I was reading through some old research on emotional expression, and in the process re-read one of the most astounding studies I've ever seen. It is by Nalini Ambady and Robert Rosenthal and is called "Half a Minute: Predicting Teaching Evaluations from Thin Slices of Behavior and Physical Attractiveness." It was published in the Journal of Personality and Social Psychology in 1993; I think you can download it here (at least it worked for me). They report three studies, but I will focus on the first, as it is makes the point.
Here is what they did. They brought in nine female undergraduates and showed each just 30 seconds of videotape (with the sound off) of 13 different instructors. For each instructor, these undergrads saw 10 seconds of them teaching at the start of class, 10 seconds in the middle, and 10 seconds at the end. On the basis of these 30 seconds, the undergraduates rated the instructors of 15 dimensions, things like competent, confident, honest, likable, professional, warm, and dominant. They then correlated these "thin slices" with the evaluations given by (different) students who had taken a semester long class from these same teachers. The correlation's were staggeringly and disturbingly high. The overall correlation for the 15 item scale was .76 — meaning that those 30 second silent slices were remarkably similar to what students concluded about their instructors after spending 16 weeks in class with them. Looking at the individual dimensions, some of the individual correlations were even a bit higher, such as confident (.82), optimistic (.84), and dominant (.79). But to me the really striking ones were judgments that seem mighty hard to make from 30 seconds of silent video, notably competent (.56), professional (.56), and supportive (.55). If you read the rest of the paper, you will see that they then replicated the study with high school teachers (achieving similar but slightly weaker results, as the overall correlation was .68) and then in a study with even thinner slices of teacher behavior (6 second and 15 second clips) they also replicated the pattern.
I was also taken with two other things about these studies. First, the researchers used statistical controls to remove any effects of physical attractiveness (it had little effect, although in one study the effects were even stronger when the controls were used). Second, all of these ratings were done by female undergraduates, and my guess is that they were female undergraduates at Harvard University. They say they used women because studies have shown that women are more skilled than men at decoding nonverbal emotions. I would also add that, as the population of Harvard undergraduates is from a similar pool as Stanford undergraduates, I expect that this is as emotionally sensitive a group of young people as you can find. To get into a university like Harvard, you need to not only have great grades and scores, you need to impress a lot of people along the way and often hold diverse leadership positions — accomplishments that depend on being able to read and respond to small emotional cues from others, especially from your teachers. So, as impressive as these results are, the question of how strong they are in broader population remains — although it does appear that this "thin slices" approach has since been replicated in more diverse samples.
These details aside, this research is stunning because –even if the effect is say only half as strong in the general population– the degree to which people can make accurate evaluations on the basis of the smallest hints is scary for job candidates and for bosses… it means that people size you up very quickly and make judgments that may turn out to be quite consistent with those who know you well. It also is related to the research I talked about on the toxic tandem and the baboon stuff — one reason I suspect these undergrads were so skilled at judging teachers was that by the time they get to the university, they had 13 years of experience watching for tiny emotional cues among those who wielded power over their grades and recommendations. This continues in the workplace, of course, as underlings become remarkably astute at making correct interpretations of their bosses non-verbal behaviors, as we saw with the "interesting shoes" story and the stuff on baboons and bosses.
Let me know what you think about this research. I know it seems hard to believe, but the power of "thin slices" has been replicated in multiple studies now.
Leave a Reply