The Grove Road Experiment

The local polis are out on the streets monitoring junctions and telling everyone off if they do something they shouldn't, which comes down to : Cyclists and red lights, cars and ASLs and one or two mobile phones.

Yet anyone who doesn't see people in hi-viz police wear at a junction is either too busy reading their facebook notifications or just dim -they should have noticed what what was happening, and behaved.

But no, people got caught by police wearing hi-viz, on a publicised operation. Which makes us wonder: would many more have been caught if the police hadn't been wearing hi-viz.

Being a data science organisation, we decided to see if people do actually change their behaviour and drive legally when they believe that they are being observed.

We chose Grove road for this, as the "no entry" sign has only been there five years and people still resist this unwanted war-on-motorists infrastructure by ignoring it completely.

Key features

1. Contraflowing it can shave about 3 minutes off the journey time from Redland Hill to Whiteladies Road, so aid the school run or the commute
2. You have to commit before you get stuck in the last twenty or so cars at the Redland Hill roundabout, so either plan to use the cut through always, or make a decision based on congestion.
3. If you change your mind you have to reverse and cut up to Redland Hill again -losing you time, especially as you will now have to wait to cut back in: it will cost you more than if you had stayed in the queue.

Being game-theory enthusiasts too. we like this, as it means

1. You have to decide whether to opt to break the law before you can see if the junction is monitored.
2. The gain of breaking the law is 2-3 minutes saved.
3. If the junction is monitored, you are left with the choice of retreat -costing you more than if you hadn't turned off, or you can continue

People aren't likely to just turn back because you point out its a no-entry sign: they know that, and they've made a conscious decision to ignore it because the gain "2-3" minutes matters to them more than the probability that they will get penalised for driving the wrong way down the road.

What then, does it take?

This experimented was conducted on a weekday between 08:28 and 08:38; this sequence shows all but one of the cars choosing to run the sign (that one was omitted as the camera team were in a conversation with a friend and not watching it properly -they shall be soundly beaten).  Only three cars came the other way, showing that that there are two cars driving illegally for every one legally. This gives all car drivers a bad name.

Experiments included:

• Point and talk: no, driver calls cyclist a git
• Wave at passing car for attention: no eye contact
• Point at sign: car does not slow down
• Stand in middle of road and act like you are texting: first car stops, but the mercedes  CLS 320 S22JBW doesn't.

What does appear to make a difference -and you can see from the final two events -is be blatantly holding a phone up as if you are about to take a photograph of the car. Once the driver decides that they are going to be photographed, then they change their minds and revert to the original route. We got a 100% success on that option.

This argues that it is not public attention that changes people's behaviour, nor is it the presence of registration numbers on the vehicle. It is the drivers' concluding that the risk of there being a penalty of of being photographed driving past the no-entry sign outweighs the cost of reversing back and continuing on the original route.

It doesn't matter if people see you, as long as the police don't.

The irony is, of course, that the entire experiment was being filmed on a helmet camera -yet somehow the sight of a phone being used as a camera changed driver behaviour in a way that a cyclist with a helmet with a camera attached did not.