There are any number of things to gaze at while walking down a city street, but you can't stop and stare at them all or you'll never get where you want to go. If you just arrived from rural Idaho and it's your first trip to Times Square, that might be the point. But how do the rest of us decide when to break stride and look at whatever everyone else is looking at?
One of the earliest studies of crowd gazing was done by Stanley Milgrim on the streets of New York in the late 1960s. Milgrim had research actors, either alone or in groups of up to 15, stand on the sidewalk and look up at the window of a nearby building. Then he and colleagues determined the effect this behavior had on other pedestrians.
A single person had a modest influence, enticing 42 percent to glance up and about 4 percent to stop and look. But a crowd of 15 had an incredible pull: 82 percent of passing pedestrians copied the glance, and a full 40 percent stopped to look. Milgrim concluded that the number of people who will "react to, and join in" a staring crowd is strongly related to the size of the crowd.
The finding passes the common-sense test, but it doesn't quite tell us why we bother to behave this way. After all, in the Milgrim study, the window that served as the object of attention was, as the researchers put it, "not a scene of compelling interest." Some behavioral scientists have suggested that passers copy gazing behavior as a form of social conformity; others believe we do it to learn something about a local environment we'll soon be entering.
Recently an international research team, led by evolutionary biologists Andrew Gallup and Iain Couzin of Princeton, reprised Milgrim's study on a crowded street in Oxford, England. (Special thanks to science writer Michael Gross, at Current Biology, for directing attention toward this work.) Gallup, Couzin, and colleagues fixed a video camera above a strip of sidewalk, then like Milgrim they sent out actors to stand below it and gaze up for 60 seconds.
Altogether the researchers captured the gazing patterns of about 2,800 pedestrians who passed through the filming zone. Across all the tests roughly 27 percent of pedestrians copied the gazing behavior of the actors looking skyward, the researchers reported earlier this year in the journal PNAS.
Gallup, Couzin, and company then took a closer look at these 760 gaze-followers. Of them, nearly half looked up multiple times during their walk, with an indecisive 4 percent or so glancing up five times. Roughly 14 percent did the full stop-and-gaze. Men tended to gaze a bit more than females; slow walkers a bit more than fast ones.
The big question, of course, was how the size of the group of actors influenced passing pedestrians. The researchers did find that the proportion of gazers increased with the number of actors. When a single actor gazed up, about 5 percent of passers did the same. When 5 to 10 actors gazed, that figure rose to roughly 30 percent. And when a dozen or 15 actors gazed, about half of pedestrians wanted to see what the fuss was.
So as Milgrim knew, stopping and looking clearly encourages more stopping and looking. But the gaze-attraction in the present study wasn't nearly as strong as Milgrim suggested. In Milgrim's test, the maximum probability that a passing pedestrian would copy someone's gazing behavior was 92 percent. In the new work the high end only reached 66 percent, and the group size needed to produce half this maximum was six times larger. A comparison of the response data reveals a striking contrast:
Having qualified Milgrim's findings, the researchers then elaborated on them in an important way: they mapped the trajectories of the pedestrians who copied the gazing behavior. Doing this showed that people were significantly more likely to look up when they crossed behind a gazer than in front. That's not what you'd expect to find if people only gazed to fit in with the crowd. After all, if you wanted other gazers to know you too were gazing, you'd presumably do it in front of them, where they could see you.
The finding suggested that conformity didn't play as great a role in sidewalk staring as some scientists believe. But the result was far from conclusive, so Gallup and Couzin designed another research study to get a better sense of precisely when (and, by extension, why) gazing behavior transfers among eyes in a crowd.
This time they hid a video camera inside a decorative fixture outside a building entrance and taped large red arrows around the fixture to entice people to gaze. As in their previous study, the researchers found that people looked when others looked. While baseline gazing behavior was around 28 percent, that rose to 49 percent when another pedestrian had looked within 3 seconds. (This test, done with a different collaborator, was published recently in the British journal Biology Letters.)
But on closer inspection the researchers realized that the social transference of staring wasn't equal for all pedestrians. Across all trials, the rate of gaze-copying was significantly higher for people walking in the same direction as a gazer (57 percent) than for people crossing paths with a gazer (20 percent). Simply put, people followed the eyes of other people in front of them, not those coming toward them.
Gallup and Couzin conclude that there is certainly a social element to sidewalk gazing, but it's not necessarily an urge for conformity. Rather, it may be that people mimic the visual behavior of those going their way to gain information about their future environment — what they term a "rearwards transfer" of attention:
That is, individuals walking in the same direction ahead of you are interacting in an environment that you will shortly experience, and thus cues relating to this context may be more important than those coming from oncoming pedestrians.
Still a lot more work to do on the subject, to be sure. But the finding at least fits with one regularly observable phenomenon: why New Yorkers don't bother turning to look at whatever tourists from Times Square are turning to look at.