Do look, don't see!

Vision scientist, magician and psychologist Gustav Kuhn from Durham University has studied where the attention of spectators is focussed when they watch a magic trick. He used eye-tracking technology to do this - technology that monitors exactly where a person's eyes are directed.

He has found that misdirection, where a magician tries to get you to pay attention to one area while they are doing some secret move in another place, is fascinatingly complex. We aren't always aware of what our eyes are looking directly at.

Looking isn't seeing

One laboratory experiment involved a trick where a cigarette vanished. In fact, it was openly dropped behind the table while the magician was looking at the lighter in his other hand.

They found that the position of the eyes when the cigarette drop took place wasn't that important. People could be looking at it fairly directly and still not know what happened. This was because their 'attention' was with the magician's other hand, i.e. where the magician was looking. Their brain missed the very visible drop! This effect can give us useful clues to understand how we perceive.

Look where I look

In another experiment he showed that so called social cues (eg we all tend to look where others are looking), were critical. If the magician didn't look directly at where the misdirection was going on, he was often found out. The magicians face is often a focus of our attention, we like faces because they tell us a lot. In a vanishing ball experiment he found that between each toss of a ball into the air the spectators looked at the magician's face. When on the third toss the ball vanished (it wasn't really flung but the magician's face and eyes followed the path the ball would have taken), many of the spectators believed they had seen the ball vanish in mid air, the magicians face had told them what to expect (even though it was wrong).

Don't Tell

Finally he showed scientifically what magicians have known for a long time, that with these sorts of misdirection tricks, you are more likely to be found out if you tell your audience what you're going to do in advance!

Human Computer Interaction

This research is very important for computer scientists to understand if they are involved in the design of interfaces. If it is important that people see certain aspects of an interface such as warnings then the designer has to design in ways to manipulate their attention.

Eye tracking technology is also starting to be used in everyday devices like cameras and cars so that the gadgets know where the person is looking. For example, cameras can focus on the thing you want them to and cars can tell if you are not looking at the road - and give you a poke if you are falling asllep at the wheel. What this research shows is such gadgets can't be sure of what you are paying attention to from eye tracking data alone.

It is common for Human-computer interaction researchers to use eye-tracking technology to evaluate their interfaces - to check that users do look at things in the order they expect, for example, and to help explain why interfaces are hard to use so they can be improved. This research shows that more care may need to be taken though. The person may be looking at the right thing (as indicated by the eye tracker) but still not be seeing it!

Social Robots

Researchers investigating social robots - those that may one day become our companions, for example, this shows some of the subtle things that need to be engineered into robot behaviour if they are going to seem natural. They should for example look at things they want their human companions to look at to help focus their attention. That also means of course that it may be important for them to have a recognizable face!