Turn Signals - An information design problem
As things stand the communication between drivers on the road is very poor. It's limited to a horn, flashing headlights, turn signals, brake lights and reversing lights.
Brake signals are a wonderful invention, highlighting very simply when to pay extra attention to speed. Without them there would undoubtedly be many more rear ending incidents.
Reversing lights are also very helpful, since reversing is a relatively unusual activity, it helps to warn people when they should be more careful. In parking lots particularly it helps pedestrians spot cars that are about to leave. This is clearly very important since the driver has reduced visibility of hazards when driving backwards.
Both brake lights and reversing lights though are automatic. The only signals used as a matter of course and manually operated are the turn signals.
These provide a valuable way of conveying a wide range of intent from one driver to another. However, they are not without problems particularly in the US where both red and amber turn signals are used. In Europe, and to the best of my knowledge most of the rest of the world, amber turn signals are the only legal colour.
The problems with red turn signals are numerous.
Most important is the ambiguity between brake lights and turn signals. It may seem that this shouldn't be a problem, and I fully expect the original car makers thought it wouldn't be since a turn indication uses one light only. However, on modern highways it is common for only one side of a vehicle to be visible and this makes it hard to rapidly distinguish between a turn signal and braking. Essentially it is impossible until the signal has flashed at least once. Given drivers impatience they frequently don't leave enough time between the turn signal and the manouver, so any lost time is vital. On busy highways driver performance is massively important, and any ambiguity is dangerous. With driver impatience and lost visibility a flashing red taillight is virtually useless.
The amber turn signal on the other hand is much less ambiguous. It can not be mistaken for a brake light, but it is possible for a hazard warning to be read as an intention to turn. This is awkward, but much less hazardous than the braking ambiguity, but it does mean that a driver who can only see one side of the signalling vehicle may not be sufficiently warned of a hazard. Of course the red turn signals suffer from this failing aswell.
All modern cars have a third brake light though. This should help to disambiguate braking from turning, but it fails for precisely the same reason, it is not always visible. Not that it's unhelpful, it does make the braking signal that much stronger, but it doesn't help with turn signals. Also older cars don't have the third brake light, so drivers are used to seeing vehicles without it. This weakens its effect in disambiguating. The important point here is not that it can be determined, but rather how long it takes. Interpretting these signals should not be a conscious process. If it were it would be much to slow.
Red turn signals also suffer from another, albeit less severe problem, which also affects brake lights, and may have been the intent behind the third light. At night when headlamps and tail lights are on there is already a red light on. Because of this there is reduced contrast between the signal being on and the signal being off. In practise this is generally unimportant, but it becomes more so when driving in poor visibility during rain or fog.
The contrast problem is also encountered with the front amber running lights common on many American cars. These provide very little utility. It could be said that they indicate the vehicles presence in the case where one or both of the headlights have failed. But this is hardly a powerful arguement. More importantly it reduces the contrast of the flashing turn signal. This loss of contrast makes the turn signal much weaker. The running light should not be co-located with the turn signal, and it should be white.
Recommendations
It's often easy to critisize, but hard to find better alternatives, but in this case it's incredibly easy to find an alternative, one which is already common on many vehicles in the US, and all in the rest of the world. Use amber turn signals! As for the hazard lights it would be easy to use a double flash or even yet another colour to disambiguate this from turning. Why this has not been legislated on I don't know. I would expect that it would help reduce fender benders by a noticable margin. It may simply be a matter of style for some designers. Having almost all the lights in the rear signal cluster one colour may look better to some people, but I wonder if this can really be considered as important as safety on our ever more crowded roads.
References
A 2008 study published by the US National Highway Traffic Safety Administration tentatively supports my observations from 2001. It's good to see rigourous work being carried out to find the best options. Hopefully this will lead to new regulations, and better turn signals.
