In partnership with
Improving road safety: new results
On 17 August 1896, Bridget Driscoll – a 44 year-old housewife from Croydon – was travelling to a folk-dancing display in Crystal Palace. On her way there she was hit by a car. She died minutes later from the resulting head injury, becoming the first pedestrian in the UK to be killed by a motorised vehicle.
At her inquest, the Coroner William Percy Morrison said he hoped that “such a thing would never happen again”. But more than a century later, over 1.25 million people globally are killed in road traffic accidents every year. In the UK, where roads are generally considered safe, five people will die in a traffic accident today.
Key to improving road safety is understanding what causes serious and fatal collisions. In work with East Sussex County Council, using cutting edge data science techniques to analyse more than a decade’s worth of data, we found that 10 per cent of all collisions, and 7 per cent of those that result in a death or serious injury, are caused by people who have at least one speeding conviction.
Despite the death toll, and despite previous brushes with the criminal justice system, it seems people continue to drive dangerously; disregarding road safety rules and failing to comply with speed limits. In the West Midlands area alone, there were more than 60,000 traffic offences in 2015.
As well as posing a risk to life, these offences create costs for criminal justice agencies, especially when drivers have to be prosecuted for non-payment of fines.
In 2015, we partnered with West Midlands Police to tackle the problem of dangerous driving.
We focused on adapting an existing point of contact with speeding drivers – the Notice of Intended Prosecution received after being caught speeding. We identified two areas for improvement:
- making it easier for drivers to comply with the sanction: simplifying the communication in the letter
- convincing drivers of the legitimacy of speeding limits so they are less likely to reoffend: explaining why speed limits exist and the dangerous consequences of breaking them
Following a clustered randomised control trial over 19 weeks, with a cohort of 15,346 drivers, we found that the intervention reduced reoffending by 20 per cent within six months of an individual’s offence in the West Midlands alone.
These results are in addition to the increased payment rate and speed previously reported, which reduced eligibility for prosecution by 41.3 per cent. Using police and Home Office data, we estimated the intervention will save the criminal justice system £1.5 million per year in the West Midlands alone – as well as reducing the numbers hurt or killed on our roads.
This trial, which won first place at Nudge Awards’ Nudge for Good category this summer, demonstrates the impact of small, low cost, changes to existing process and communications. It is a classic nudge, applied to a hard to shift behaviour. We think this approach – making the rationale behind laws more salient – is a useful tool for law enforcement, which we hope to test in other domains.
Back in 1896, witnesses to Bridget Driscoll’s death said that Arthur Edsel, the driver of the vehicle that killed her, was travelling at a reckless pace. And in present day East Sussex, our findings were very similar: that a driver being careless, reckless or in a hurry was the most common recorded factor of collisions resulting in deaths or serious injuries. A century has passed and some things have changed little.
Fortunately, now, we have new tools and techniques at our disposal to make our roads safer for the next century.
Want to work with us on similar trials? Get in touch: email@example.com
Subscribe to our mailing list
- Australia (11)
- BI Ventures (2)
- BX Conference (11)
- Charitable giving (8)
- Consumer affairs (2)
- Data science (1)
- Education and skills (17)
- Energy and sustainability (6)
- Equality and discrimination (3)
- Health (17)
- Home affairs and security (9)
- International (6)
- Labour market & economic growth (8)
- New Zealand (1)
- News (52)
- North America (14)
- Reports (19)
- Tax (4)
- Trial design (6)
- Trial results (24)
- Uncategorized (96)
- Vacancies (11)