1 00:00:00,652 --> 00:00:06,862 So what we thought, was that maybe what's going on here, and there's a fudge factor, 2 00:00:06,862 --> 00:00:12,722 that people try to balance two forces. On one hand, we want to look at the mirror 3 00:00:12,722 --> 00:00:18,863 and feel that we're good honest wonderful moral people, that's the psychological 4 00:00:18,863 --> 00:00:24,977 benefit of viewing ourself positively. On the other hand, we want to benefit from 5 00:00:24,977 --> 00:00:30,080 cheating, that's the economic incentive. Now usually you would say, you can have 6 00:00:30,080 --> 00:00:32,749 one or the other. How can you both think of yourself as a 7 00:00:32,749 --> 00:00:36,180 wonderful honest person, and benefit from cheating at the same time? 8 00:00:36,180 --> 00:00:39,772 It doesn't work. Well, maybe, maybe due to our flexible 9 00:00:39,772 --> 00:00:45,148 cognitive ability, maybe due to the fact that we can rationalize things very 10 00:00:45,148 --> 00:00:50,776 quickly, as long as we cheat just a little bit, we can both benefit from cheating, 11 00:00:50,776 --> 00:00:55,610 just a little bit, and we can still view ourself as honorable people. 12 00:00:55,610 --> 00:01:00,018 And that's what we call the fudge factor. That the moment we start creating a 13 00:01:00,018 --> 00:01:04,698 dishonest behavior, there's a range before it invokes a different concept of 14 00:01:04,698 --> 00:01:07,746 ourselves. And in that range, we could both be 15 00:01:07,746 --> 00:01:11,617 dishonest, and still think of our self as honest people. 16 00:01:11,618 --> 00:01:17,550 So how do we test this kind of a theory? The first thing we wanted to do, is we 17 00:01:17,550 --> 00:01:21,080 wanted to take some approaches and shrink the fudge factor. 18 00:01:21,080 --> 00:01:26,136 We want to get people to cheat less by, but not by manipulating their economic 19 00:01:26,136 --> 00:01:29,256 variables. And the first idea for this came from a 20 00:01:29,256 --> 00:01:31,824 little joke. And the joke is that a guy goes to the 21 00:01:31,824 --> 00:01:35,982 Rabbi and he said, Rabbi you wouldn't believe what happened to me, somebody 22 00:01:35,982 --> 00:01:39,656 stole my bicycle from Synagogue. And the Rabbi is appalled, he say I can't 23 00:01:39,656 --> 00:01:42,260 believe this is happening, this is awful, this is terrible. 24 00:01:42,260 --> 00:01:46,050 I tell you what you do, come to synagogue next week and sit in the front row. 25 00:01:46,050 --> 00:01:51,528 And as we go over the Ten Commandments, turn around, and look at everybody in the 26 00:01:51,528 --> 00:01:54,304 aisles. And when we get to Thou Shalt Not Steal, 27 00:01:54,304 --> 00:01:57,992 see who can't look you in the eyes, you'll know that's your thief. 28 00:01:57,993 --> 00:02:01,067 The guy's very excited about the advice. The rabbi is very excited about the 29 00:02:01,067 --> 00:02:03,412 advice. He comes to synagogue, sit in the front 30 00:02:03,412 --> 00:02:05,915 row, turns around during the Ten Commandments. 31 00:02:05,915 --> 00:02:08,408 He comes to the Rabbi at the end of the service. 32 00:02:08,408 --> 00:02:13,745 He says, "Rabbi, you wouldn't believe it. What a fantastic idea, it worked just like 33 00:02:13,745 --> 00:02:17,927 a charm. The moment, the moment we got to "Thou 34 00:02:17,927 --> 00:02:23,957 shall not commit adultery" I remembered where I left my bike. 35 00:02:23,957 --> 00:02:27,240 Now where, [laugh] where is this joke leading us? 36 00:02:27,240 --> 00:02:31,310 You probably think to yourself where is the experiment in this thing? 37 00:02:31,310 --> 00:02:35,918 Well, we went to UCLA, and we asked about 500 students to try and recall the ten 38 00:02:35,918 --> 00:02:38,576 commandments. Interestingly by the way, none of them 39 00:02:38,576 --> 00:02:42,074 could recall all ten commandments and quite a few of them invented new ones, but 40 00:02:42,074 --> 00:02:47,924 that's a topic for a different discussion. But after getting them to try and recall 41 00:02:47,924 --> 00:02:51,376 the ten commandments we gave them the same math task. 42 00:02:51,376 --> 00:02:55,938 And we were wondering how much would they cheat after just thinking about the ten 43 00:02:55,938 --> 00:02:57,489 commandments. What did we see? 44 00:02:57,489 --> 00:03:01,438 Zero cheating. Now it wasn't as if the people remembered 45 00:03:01,438 --> 00:03:05,425 moral commandments. Presuming the more religious people 46 00:03:05,425 --> 00:03:09,110 cheated less, and the people who could hardly remember any commandment cheated 47 00:03:09,110 --> 00:03:11,938 more? No, nobody cheated. 48 00:03:11,938 --> 00:03:16,774 In fact even when we take self-declared atheist, and we asked them to swear on the 49 00:03:16,774 --> 00:03:21,382 Bible, and then we give them a chance to steal from us in the same way in the same 50 00:03:21,382 --> 00:03:26,850 task, we see no cheating whatsoever. So this suggests that it's not about 51 00:03:26,850 --> 00:03:30,945 cost-benefit, it's not about people thinking about heaven or hell and this is 52 00:03:30,945 --> 00:03:35,136 why they should be honest or not. It's about reminding people about their 53 00:03:35,136 --> 00:03:39,759 own morality, because even if you remind people about morality that is not their 54 00:03:39,759 --> 00:03:43,662 own, like atheists are being reminded about the Ten Commandments. 55 00:03:43,662 --> 00:03:48,422 Being, thinking, just thinking about morality, shrinks our fudge factor, get us 56 00:03:48,422 --> 00:03:52,902 to be a bit more careful about our own behavior and therefore allowing us to be 57 00:03:52,902 --> 00:03:56,665 more, more honest. Now, eh, this was kind of an interesting 58 00:03:56,665 --> 00:03:59,700 idea. But you know religion and civic society 59 00:03:59,700 --> 00:04:05,220 and there's all kinds of issues whether we want to involve religion in many, in many 60 00:04:05,220 --> 00:04:10,180 places, so the next thing we decided to do was to try a secular version of it. 61 00:04:10,180 --> 00:04:13,230 And at the university it's very natural to use the honor code. 62 00:04:13,230 --> 00:04:16,560 So we did this model at MIT and at Yale. We went to the students. 63 00:04:16,560 --> 00:04:19,860 And we edit one sentence to the beginning of the test. 64 00:04:19,860 --> 00:04:24,496 I understand that this short survey falls under the MIT or Yale honor code 65 00:04:24,496 --> 00:04:27,476 signature. What happened? 66 00:04:27,476 --> 00:04:32,357 People signed, they did the test, they shredded, no cheating whatsoever. 67 00:04:32,358 --> 00:04:37,188 And no cheating whatsoever despite the fact that neither MIT nor Yale actually 68 00:04:37,188 --> 00:04:40,780 have an honor code. Now this means that those two universities 69 00:04:40,780 --> 00:04:44,064 don't have specific costs associated with cheating. 70 00:04:44,065 --> 00:04:49,851 But nevertheless, being reminded about some moral element decreased cheating so 71 00:04:49,851 --> 00:04:54,075 dramatically we couldn't detect any. Now in contrast, we went to another 72 00:04:54,075 --> 00:04:56,220 university that has a very strong honor code. 73 00:04:56,220 --> 00:04:59,150 We went to Princeton. Princeton has a very strong honor code. 74 00:04:59,150 --> 00:05:05,463 They take the freshmen and they give them a, a week long, a lesson on honesty. 75 00:05:05,463 --> 00:05:09,694 There's lectures and discussion, the a capella group has a song about it, it's, 76 00:05:09,694 --> 00:05:12,102 the whole week is a crash course on morality. 77 00:05:12,102 --> 00:05:16,400 We waited two weeks, after these freshmen finished with their crash course on 78 00:05:16,400 --> 00:05:20,822 morality, and we tempted them to cheat in the same way, some of them with signing, 79 00:05:20,822 --> 00:05:24,361 some of them without signing. What do you think happened? 80 00:05:24,361 --> 00:05:27,840 Would the Princeton student be different than the Yale, and MIT student? 81 00:05:27,840 --> 00:05:30,550 They were not. When they did not sign the code, they 82 00:05:30,550 --> 00:05:33,180 cheated just like the MIT and Yale student. 83 00:05:33,180 --> 00:05:37,982 When they signed, they didn't cheat just like the MIT and Yale students. 84 00:05:37,982 --> 00:05:42,228 So this is kind of a good news, bad news. The bad news, is that a week-long crash 85 00:05:42,228 --> 00:05:47,940 course on morality doesn't seem to have any measurable effects two weeks down the 86 00:05:47,940 --> 00:05:51,320 road, at least not the Princeton version of it. 87 00:05:51,320 --> 00:05:57,638 And I think the same thing applies to many corporate lessons and, university, classes 88 00:05:57,638 --> 00:06:01,680 and so on, in morality. Morality training just doesn't seem to be 89 00:06:01,680 --> 00:06:06,052 that effective, that's the bad news. The good news, is that reminding people 90 00:06:06,052 --> 00:06:10,532 about their own morality, whether it has consequences or not when, like Princeton, 91 00:06:10,532 --> 00:06:15,012 it has really huge penalty if you're being caught cheating versus MIT and Yale, that 92 00:06:15,012 --> 00:06:18,252 you don't, that seemed to actually make a big difference. 93 00:06:18,252 --> 00:06:22,352 So rather than thinking about global cultural changes, that might be very nice 94 00:06:22,352 --> 00:06:25,884 but very hard to achieve. Maybe what we need to focus on, is to 95 00:06:25,884 --> 00:06:31,060 think about how we change things just before people are tempted to be dishonest. 96 00:06:31,060 --> 00:06:35,941 In the next version of this we tried one more thing, we tried some people signed 97 00:06:35,941 --> 00:06:40,257 first, before they did the test, and some people signed afterward. 98 00:06:40,258 --> 00:06:43,440 What did we find? If you signed first, people don't cheat. 99 00:06:43,440 --> 00:06:47,235 If you sign at the end, people cheat just a lot, because it's finished, you finished 100 00:06:47,235 --> 00:06:50,127 cheating. And so with this idea we went to the IRS 101 00:06:50,127 --> 00:06:53,407 and we said to the IRS think about all the IRS forms. 102 00:06:53,408 --> 00:06:56,456 Right. People finish the IRS form and then at the 103 00:06:56,456 --> 00:07:00,898 end, many pages later you sign. At that point your signature is not doing 104 00:07:00,898 --> 00:07:05,228 anything to prime morality. And nobody's going to say, oh, yes, let me 105 00:07:05,228 --> 00:07:09,150 sign, let me go back and fix things. No, cheating is over by that time. 106 00:07:09,150 --> 00:07:12,882 The form is being finished. So we said to the IRS, why don't you put 107 00:07:12,882 --> 00:07:16,040 the signature up front? Why don't you start by signing? 108 00:07:16,040 --> 00:07:20,804 Eh, well the people at the IRS said this is illegal, because signature is meant for 109 00:07:20,804 --> 00:07:25,500 verification, not for priming morality. And you can actually see how signature has 110 00:07:25,500 --> 00:07:29,217 these two very different functions, if it's in the front or in the back. 111 00:07:29,218 --> 00:07:33,300 So we said okay, we understand this. Why don't you put both the signature in 112 00:07:33,300 --> 00:07:37,814 the front and the signature in the back? Well, they said, the people at the IRS 113 00:07:37,814 --> 00:07:41,872 said, well, that's confusing. I don't know if you've seen the IRS forms 114 00:07:41,872 --> 00:07:46,142 recently, or the IRS tax a code, but you know the idea that signing twice is 115 00:07:46,142 --> 00:07:49,788 confusing, baffled me a little bit. I was speechless. 116 00:07:49,788 --> 00:07:54,508 But then I said okay, if that's confusing how about if the first item on the tax 117 00:07:54,508 --> 00:07:59,814 return, would ask people would you give $25 to a task force to fight corruption. 118 00:07:59,815 --> 00:08:04,030 And we said this is not just a signature, this is asking people whether they want to 119 00:08:04,030 --> 00:08:07,476 put money towards something. And if people commit, and commit money 120 00:08:07,476 --> 00:08:11,823 towards something, there's a good chance that later on, they will continue acting 121 00:08:11,823 --> 00:08:14,367 in a more, eh, appropriate, honest, behavior. 122 00:08:14,367 --> 00:08:18,088 Plus, we said, the people that don't want to give money a task force to fight 123 00:08:18,088 --> 00:08:21,231 corruption, maybe you can audit them, to a higher degree. 124 00:08:21,232 --> 00:08:23,530 Anyway, the IRS did not want to talk to us anymore. 125 00:08:23,530 --> 00:08:26,302 But we went to talk to an insurance company. 126 00:08:26,302 --> 00:08:30,446 Interesting, this was insurance company that insures, eh, homes and houses, lots 127 00:08:30,446 --> 00:08:33,379 of personal things. And they told us, that they think that 128 00:08:33,379 --> 00:08:37,181 their cheating, that they are seeing, is very much like in our experiments. 129 00:08:37,181 --> 00:08:42,434 They think that very few people invent burglaries, and arsenies, and larceny, and 130 00:08:42,434 --> 00:08:47,070 car theft, but lots and lots and lots of people, when something bad happen 131 00:08:47,070 --> 00:08:51,548 exaggerated by 10, 15%. And they estimated that these exaggeration 132 00:08:51,548 --> 00:08:56,180 cost industry about $24 billion, a year which is lots and lots of money. 133 00:08:56,180 --> 00:08:59,393 I know we through lots of big numbers these days, but 24 billion is still a big 134 00:08:59,393 --> 00:09:03,927 number. So we came up with all kinds of ideas. 135 00:09:03,928 --> 00:09:08,764 We said, what if when people were waiting on hold, you would play religious music 136 00:09:08,764 --> 00:09:11,518 for them? What would happen if you would put images 137 00:09:11,518 --> 00:09:14,120 of eyes, as if they're being watched, on the form? 138 00:09:14,120 --> 00:09:18,644 What if you got them to sign first? What if you got some declaration that had 139 00:09:18,644 --> 00:09:22,925 to do something with religion as well? We had all kind of ideas about how you 140 00:09:22,925 --> 00:09:26,630 increase, em, honesty. But you know, we had lots of good ideals 141 00:09:26,630 --> 00:09:31,229 and they had lawyers and when you have good ideas and lawyers together in the 142 00:09:31,229 --> 00:09:35,610 same room nothing ever happen. So after six months of trying very hard to 143 00:09:35,610 --> 00:09:40,760 get this, nothing was done. But my contact in this insurance company 144 00:09:40,760 --> 00:09:44,086 felt very bad for me. So he came to me and he gave me this form, 145 00:09:44,086 --> 00:09:48,370 where you mail it to people and you ask them to report how many miles you drove 146 00:09:48,370 --> 00:09:50,932 last year. Now imagine you got one of those form, 147 00:09:50,932 --> 00:09:53,588 what would you want to shade the number, up or down? 148 00:09:53,588 --> 00:09:57,116 If you shade it up, it would mean your premium would go up, if you shade it down 149 00:09:57,116 --> 00:10:00,709 your premium will go down, so you'll probably want to shade it down. 150 00:10:00,710 --> 00:10:05,512 They gave us 20,000 forms and we did one thing, we moved the signature line to the 151 00:10:05,512 --> 00:10:08,640 top. Half the people got the signature on the 152 00:10:08,640 --> 00:10:12,221 top first and half the people got it at the bottom. 153 00:10:12,221 --> 00:10:15,800 Some people signed first, and then filled the mileage, some people fill the mileage 154 00:10:15,800 --> 00:10:17,660 then sign. What did we see? 155 00:10:17,660 --> 00:10:22,060 There was a difference of about 2400 miles on average. 156 00:10:22,060 --> 00:10:27,040 The people who signed first reported driving 2400 miles more. 157 00:10:27,040 --> 00:10:31,584 Now we don't know if they didn't cheat at all, but we know the cheating was much 158 00:10:31,584 --> 00:10:34,465 reduced. And that actually gives me great hope that 159 00:10:34,465 --> 00:10:38,689 we can create those small interventions and get them to work not just in the lab, 160 00:10:38,689 --> 00:10:42,222 but get them in the field in all kinds of other versions like that. 161 00:10:42,222 --> 00:10:46,138 Okay, so all that was about how we shrink the fudge factor. 162 00:10:46,138 --> 00:10:50,761 How do we get people to, at the moment, tolerate less of their own immortality and 163 00:10:50,761 --> 00:10:56,452 be more careful about what they're doing? The next question we wanted to ask, was 164 00:10:56,452 --> 00:11:00,047 how do you get people to cheat more, how do you expand the fudge factor? 165 00:11:00,047 --> 00:11:04,118 And of course, this is not an objective to increase cheating in society, but for the 166 00:11:04,118 --> 00:11:08,189 purpose of understanding it, what kind of things would get people to cheat more and 167 00:11:08,189 --> 00:11:10,707 nevertheless feel comfortable with themselves. 168 00:11:10,708 --> 00:11:12,580 And the idea here also came from a little joke. 169 00:11:12,580 --> 00:11:16,935 And the little joke is, that little Johnny comes home from school with a note from 170 00:11:16,935 --> 00:11:20,965 the teacher that says that little Johnny stole a pencil from the kid who was 171 00:11:20,965 --> 00:11:24,103 sitting next to him. And Johnny's father is furious and he 172 00:11:24,103 --> 00:11:27,260 said, Johnny I'm humiliated and embarrassed and I'm angry. 173 00:11:27,260 --> 00:11:31,080 And you're grounded for three weeks and just wait until your mother comes. 174 00:11:31,080 --> 00:11:34,849 You never, never, never steal a pencil from the kid who is sitting next to you. 175 00:11:34,850 --> 00:11:39,000 And beside Johnny if you need something, if you need a pencil you know very well 176 00:11:39,000 --> 00:11:43,591 you can just say something. You can just ask and I will bring you 177 00:11:43,591 --> 00:11:48,310 dozens of pencils from work. If you find this funny it's probably 178 00:11:48,310 --> 00:11:52,798 because you realize that taking a pencil from work feels very differently than 179 00:11:52,798 --> 00:11:57,354 taking 10 cents from petty cash box, even if you were going to use those 10 cents to 180 00:11:57,354 --> 00:12:01,596 buy yourself a pencil. Perhaps the moment something becomes 181 00:12:01,596 --> 00:12:05,958 multiple steps removed from money, there are more stories we can tell ourselves. 182 00:12:05,958 --> 00:12:08,885 Right, you could say the workplace wants us to have these pencils, otherwise they 183 00:12:08,885 --> 00:12:11,724 wouldn't put them there. Or you could say other people take them, 184 00:12:11,724 --> 00:12:13,920 as well. Or you could say, even if I take them and 185 00:12:13,920 --> 00:12:17,440 give them to my kids, they will leave me alone and that would allow me to answer 186 00:12:17,440 --> 00:12:20,230 two more emails. Good for the workplace. 187 00:12:20,230 --> 00:12:25,351 And all of those distance allow us to tell ourselves stories, allow us to rationalize 188 00:12:25,351 --> 00:12:30,319 things that basically create this ability to cheat more, but nevertheless to feel 189 00:12:30,319 --> 00:12:34,738 better about ourselves. So the first experiment we did on this, is 190 00:12:34,738 --> 00:12:40,349 I walked around campus and I put six packs of Cokes, in some common refrigerators. 191 00:12:40,350 --> 00:12:43,492 Now these were refrigerators that everybody had their own stuff, but 192 00:12:43,492 --> 00:12:46,200 multiple people had their stuff in the same refrigerator. 193 00:12:46,200 --> 00:12:50,560 And I basically wanted to see what is the half lifetime of Coke. 194 00:12:50,560 --> 00:12:53,130 How long will it take for my Cokes to disappear? 195 00:12:53,130 --> 00:12:58,489 And basically within 72 hours no Cokes were left there, everything was gone. 196 00:12:58,490 --> 00:13:03,770 Now in contrast, I took a plate with six $1 bills, I put in the same refrigerators 197 00:13:03,770 --> 00:13:06,785 and I waited. And now what happened, nobody touched a 198 00:13:06,785 --> 00:13:09,864 single dollar. Now the students could have taken a 199 00:13:09,864 --> 00:13:15,032 dollar, putting it in the nearby vending machine, get coke and change, but nobody 200 00:13:15,032 --> 00:13:17,412 did. When something is directly about money, we 201 00:13:17,412 --> 00:13:22,168 feel very differently about it. When something is removed from money, more 202 00:13:22,168 --> 00:13:27,160 abstract, there are some stories we can tell ourselves, rationalizing and, 203 00:13:27,160 --> 00:13:31,650 therefore, we can be more dishonest. Now this is not an ideal experiment, 204 00:13:31,650 --> 00:13:35,790 because it's often the case that you open the refrigerator and you find Coke. 205 00:13:35,790 --> 00:13:39,255 It's not so common that you open the refrigerator and find a plate with six $1 206 00:13:39,255 --> 00:13:41,590 bills, so something must have been very strange. 207 00:13:41,590 --> 00:13:46,920 So we did the same experiment, the same principle with the same matrix test and 208 00:13:46,920 --> 00:13:51,530 here's the experiment. Some people got the sheet, solved it, came 209 00:13:51,530 --> 00:13:56,822 to the experimenter, looked him in the eyes and said, Mister experimenter I 210 00:13:56,822 --> 00:14:01,039 solved x problems, give me x dollars. So they looked him in the eyes and they 211 00:14:01,039 --> 00:14:04,892 cheated for dollars. The second group did the same thing, but 212 00:14:04,892 --> 00:14:09,244 when they came to the experiment, the guy looked him in the eyes, and they said, "Mr 213 00:14:09,244 --> 00:14:12,106 Experimenter, I solved x problems; give me x tokens. 214 00:14:12,106 --> 00:14:16,201 So they were lying not for money, they were lying for something that is tokens. 215 00:14:16,201 --> 00:14:19,916 And we gave them tokens. They took 12 feet to the side; they walked 216 00:14:19,916 --> 00:14:23,980 12 feet to the side, and they replaced these tokens for money. 217 00:14:23,980 --> 00:14:28,877 Now they were lying for a pencil, but it was a pencil for only a few seconds. 218 00:14:28,878 --> 00:14:32,902 It was one step removed from money, but for a very, very short duration. 219 00:14:32,903 --> 00:14:35,920 What happened? Our participant doubled their cheating in 220 00:14:35,920 --> 00:14:39,390 that condition. And by the way this is for me one of the 221 00:14:39,390 --> 00:14:41,846 most worrisome experiments we've ever done. 222 00:14:41,846 --> 00:14:45,746 Because if you think about it, we're moving to becoming a cashless society, as 223 00:14:45,746 --> 00:14:49,084 consumers there's all these discussions about digital wallets. 224 00:14:49,084 --> 00:14:53,415 On Wall Street there's of course all these complex financial tools that are becoming 225 00:14:53,415 --> 00:14:57,390 more steps removed from money. People own more and more things like stock 226 00:14:57,390 --> 00:14:59,640 options and things that are removed from money. 227 00:14:59,640 --> 00:15:03,910 And if being close to money and observing it directly is what's keeping us 228 00:15:03,910 --> 00:15:08,670 relatively honest, is it the case that in all kinds of domains of life, electronic 229 00:15:08,670 --> 00:15:13,500 wallets, financial markets, the more this distance is increasing, the more we can 230 00:15:13,500 --> 00:15:17,910 allow ourselves to both be dishonest and think of our self as hon-, as honest 231 00:15:17,910 --> 00:15:21,198 people. And that really, really worries me. 232 00:15:21,199 --> 00:15:26,357 So that was the first thing that got people to cheat more, distance from money. 233 00:15:26,358 --> 00:15:30,594 The next thing we wondered, is we wondered what happened when you watched other 234 00:15:30,594 --> 00:15:34,938 people cheat in an egregious way? After all we open the newspaper everyday 235 00:15:34,938 --> 00:15:38,025 and we see all kinds of stories about big cheaters. 236 00:15:38,025 --> 00:15:42,610 What is the affect of observing other people in society does to our own 237 00:15:42,610 --> 00:15:45,220 morality? So we repeated this same experiment with 238 00:15:45,220 --> 00:15:48,410 two changes. The first one was that we prepaid people, 239 00:15:48,410 --> 00:15:53,417 everybody got an envelope with all the money as if they solved all the problems. 240 00:15:53,418 --> 00:15:57,212 And we asked them at the end, to pay us back the money they didn't make and to 241 00:15:57,212 --> 00:16:01,140 take only the money they deserve. That was the first change. 242 00:16:01,140 --> 00:16:04,310 The second change was that we hired an acting student. 243 00:16:04,310 --> 00:16:08,888 That acting student sat in the front seat in the, in the classroom. 244 00:16:08,888 --> 00:16:13,554 And 30 seconds into the experiment, he raised his hand and said, excuse me, I 245 00:16:13,554 --> 00:16:18,090 solved everything, what do I do now? Now imagine you were in that experiment. 246 00:16:18,090 --> 00:16:21,402 You're solving, you're working very hard, you're still on problem number one, and 247 00:16:21,402 --> 00:16:24,587 all of a sudden you see somebody who claimed that they've solved everything. 248 00:16:24,588 --> 00:16:27,487 You know for sure, without cheating, there's no other way. 249 00:16:27,488 --> 00:16:30,180 And the experimenter said, "You've solved everything? 250 00:16:30,180 --> 00:16:33,776 You're free to go." And you see that person taking the envelope with all the 251 00:16:33,776 --> 00:16:38,052 money and leaving the room. What would happen now to your own 252 00:16:38,052 --> 00:16:41,120 morality? Lots and lots of moral people cheat. 253 00:16:41,120 --> 00:16:44,700 Now there's two ways to interpret this result. 254 00:16:44,700 --> 00:16:48,220 One is to say this is about cost/benefit analysis. 255 00:16:48,220 --> 00:16:51,919 Before you saw that guy cheating in an egregious way, you thought to yourself, I 256 00:16:51,919 --> 00:16:54,060 don't know what's the cost of cheating here. 257 00:16:54,060 --> 00:16:56,540 Maybe they will catch me. Maybe somebody would run after me. 258 00:16:56,540 --> 00:17:00,892 But after you saw a demonstration of somebody cheating in this egregious way 259 00:17:00,892 --> 00:17:04,900 and nobody going after them, you say there's no cost here let me cheat. 260 00:17:04,900 --> 00:17:09,661 The other possibility, is that it's not about the cost benefit analysis, but its 261 00:17:09,661 --> 00:17:13,676 about what you call social proof. Its about giving people a demonstration 262 00:17:13,676 --> 00:17:17,924 that people like them, from their society, from their community, students at the same 263 00:17:17,924 --> 00:17:20,694 university are cheating in such circumstance. 264 00:17:20,695 --> 00:17:24,916 And maybe because of that it is okay to cheat in those things, not from a moral 265 00:17:24,916 --> 00:17:29,010 perspective, but from a social agreement persepctive this is the behavior. 266 00:17:29,010 --> 00:17:33,962 Here is someone who has done it and is not even feeling sorry about that. 267 00:17:33,962 --> 00:17:37,630 As an analogy you can think about what if everybody around you is downin-, 268 00:17:37,630 --> 00:17:41,590 downloading illegal content, at some point, do you start saying, okay, this is 269 00:17:41,590 --> 00:17:46,180 just, I guess what is okay to do. To separate those two things, we tried 270 00:17:46,180 --> 00:17:48,300 these two explanations, we tried two things. 271 00:17:48,300 --> 00:17:52,640 The first one, is that instead of the student cheating, we asked him to just 272 00:17:52,640 --> 00:17:57,050 raise his hand and said excuse me, but with this set of instructions, can't I 273 00:17:57,050 --> 00:18:00,857 just say I solved everything and take all the money and go home? 274 00:18:00,858 --> 00:18:04,359 And the instructor said you know what? You can, it's up to you. 275 00:18:04,360 --> 00:18:09,180 So now it was clear that from a cost benefit analysis you can cheat and get 276 00:18:09,180 --> 00:18:13,433 away with it, but there was no example of somebody actually doing it. 277 00:18:13,433 --> 00:18:18,144 What happened now when we did this? Cheating did not go up it actually went 278 00:18:18,144 --> 00:18:20,554 down. We did another version of this and in the 279 00:18:20,554 --> 00:18:23,956 other version we changed the outfit of this acting student. 280 00:18:23,957 --> 00:18:27,820 We ran this experiment at Carnegie Mellon. All the students were Carnegie Mellon. 281 00:18:27,820 --> 00:18:29,760 The acting student was a Carnegie Mellon student. 282 00:18:29,760 --> 00:18:34,460 They were all from that group. In the second experiment, we ran this, eh, 283 00:18:34,460 --> 00:18:39,948 version, but we dressed the acting student with University of Pittsburgh outfit. 284 00:18:39,948 --> 00:18:44,440 Now, when they stood up and say, excuse me, I solved everything, and walked home 285 00:18:44,440 --> 00:18:49,064 with all the money, you would still know that it's the same cost benefit analysis, 286 00:18:49,064 --> 00:18:53,348 that now in this experiment you're not going to be punished, but there was no 287 00:18:53,348 --> 00:18:56,266 social proof. It wasn't as if somebody from your social 288 00:18:56,266 --> 00:19:00,160 group, from your friends, from your community is cheating this way, it's those 289 00:19:00,160 --> 00:19:03,500 other people from this other university that we don't like so much. 290 00:19:03,500 --> 00:19:06,170 It's those people who are immoral. What happened now? 291 00:19:06,170 --> 00:19:10,703 Cheating did not go up, it went down. And basically I think that two things 292 00:19:10,703 --> 00:19:13,416 happened. The first one, was that people were 293 00:19:13,416 --> 00:19:18,448 reminded about morality and honesty by seeing an egregious act of misbehavior, a 294 00:19:18,448 --> 00:19:21,471 little bit like the ten commandments version. 295 00:19:21,472 --> 00:19:25,634 But it wasn't their group, it was what we called an out group, it was somebody else 296 00:19:25,634 --> 00:19:28,643 behaving this way. And therefore they did not take it as a 297 00:19:28,643 --> 00:19:31,242 signal that it's okay for them to do this as well. 298 00:19:31,243 --> 00:19:34,430 So, so far we found two things that make people cheat more. 299 00:19:34,430 --> 00:19:38,522 We found that the distance for money and we found the social acceptance, kind of 300 00:19:38,522 --> 00:19:40,980 seeing other people behave in egregious way. 301 00:19:40,980 --> 00:19:46,524 But we also wanted to find out what are personality traits, which are the people 302 00:19:46,524 --> 00:19:50,100 that would cheat more? Now you could speculate a lot of things. 303 00:19:50,100 --> 00:19:53,935 You can say maybe the people who are more egocentric or maybe the people who are 304 00:19:53,935 --> 00:19:56,410 more competitive, maybe the people hate losing. 305 00:19:56,410 --> 00:20:00,638 You can speculate on a lot of things, but if you think about our theory. 306 00:20:00,638 --> 00:20:04,800 Our theory is that people want to view themselves as honest people and also want 307 00:20:04,800 --> 00:20:08,030 to benefit from cheating. We want to have the cake and eat it too. 308 00:20:08,030 --> 00:20:11,670 We want both of those things. Who are the people that could do that? 309 00:20:11,670 --> 00:20:14,669 And what we thought was that maybe those are creative people. 310 00:20:14,670 --> 00:20:19,347 Because if it's all about rationalization, if this kind of cheating we discover is 311 00:20:19,347 --> 00:20:23,190 all about being able to cheat and rationalize it, so it doesn't affect our 312 00:20:23,190 --> 00:20:27,540 own image, people who are more creative presumably can just tell better stories. 313 00:20:27,540 --> 00:20:32,921 They can cheat more and rationalize it to an even higher degree, so we tried that. 314 00:20:32,921 --> 00:20:37,224 We tried to measure people creativity and measure how much they cheat, and we found 315 00:20:37,224 --> 00:20:41,575 that more creative people cheat more. We tried to increase creativity with some 316 00:20:41,575 --> 00:20:45,220 creativity enhancing exercises. And we found that as we were increase 317 00:20:45,220 --> 00:20:48,419 creativity slightly, we also got cheating to increase slightly. 318 00:20:49,590 --> 00:20:54,079 We also went to a big advertising agency, and we checked the moral flexibility of 319 00:20:54,079 --> 00:20:57,190 people in different jobs in that advertising agency. 320 00:20:57,190 --> 00:21:01,100 And we saw that as the job required more creativity, people cheated more. 321 00:21:01,100 --> 00:21:04,812 There was less cheating in accounting, accounting much more cheating in the 322 00:21:04,812 --> 00:21:08,101 creative professions. And, you know, maybe it is good that 323 00:21:08,101 --> 00:21:13,000 there's less cheating in accounting, but it clearly supports this theory that our 324 00:21:13,000 --> 00:21:16,460 version of cheating, it's all about this rationalizing. 325 00:21:16,460 --> 00:21:21,220 It's about acting badly and rationalizing, and the question is to what extent can we 326 00:21:21,220 --> 00:21:23,803 do it. If we're multiple distance, multiple steps 327 00:21:23,803 --> 00:21:26,320 removed from money, we can do it to a higher degree. 328 00:21:26,320 --> 00:21:29,984 If people around us are doing it more, we can do it to a higher degree. 329 00:21:29,984 --> 00:21:33,923 If we ourselves are better storytellers, more creative storytellers, we can do it 330 00:21:33,923 --> 00:21:37,296 to a higher degree. All of those things are contributing to a 331 00:21:37,296 --> 00:21:40,704 very high degree. So what have we learned up to this point? 332 00:21:40,704 --> 00:21:45,376 I think if we think about the Becker model, that cheating is basically driven 333 00:21:45,376 --> 00:21:49,632 by how much to I stand gain? What's the probability of being caught? 334 00:21:49,632 --> 00:21:54,168 What kind of punishment will I get? And doing some cost benefit analysis. 335 00:21:54,168 --> 00:22:00,344 That doesn't explain much of our behavior. It's also not the case that we have bad 336 00:22:00,344 --> 00:22:02,937 apples. If you think about what is real cheating 337 00:22:02,937 --> 00:22:05,329 due to? In Enron, and other cases, is it due to 338 00:22:05,329 --> 00:22:09,197 bad apples, or is it due to lots and lots of people cheating a little bit? 339 00:22:09,198 --> 00:22:13,089 I think it's much more due to lots and lots of people, cheating a little bit. 340 00:22:13,090 --> 00:22:17,567 And specifically we learned a few things that either hinder it or increase it. 341 00:22:17,568 --> 00:22:22,860 So we learned that getting people to think more carefully about their morality 342 00:22:22,860 --> 00:22:27,347 shrinks the fudge factor and gets people to behave more honestly. 343 00:22:27,348 --> 00:22:29,576 We also learned that it's all about rationalization. 344 00:22:29,576 --> 00:22:33,939 So, when we have something that is multiple steps removed from money, we can 345 00:22:33,939 --> 00:22:36,733 cheat more, and at the same time rationalize it. 346 00:22:36,733 --> 00:22:41,576 When you see other people around us behaving in immoral ways, we can cheat a 347 00:22:41,576 --> 00:22:45,765 bit more, and rationalizing it. And when we ourselves are creative, we can 348 00:22:45,765 --> 00:22:49,832 rationalize, we can tell better stories, cheat more, and feel good about ourselves. 349 00:22:49,832 --> 00:22:54,359 And I think that if we have this perspective, it means we should not direct 350 00:22:54,359 --> 00:22:57,651 much of our attention to the cost-benefit analysis. 351 00:22:57,652 --> 00:23:02,925 We should not think about something like, how do we expose people, and how do we 352 00:23:02,925 --> 00:23:06,382 punish them, and how do we, eh, surveillance them? 353 00:23:06,382 --> 00:23:10,332 Because all of those approaches are driven to this idea that people are just doing 354 00:23:10,332 --> 00:23:13,891 the cost-benefit analysis. And if you say no, most of the cheating is 355 00:23:13,891 --> 00:23:18,000 not about that, it's about what people can feel comfortable with themselves. 356 00:23:18,000 --> 00:23:20,958 Now we have to ask with different approaches, how do we get people to be 357 00:23:20,958 --> 00:23:24,559 comfortable with less cheating? How do we get people to become aware of 358 00:23:24,559 --> 00:23:28,110 their own traps? How do we create rules that don't allow 359 00:23:28,110 --> 00:23:31,313 this flexibility in terms of rationalization? 360 00:23:31,313 --> 00:23:34,573 And I think this will be the actual Improvement. 361 00:23:34,574 --> 00:23:39,341 You know, across all of our experiments, we ran about 30,000 people, we found about 362 00:23:39,341 --> 00:23:43,189 12 people that cheated a lot, and we lost about $150 to these guys. 363 00:23:43,189 --> 00:23:48,859 We found about 18,000 people who cheated just a little bit, and we lost about 364 00:23:48,859 --> 00:23:52,444 $36,000 to those people. And I think that this is probably a good 365 00:23:52,444 --> 00:23:55,890 reflection of what's happening in society. Sure, there are bad apples. 366 00:23:55,890 --> 00:23:58,974 There are people who steal, and so on. And it's terrible. 367 00:23:58,974 --> 00:24:03,549 But economically, it's a small potato, these are small things in terms of 368 00:24:03,549 --> 00:24:06,885 economic impact. The big economic impact comes from lots of 369 00:24:06,885 --> 00:24:11,043 people cheating a little bit, but because there are so many of them the economic 370 00:24:11,043 --> 00:24:14,514 impact is much higher. And I don't think we need to stop thinking 371 00:24:14,514 --> 00:24:18,804 about the crooks, but I think we need to pay more attention to the people who can 372 00:24:18,804 --> 00:24:21,982 change just a little bit and feel good about themselves. 373 00:24:21,982 --> 00:24:31,483 Now, I think if we do it, everybody would be better off.