1 00:00:01,276 --> 00:00:07,076 The first experiments we did were with financial bonuses. 2 00:00:07,076 --> 00:00:13,759 But what about other forces? For example, social recognition. 3 00:00:13,760 --> 00:00:20,692 We did a task of anagrams. We asked people to solve some anagrams and 4 00:00:20,692 --> 00:00:25,630 we paid them by their success. And people were scattered around the room 5 00:00:25,630 --> 00:00:30,447 and every person had the little cubicle. And they were solving anagrams for pay. 6 00:00:30,448 --> 00:00:34,876 But from time to time, we asked somebody to get up, come to the board in front of 7 00:00:34,876 --> 00:00:38,364 everybody, and solve the anagram in front of everybody. 8 00:00:38,365 --> 00:00:42,445 They still got paid in the same way. But now they solved it in front of 9 00:00:42,445 --> 00:00:44,925 everybody. Now would this be more important to 10 00:00:44,925 --> 00:00:48,700 people, or less important to people? It's more important because you take the 11 00:00:48,700 --> 00:00:53,322 financial motivation to solve anagram. Plus you add to it the social desire to 12 00:00:53,322 --> 00:00:56,530 have your friends think you're a smart person. 13 00:00:56,530 --> 00:01:01,139 And to avoid looking foolish. But what actually happened to performance? 14 00:01:01,140 --> 00:01:07,285 Performance was about half as high in the public condition than in the private 15 00:01:07,285 --> 00:01:10,660 condition. When people wanted the most to perform 16 00:01:10,660 --> 00:01:15,502 well, they actually worked the hardest. And this is of course things like test 17 00:01:15,502 --> 00:01:18,887 anxiety and so on. When we want to perform to the best of our 18 00:01:18,887 --> 00:01:25,123 ability, it doesn't mean that we can. It means we want to but this wanting to 19 00:01:25,123 --> 00:01:29,532 can actually overwhelm our ability. So, what do we have? 20 00:01:29,532 --> 00:01:33,820 When we think about the relationship between incentive and performance, we 21 00:01:33,820 --> 00:01:36,577 usually think about just an increasing amount. 22 00:01:36,578 --> 00:01:39,970 Sometimes we think about diminishing returns. 23 00:01:39,970 --> 00:01:45,361 Very rarely do we think that performance would go up and then reverse, and go down. 24 00:01:45,361 --> 00:01:48,932 But so far when we've tried things, this is what we see. 25 00:01:48,932 --> 00:01:54,767 And I should point that somebody did an analysis of how people play golf when they 26 00:01:54,767 --> 00:01:58,796 play next to Tiger Woods. This was in the period before he had all 27 00:01:58,796 --> 00:02:01,750 these affairs and he was the top player in the world. 28 00:02:01,750 --> 00:02:04,724 And what do you think would happen to somebody who's playing next to him? 29 00:02:04,725 --> 00:02:10,347 Would Tiger Woods push them to be better or stress them to perform worse? 30 00:02:10,348 --> 00:02:14,157 Well, the results show that he stresses them to perform worse. 31 00:02:14,157 --> 00:02:18,987 By the way, from this perspective, it is also interesting to speculate about Tiger 32 00:02:18,987 --> 00:02:22,451 Woods. He was a guy who was the most important I 33 00:02:22,451 --> 00:02:28,070 guess, successful golf player ever. And all, he was at the height of his 34 00:02:28,070 --> 00:02:33,350 career, all of a sudden he found out that he has some affairs and his performance 35 00:02:33,350 --> 00:02:35,453 drops down. How can it be? 36 00:02:35,454 --> 00:02:38,067 Right? I mean, could it be that he couldn't 37 00:02:38,067 --> 00:02:41,623 concentrate, that he couldn't think very much, that his mind was preoccupied? 38 00:02:41,624 --> 00:02:46,854 I mean, that's one interpretation, but it tells you how easy it is to get people to 39 00:02:46,854 --> 00:02:51,924 being distracted and by thinking about some other things not performing very 40 00:02:51,924 --> 00:02:53,900 well. One of my friends had a different theory. 41 00:02:53,900 --> 00:02:57,728 He said that maybe what he was doing is he basically had no release for his sexual 42 00:02:57,728 --> 00:03:00,790 energy so he was all pent up and he couldn't perform very well. 43 00:03:00,790 --> 00:03:05,757 And I don't know how to differentiate between these two, these two theories. 44 00:03:05,758 --> 00:03:09,337 So now, let's go back to Wall Street executives. 45 00:03:09,338 --> 00:03:12,954 I meet these Wall Street executives, and they say, oh, your results are very, very 46 00:03:12,954 --> 00:03:15,824 nice, but they don't apply to us, to the Wall Street executives. 47 00:03:15,824 --> 00:03:18,755 Why? Because your experiment, dear Dan, are 48 00:03:18,755 --> 00:03:23,250 carried on regular normal people, students, people in India and so on. 49 00:03:23,250 --> 00:03:27,865 We are super special people. If you hide the distribution of people, we 50 00:03:27,865 --> 00:03:32,222 would be at the edge. Regular people would stress, get stressed 51 00:03:32,222 --> 00:03:36,410 by a thousand, 2000, $50 million. We don't worry about this. 52 00:03:36,410 --> 00:03:40,502 In fact, there's not an amount of money in the world that would get us to go off the 53 00:03:40,502 --> 00:03:43,714 continuous improvement. And you can pay us as much as you want and 54 00:03:43,714 --> 00:03:46,585 we would always improve in quality and never deteriorate. 55 00:03:46,586 --> 00:03:50,560 Now, of course, what I do when they say something like this is to invite them to 56 00:03:50,560 --> 00:03:52,770 the lab. And I say you know what, maybe you're 57 00:03:52,770 --> 00:03:56,475 right, why don't you come to the lab and let's test things out and if you'll give 58 00:03:56,475 --> 00:04:00,408 me a research budget, we could see whether you indeed don't get stressed with these 59 00:04:00,408 --> 00:04:03,916 large amounts of money. But none of them is truly that interested 60 00:04:03,916 --> 00:04:06,739 in doing it and nobody has ever appeared in the lab. 61 00:04:07,850 --> 00:04:10,266 But nevertheless, it was an important point. 62 00:04:10,266 --> 00:04:13,416 I don't believe that. I don't believe that all the stress free 63 00:04:13,416 --> 00:04:17,322 people are going to Wall Street and these are not normal people and so on, but 64 00:04:17,322 --> 00:04:20,146 maybe. Now testing people in Wall Street is very 65 00:04:20,146 --> 00:04:22,627 tough. So, instead what we did was we went to 66 00:04:22,627 --> 00:04:28,416 look at professional basketball players. And together with Coach Kane the head of 67 00:04:28,416 --> 00:04:35,059 the university basketball team and we got his help to contact professional coaches 68 00:04:35,059 --> 00:04:39,761 of teams at the NBA, the, the national league for basketball. 69 00:04:39,761 --> 00:04:43,940 And we asked him to tell us, who are the clutch players in the league? 70 00:04:43,940 --> 00:04:47,768 Clutch players are the people who get the ball in the last second, and then shoot 71 00:04:47,768 --> 00:04:51,024 the ball. And everybody thinks that they are doing 72 00:04:51,024 --> 00:04:55,160 fantastic job under stress. And what was interesting is that all these 73 00:04:55,160 --> 00:04:59,000 coaches basically agreed very well about who were the clutch players. 74 00:04:59,000 --> 00:05:04,312 They named 17 people, and the agreement was virtually 100% about who those 17 75 00:05:04,312 --> 00:05:08,587 people are. And then we went back and we looked at the 76 00:05:08,587 --> 00:05:13,126 whole season of basketball games. And we looked at what was the performance 77 00:05:13,126 --> 00:05:17,004 of these clutch players? So, we looked at how many points these 78 00:05:17,004 --> 00:05:21,740 clutch players em, eh, got in the last five minutes of the game versus the last 79 00:05:21,740 --> 00:05:25,660 five minutes of the first half. Two five-minute periods. 80 00:05:25,660 --> 00:05:30,242 One is crucial, one less important. And then we also looked at the regular 81 00:05:30,242 --> 00:05:32,967 team members. So, what do you think happened? 82 00:05:32,968 --> 00:05:37,040 Did the clutch players have more points in the last five minutes of the game compared 83 00:05:37,040 --> 00:05:39,117 to the last five minutes of the first half? 84 00:05:39,118 --> 00:05:42,742 The answer is yes. And did the team members get more points 85 00:05:42,742 --> 00:05:46,834 in the last five minutes compared to the last five minutes of the first half? 86 00:05:46,834 --> 00:05:50,275 No. So the first look it seems like regular 87 00:05:50,275 --> 00:05:55,945 people, nonclutch players have a flat effect from the 5 minutes at the end of 88 00:05:55,945 --> 00:06:01,197 the first half to the last. And clutch players have an improvement. 89 00:06:01,198 --> 00:06:06,522 But there two ways you could get more points, you can get the ball passed to you 90 00:06:06,522 --> 00:06:09,937 more, you can try more often, but just do the same. 91 00:06:09,937 --> 00:06:13,387 Or you can actually improve your percentage average. 92 00:06:13,388 --> 00:06:17,489 Which one do you think they did? Did they just try more frequently or did 93 00:06:17,489 --> 00:06:21,368 they just improve their point average, or a little bit of both? 94 00:06:21,368 --> 00:06:24,942 Well, it's just that they tried more frequently. 95 00:06:24,942 --> 00:06:28,262 There was no higher success in terms of percentage. 96 00:06:28,262 --> 00:06:30,967 Other people seem to think the will try to play him. 97 00:06:30,967 --> 00:06:33,860 They passed him the ball. They think that they're a clutch player. 98 00:06:33,860 --> 00:06:37,094 They try more frequently. Success is just the same. 99 00:06:37,094 --> 00:06:41,504 Now, these are good players. They're fantastic players, but they don't 100 00:06:41,504 --> 00:06:45,371 have this improvement when you look at percentage score. 101 00:06:45,371 --> 00:06:49,596 By the way, when we ask coaches, coaches believe that they actually have an 102 00:06:49,596 --> 00:06:53,002 improvement in percentage, not just in absolute numbers. 103 00:06:53,002 --> 00:06:57,082 So, what happens is we have a team. And we all think that I'm the clutch 104 00:06:57,082 --> 00:06:59,737 player. So, therefore, you pass me the ball, I 105 00:06:59,737 --> 00:07:04,702 take more shots, everybody's happy. There's a self-fulfilling prophecy, in 106 00:07:04,702 --> 00:07:08,967 which I act on our belief, but I actually don't get any better. 107 00:07:08,968 --> 00:07:11,103 Now, you couldn't argue, you could argue against it. 108 00:07:11,103 --> 00:07:15,450 And you could say you know what? These clutch players are really getting 109 00:07:15,450 --> 00:07:19,020 better, but what happen is the defense also knows about. 110 00:07:19,020 --> 00:07:22,616 So, now if I'm the clutch player, everybody passes to me, the defense knows 111 00:07:22,616 --> 00:07:25,557 about it and they're going to protect me to a higher degree. 112 00:07:25,558 --> 00:07:30,260 So, the fact I don't succeed more is not because I didn't get better, it's because 113 00:07:30,260 --> 00:07:33,454 the defense was ready for it and they got better. 114 00:07:33,455 --> 00:07:36,857 So, we looked at that as well. We looked at free throws. 115 00:07:36,857 --> 00:07:41,209 And now there's no role for defense, you're just standing there and would you 116 00:07:41,209 --> 00:07:42,691 do more? So what happened? 117 00:07:42,691 --> 00:07:46,129 In the last five minutes, the clutch players get more points? 118 00:07:46,130 --> 00:07:50,594 Absolutely. Do they get better percentage? 119 00:07:50,594 --> 00:07:53,990 Not at all. They get the ball passed. 120 00:07:53,990 --> 00:07:56,956 They try more frequently. They get far more frequently, they go to 121 00:07:56,956 --> 00:08:00,218 the line more often. They take more shots, but do they do any 122 00:08:00,218 --> 00:08:03,522 better? Only in an absolute way, not in the 123 00:08:03,522 --> 00:08:04,618 perceptive.