The Say-Do Gap in Driver Behaviour

We recently finished a project about speeding and it was one of those projects where counter-intuitive findings came thick and fast.

Our goal was to examine ways to encourage drivers to keep to speed limits, so we needed to really get to the bottom of the behaviour. 

We recruited two groups and put dashcams in their cars for a week: 

  • Those who said they kept to speed limits
  • Those who admitted speeding from time to time 

The surprising thing? Everyone broke speed limits. In fact, some of those who claimed not to speed had more incidences of speeding than those who admitted it. 

Attitudes didn’t correlate with behaviour one bit. 

We also spent time driving with people on their regular journeys. Getting to work, picking kids up from school going to the shops. We thought the interviewer effect could mean people would be on their best driving behaviour with us present. Truth be told, after 5 minutes of driving people forgot we were there. 

This accompanied driving really helped our contextual understanding of where people drive too fast. 

Real-world driving is often automatic. Most of us will have experienced a time when we’ve been driving home – then suddenly realised we can’t remember driving for the past 5 minutes. You can drive the same route so frequently that you’re able to respond to it automatically. Here driving is a trade-off between reacting to unexpected occurrences (someone cut me up!) and habit (context-triggered actions when driving a known route).

This is why tactics like vehicle activated signage and road redesign make a real difference to speeding behaviours. In observations, we saw how they re-focus driver attention.  

We used the ISM model to structure our analysis. It breaks down behaviour into three component parts: 


  • Individual factors relate to the person: their values, attitudes, skills, and evaluations 
  • Social factors relate to the presence of other people on the road
  • Material factors relate to aspects of the environment and infrastructure which affect how the driver interacts with the road 

Whilst we uncovered lots relating to the individual and their personal circumstances, many of the surprising and counter-intuitive findings related to material factors. Reviewing driving footage showed us how much people reduced their speeds on narrower roads for example. 

Our analysis of the cues in the built and social environment – combined with individual beliefs and habits – formed the basis of a new strategy to encourage behaviour change, which we developed in interviews and group workshops.

This project has been a reminder of some of the principles we abide by:

  • We define the behaviour carefully 
  • We observe, as well as ask 
  • We triangulate our methods to get the complete picture 

The last one is particularly important. Attitudes are a poor predictor of driving behaviour. When it comes to a complex issue like speeding, researching attitudes alone is not enough.