compliments our in-house systems and like all data should be used in context.
Nick Hughes – Lancashire
0
I’d like to be more enthusiastic about this, but I don’t believe any amount of analysis can help when you factor in such things as chance, luck, recklessness and carelessness. The stats are simply the evidence of these elements coming together somewhere, sometime.
Hugh Jones, Cheshire
0
Thank you both for your continuing thoughts.
@Hugh: Your question is a good one, and goes right to the heart of some of MAST’s most important features. This is rather a long post, even though it only scratches the surface. Please feel free to get in touch with me directly if you would like further information.
MAST facilitates risk profiling with key variables which go far beyond straightforward agegroup-and-gender analyses based on STATS19 alone.
The relevant variables, provided for every driver and casualty with a known postcode (which includes over 80% of the dataset) are:
* Detailed area of residence, by Small Area (MSOAs)
* Socio demographic profiling using the Mosaic Public Sector classification system, supported by extensive road safety specific documentation
* Index of Multiple Deprivation
* Rurality of home area (independent of crash location rurality)
* Distance from home at time of crash
Because this information is provided on a national basis, it is possible for authorities to see the whole picture of road risk to which their residents are exposed (not only within the authority’s borders). It is also possible to compare local trends to the regional or national picture, and also to compare with other area on the basis of both proximity and socio demographic similarity. This helps to ensure that data is set in its proper context and makes it easier to prioritise important trends, especially when working in multi agency partnerships.
@Duncan: Yes, there is a need for theoretical as well as practical progress. It is important that we recognise and share both best practice and worst pitfalls, and identify which metrics are of the most practical value for which purpose. Theoretical frameworks do not have to be handed down from on high though – we as a profession should be proactive in developing them. In order to support this process, we hold an annual MAST Users conference, free of charge to all MAST member organisations: this year, there will be lots of in depth and practical presentations and discussions about CFs, for instance. Making sensible and practical use of data is what this process is all about, and the more road safety stakeholders who contribute to that the better the outcomes will be. You might be surprised at how much has already been achieved.
Bruce Walton, Oxfordshire
0
Yes, there’s data which is clearly useful – essential even – and data which is of academic interest only and without wishing to sound too cynical, from what I can glean, MAST is taking data which might otherwise be of academic interest only and presenting it as being ‘must have data’ which could lead authorities in the wrong direction. I’m willing to admit to being wrong however, so perhaps some examples would help? Also, from the main news story “… offers unique insight into people involved in crashes..” How is this possible and what is revealed?
Hugh Jones, Cheshire
0
Lots of data, but no theory. Without a theoretical framework in which to determine which data is valid we end up trying to evaluate all of the data rather than that which actually leads to a solution or at least points the way to a solution.
Duncan MacKillop, Stratford on Avon
0
Hugh, you raise a very valid point – data can be dangerous in the hands of people who don’t know what they are doing. So our projects involve a range of people with specialist skills, and if we haven’t got the skills between the various partner organisations, we will find the skills elsewhere. Data is important but it always has to be in context. But it is important to be data led where possible.
Ruth Gore, Safer Roads Humber
0
It is certainly possible to make the mistake of reading significance into data which is not in fact there. Correlation is not causation, as they say. Anybody who works with statistics should be cautious about jumping to conclusions from them, particularly when working with small sample sizes or taking data out of context. Have you had some bad experiences of this kind of thing, Hugh?
However, just because it is possible to use data unwisely does not mean it is impossible to use it well, nor does it make the data itself a bad thing. As a profession, we should be intelligence led – which does not require becoming a slave to data, but does require objective measurement of what is going on. Data should be used sensibly, as a tool to gain insight into trends and priorities. Nobody here has suggested treating contributory factors “as gospel” – but opinions of attending officers should not be dismissed as entirely worthless either.
Bruce Walton, Oxfordshire
0
Ruth and Bruce: Your responses are appreciated and if the system has proved to be useful for some, then fine, I would not wish to detract from that. However, I feel road safety and accident prevention have become too data-led and there is a danger in reading into statistics some significance which isn’t warranted – particularly with contributory factors which should never be taken as gospel, not until all vehicles have black boxes and dash-cams anyway!
Hugh Jones, Cheshire
0
To answer the question “Why Mast?” you first have to understand who’s actually involved in road safety delivery. Whilst the LA still have the statutory responsibility they would be the first to admit that road safety is delivered by many agencies.
Today’s road safety world includes council road safety teams, highways teams and sustainable transport teams. It includes the police, the fire service, the public health teams, for profit and not for profit companies, charities, the voluntary sector, academics and universities to name but a few.
Mast allows all these agencies to look at collision data at a click of a button. Road safety isn’t just confined to tarmac and roads, it’s also about understanding the behaviour of road users. And this will be different depending if you’re interested in the old, the young, the car driver, the motorcyclist etc. Mast helps you understand what is happening.
Road safety isn’t confirmed to geographical boundaries. Road users may travel in one area and live in another area. The Yorkshire and Humber region has recognised this and have been working together for a number of years on projects that Mast has provided some of the data to build up the road safety picture.
The mosaic side of Mast is excellent as it draws you a picture of who is at most risk. This is used to direct information campaigns, so that they can be delivered to the right person in the right way.
In today’s environment of diminishing resources you need to ask yourselves why aren’t you using tools like Mast to help understand your casualty problem? And if you’re not working with any of the agencies mentioned above, you should be.
Lastly, let’s not forget that behind every one of these faceless numbers there has been a person injured. Someone’s loved one. And that’s why a lot of us do road safety.
Ruth Gore, Safer Roads Humber
0
Thanks everyone for some interesting points.
@Duncan: I agree that contributory factors are not “causes”, and that considerable care has to be taken when publishing such information or drawing conclusions from it. I hope that the information provided in MAST Professional will allow stakeholders to set this information in a national context and therefore find better ways to use it and more meaningful ways of interpreting it.
@Dave: Much of the raw data which lies behind MAST is publicly available and can be downloaded for free. The only reason information is redacted is data protection (such as exact ages, postcodes and CFs). Anyone can access MAST Core, it is available to any organisation or individual (including yourself) who may wish to subscribe. MAST is supplied at cost by a not for profit company, the ‘costs’ being the licenses, hosting, expertise and software required for providing online analysis and socio-demographic profiling. Local authorities could download and analyse public STATS19 data themselves of course – but the resources required would be considerable so in reality it’s far more efficient for them to subscribe to MAST. On your point about interventions, MAST is a trend analysis and strategic planning resource not a GIS tool, so it is not suitable for assessing individual sites.
@Hugh: The important point here is that no authority is an island. Authorities generally only have access to STATS19 information for roads within their borders. This means that (1) they cannot compare their networks meaningfully to trends elsewhere and (2) they do not have complete insight into the road safety risks to which their residents are exposed (given that nearly 40% of all casualties are outside their home authority when they suffer injury). Road safety is not just about roads, it is about people and behaviour. MAST makes it possible for authorities to look at the bigger picture and devise strategies accordingly. Of course analysis for its own sake would be pointless – but that is not what MAST is. I agree that CF data is complex and interpreting it is challenging, but its inclusion in MAST will hopefully help more stakeholders to understand it and use it well. Better that than to go to the considerable trouble of collecting it, only not to use it at all!
Bruce Walton, Oxfordshire
0
Bruce: I did actually get the point of MAST and I understand what it can do, but just wondered if there’s a need for road accident analysis of this sort and on this scale anyway? Is it not just a case of data being analysed because it can be?
I’m sure it’s a clever bit of kit and I don’t wish to detract from the efforts of those involved, but based on my own experience in this field, (obviously other LAs may be different) all the accident info we ever needed was already at out fingertips and could be analysed and filtered every which way and with all due respect, I don’t recall any occasion when we felt further info or analysis, such as that which could be provided by MAST for example would have been useful.
The collating and presenting in statistical form of contributory factors (which are subjective anyway) on a national basis, but without knowing anything about the accidents they MIGHT have contributed to, is surely of limited use to a LA?
Hugh Jones, Cheshire
0
I can see Hugh’s point, that the data is already available to LAs etc, but I can also see Bruce’s point that MAST might show relative data and trends in a way LAs might struggle to find (I don’t know, I am independent and therefore cannot access MAST). What I would like to see is an example of an intervention that had been demonstrated to have produced a road safety effect and which had been, and could only have been, identified using MAST. Has this happened yet?
There may be a concern that private companies can sell data collected using taxpayers money, yet that data is not available to the public, independent researchers or, it would appear, the authorities.
What safeguards exist to allow the public to ensure private companies are accountable?
Should private companies holding publicly funded data be subject to FOI laws?
Dave Finney, Slough
0
Maybe one day people might realise that there is a profound difference between a contributory or ‘causal factor’ and a ’cause’. At the moment contributory factors are used as satisficing labels that are cherry picked to support whatever argument is being put forward without even considering what the true cause might be. A statement like “speed is a significant contributory factor in road accidents” is often reported as “speed is a significant cause in road accidents” and this leads to a great deal of misunderstanding by those who don’t know the difference between the two.
Duncan MacKillop, Stratford on Avon
0
First of all, I should declare a vested interest: I am responsible for the design of the databases which lie behind MAST. I would not normally comment on a story with which I am personally connected, but I feel Hugh’s comment illustrates how easy it is to miss the point of MAST, in quite a number of different ways. I would like to point out why.
The most basic reason MAST exists is because it provides information which *cannot* be accessed in *any* other way by *anybody*. This includes police forces and highway authorities, which typically have access to local area data only, often inconsistently presented in disparate systems. Under these circumstances meaningful, comparative and co-operative intelligence would be very hard if not impossible to come by, if MAST was not there to provide it.
MAST also includes elements that go considerably beyond detailed access to STATS19. It provides stakeholders with unique residency and socio demographic analyses which would be impossible to replicate without full access to the national dataset. It includes detailed coverage of the entire British Strategic Road network. I could go on.
MAST is provided by a “third party company” because that is how the Department for Transport want it done. It is far and away the cheapest, most convenient, and most responsive way to deliver such a service.
Finally, accident descriptions, while very valuable for small scale analyses of individual junctions or short routes, are not a practical resource for quantitative study of trends. Yes, every incident is different to some degree – but if we concentrated purely on reading descriptions and making subjective judgements on them, it would be impossible to analyse trends consistently. If that was all anyone ever did, none of us would be able to see the wood for the trees.
Bruce Walton, Oxfordshire
0
I still don’t get this recent emergence of third party companies/organisations providing and analysing accident data for the benefit of those who would surely have it already (and in more detail) i.e the LAs.
Even for non-LAs, if it’s purely data that’s being supplied, then the most important information i.e the description of what actually happened, can’t be provided anyway can it?
Hugh Jones, Cheshire
0