An attentive glance, an affirming nod, a warm smile--these may sound like the makings of a love sonnet, but in a groundbreaking study on instantaneous impressions, they're the fleeting images of a Harvard teaching fellow (TF) leading a section. Glimpses such as these, captured on soundless video clips only two, five, or 10 seconds long, were rated by student judges who had never met or taken classes with the TFs on tape. "The students see this flash of a person, and it's silent, so they're just getting what we call 'expressive behavior,'" explains Hazel associate professor of the social sciences Nalini Ambady, Ph.D. '91. "Yet the impressions they form based on that behavior are pretty highly correlated with end-of-the-semester ratings by students who have actually taken the class." Contrary to expectation, first impressions--based on extremely brief moments of behavior--can be remarkably accurate.
"I was totally surprised. I really did not expect any results at all," Ambady recalls. Previous "thin-slice" research had used five-minute video clips, but Ambady and coauthor Robert Rosenthal, Pierce professor of psychology emeritus, wanted to know how thin a slice of behavior could be and still yield accurate impressions. Ambady started with 10-second clips, then "sliced it down" to five- and then two-second clips, having her judges rate the same 15 characteristics in every case, including how accepting, empathic, optimistic, professional, or supportive the teacher seemed. No matter how thinly Ambady sliced the behavior, the first impressions of strangers successfully predicted the evaluations of students who had worked with a teacher for months. The correlation coefficient was .76, which is "as strong as you can get from these types of judgments," she says. In short, the more positive and likable the TFs, the higher their evaluations.
Ambady was thrilled but somewhat disconcerted by the results, which were published in the Journal of Personality and Social Psychology. "One would think that preparation and organization should count--and I'm sure it does to some extent," she says, "but behavior, charisma, and the factors that go into holding an audience count, too."
Rather than boiling charisma down to specifics, Ambady prefers to talk of "constellations of behaviors" that signal what she calls "'warm fuzzy' concepts, like how enthusiastic the person is, how warm, how likable." Since people perceive this "likability/positivity" dimension very quickly, its presence or absence helps predict performance in contexts where such characteristics count, such as sales, psychotherapy, healthcare, or teaching. "How expressive a teacher is really matters. It's what holds a student's attention and motivates a student to work hard, so it might be directly correlated to student learning," she explains. "If, on the other hand, a teacher's style turns a student off, then it's unlikely the student is going to learn a lot, no matter how well organized a teacher is, or how clear."
The effects of a surgeon's expressive style are the subject of Ambady's recent study with Wendy Levinson, a physician at the University of Chicago Medical School, and others. Could thin-sliced observations of a doctor relating to a patient "postdict" past malpractice suits against that doctor? To answer this question, they chose a different mode of nonverbal behavior than facial expressions and physical gestures, Ambady says. "One channel that most people aren't able to control and don't pay a lot of attention to--and that's thought to be 'leaky' of intentions or feelings--is the voice."
Ambady pared down Levinson's audio data from 114 surgeon-patient interviews to thin slices of two 10-second sound clips per doctor. She then narrowed the input further by removing the sound frequencies necessary for word recognition. As a result, her judges heard no verbal content, only the tone of a doctor's voice. "We were really amazed," she recalls, "because we found that with just 20 seconds of each doctor's voice, you could postdict malpractice claims. For instance, surgeons who sounded more unfeeling or dominant were more likely to have been sued in the past. This is very interesting, because it feeds into lay stereotypes about surgeons as being cold and uninterested in people--and it turns out that doctors who fit that stereotype are more likely to be sued."
Medical schools may try to train physicians-to-be in good bedside manners, but Ambady is dubious about such coaching. "If you're multitasking or under stress, you tend to revert to your mean, whatever your average behavior is," she says. Besides, "People are very good at picking up the meaning underlying behavior. Someone could smile, make eye contact, and shake your hand, and you could think, 'Oh, this person's faking it.' We have this intuitive radar."
Ambady's research pinpoints when that radar works and when it doesn't. Whereas accurate analytic judgments depend on careful reflection, "thin-sliced judgments" suffer under such scrutiny. "If I tell people, 'I want you to write down every reason why you're making these judgments,' then that completely stifles and inhibits accuracy," she explains. "These thin-sliced judgments are like riding a bicycle--if you start paying attention to every move you're making, you're going to fall off."
Despite the remarkable accuracy of some impressions, "I don't want the message to get out that I think that all first impressions are right," Ambady says. "There are times when it is very dangerous to make these first-impression-type judgments, especially when using stereotypes." First impressions clearly have other limits as well. "I don't think you can pick up the deeper layers of the onion," she notes. "If you think of the human personality as consisting of different layers, it's the outer layers--extroversion and warmth and expressiveness--that people can pick up. It would be pretty hard to pick up intelligence or creativity. People claim that you can, but I think the jury's still out on that."
