The skill a test pass doesn't give you: reading the road early

A large review by the Cochrane Collaboration pulled together 24 trials covering close to 300,000 drivers. It found no evidence that formal post-licence driver training reduces crashes or injuries. Across the 15 trials that recorded actual crashes, the difference between trained and untrained drivers was almost exactly nothing (a relative risk of 0.98, which for practical purposes means "the same"). A separate 2018 review for the European Commission reached a matching conclusion: learning formally to pass the test leaves a driver no safer afterwards than learning informally with family or friends.

None of this means lessons are pointless. It means the usual lesson content isn't aimed at the thing that causes crashes. The European review is blunt about the gap. Traditional training, it says, is "about vehicle handling, applying the rules of the road and mastering basic traffic situations", with "not much attention" paid to hazard perception, risk assessment, self-assessment, fatigue and distraction. Those last few are the factors that put young drivers in the ditch. In other words, most training teaches the car well and the road poorly.

Put a new driver and an experienced one on the same road and the big difference isn't reflexes. It's where the eyes go and how early the brain flags a problem. Experienced drivers scan further ahead, check back and to the sides, and read small cues: a ball bouncing out between parked cars, a pedestrian who hasn't looked up from their phone, brake lights three cars in front. They've turned "what might happen next" into a habit, so by the time a hazard develops they've already eased off and left themselves room.

A new driver, by contrast, tends to look at what is directly in front of the car and deal with things as they arrive. That's not carelessness. It's just what an under-trained eye does. The good news is that this is a pattern of looking and thinking, and patterns can be taught.

In 2024 a team led by Prabhakharan pooled 57 separate studies of hazard perception training. The result was a clear, large improvement in drivers' hazard perception (in the technical terms researchers use, an effect size of Hedges' g = 0.78, which is a big one). The same benefit showed up for cyclists (0.97), pedestrians (0.64) and motorcyclists (0.42). Very little else you can add to a lesson has evidence this consistent behind it.

One detail from that review shapes how we teach it. The gains were most consistent when the training was active, meaning the learner is doing something rather than watching. Sitting a young driver in front of a screen of clips is the weak version. Getting them to talk the road out loud, and to call what might happen before it does, is the strong version.

Keep it low-pressure. This is about building a habit, not marking a test. Ease off the commentary when the road gets busy, because safety comes first and a learner shouldn't be narrating through a demanding junction. This pairs naturally with our work on judging your own skill honestly and with our coaching approach, where the learner does the thinking rather than being told the answer.

1 · Making it passive. A learner watching you point out every hazard is getting the weak version of the training. They have to be the one talking, looking and predicting. If they've gone quiet, the eyes have usually stopped moving, so a simple "and ahead?" gets the scan going again.

2 · Settling for description instead of prediction. "There's a car" is a start, but it's the easy half. The value is in "that car might pull out, so I'm ready." Always push the learner one step past what's there to what could happen.

3 · Forcing it through a busy moment. When the road demands full attention, stop the commentary and let them drive. The habit is built on the quieter stretches, not by talking over a hard piece of road. Safety always wins the argument.