Building It into Every Drive

There are four kinds of learning, and they explain how hazard perception is built:

• Superficial learning memorises answers to pass a test and forgets them. It is useless here: no list covers every road.
• Deep learning understands why: why the parked lorry hides a child, why the herder means cattle. It is what lets you handle a scene nobody taught you.
• Over-learning makes something automatic. The point is that car control should be over-learned, meaning clutch, gears, steering and mirrors done without thinking, so that your attention is free for everything around you. That is why new drivers fix their eyes (Where Your Eyes Go): their attention is still inside the car.
• Imitation learning copies someone else. It is good or bad depending on who you copy, which is worth thinking about when you choose who supervises your practice.

So practise on purpose:

• As a passenger, every journey. Find the developing hazard before the driver reacts. Commentate quietly.
• On routes you know, look for what you have stopped noticing. Familiarity is where the bad kind of probability learning grows: "nothing ever happens here."
• After a near-miss, ask what the first clue was and when you could have seen it. That is the good kind of probability learning: experience that makes you more careful, not less.
• For the theory paper, NTSA asks hazard perception as multiple-choice questions ("what is the developing hazard?") and the practical test takes it onto the road. Learn the method, not the answers, and both will follow.

You will never know everything, so aim for lifelong learning. Experienced drivers who stop practising drift back towards the inexperienced pattern.