News & Events

Print/PDF this page:

Print Friendly and PDF

Share this page:

Are Public Commitments Counterproductive?—A Commentary by Ian Ayres ’86

The following commentary was posted on newyorktimes.com on October 28, 2010.

Are Public Commitments Counterproductive?
By Ian Ayres ’86

The music entrepreneur Derek Sivers gave a TED Talk with the provocative claim that you’re more likely to reach your goals if you keep them secret:

http://freakonomics.blogs.nytimes.com/2010/10/28/are-public-commitments-counterproductive/  

Sivers claims that the benefits of secret goal-setting are proven by hard social science testing and “social reality” theory. The idea he presents is that when we announce a goal, our mind is tricked into thinking that we’ve already achieved it and so we are less likely to do the work that is required to actually accomplish the goal. As he memorably puts it, “Your mind mistakes the talking for doing.” You can read more about his thinking in his blog post on the subject.

I’m attracted to provocative claims as much as the next person. But this talk is reckless and at times just plain wrong. Sivers tells us in the talk that his “Keep Your Goals to Yourself” evidence “goes against the conventional wisdom that we should tell our friends our goals right so that they hold it to us, err, hold us to it, yeah.”

His inability to get out the last bit at the end, I think, might be a Freudian slip. You see, the experiments that he is alluding to don’t test whether “telling our friends our goals so that they can hold us to the goals” works or not. In fact, the experiments cited are not about announcing your intentions to your friends, but about whether or not your written intentions are noticed by an experimenter. And most crucially they are not testing whether “announcing so that your friends can hold you to your goals” works. If your friends are going to have any shot of holding you to your goals, they must not only know your goals, but they must know later on whether or not you achieved them. This subsequent feedback is crucial, and it is crucially missing from the experiments upon which he relies.

About a minute and 50 seconds into his Ted talk, Sivers describes a 2009 experiment conducted by Peter Gollwitzer (also available here) that provides the primary and most recent experimental evidence for his thesis. Here is how Sivers describes the experiment:

163 people across four separate tests — everyone wrote down their personal goal. Half of them announced their commitment to this goal to the room, and half didn’t. Then everyone was given 45 minutes of work that would directly lead them towards their goal, but they were told that they could stop at any time. Now, those who kept their mouths shut worked the entire 45 minutes, on average, and when asked afterwards, said that they felt that they had a long way to go still to achieve their goal. But those who had announced it quit after only 33 minutes, on average, and when asked afterwards, said that they felt much closer to achieving their goal.

To be frank, there are many discrepancies between Sivers’s description and the published description of experiments in this paper. (For example, Sivers claims in his talk that there were 163 subjects, and in his post that there were 63 subjects). But I’m pretty sure his talk is describing the second experiment, where there just 32 subjects. Here’s how Gollwitzer (and his coauthors) describe the same experiment:

[Law students at a German university were] informed that the experiment consisted of two independent parts. The first was introduced as an assessment of students’ willingness to intensify their study of law. Participants were asked to answer a four-page questionnaire. On the first page, the following critical intention item was presented: ‘‘I intend to make the best possible use of educational opportunities in law.’’ Participants responded on a 9-point scale ranging from 1, definitely not, to 9, definitely yes. In the social-reality condition, after a participant completed the questionnaire, the experimenter looked at this item and asked whether the number circled on the answer scale was the one the participant actually wanted to circle. Then the experimenter dropped the questionnaire into a box. In the no-social-reality condition, participants were simply asked to drop the questionnaire into a prepared box.

There are important structural differences between these two descriptions. Sivers says that half the subjects “announced their commitment . . . to the room,” but in the study there is no oral announcement to the room. Instead, some of the subjects were asked whether their written level of commitment (which, by the way, might have been low) was correct.

In the second part of the experiment, all of the law students were asked to help the experimenter develop a computer-based study package for law students. Gollwitzer, et al. wrote:

Participants were asked to help her find which cases to select for the package by trying hard to solve each case. The students were given 45 min to work on the prepared cases (plus the time needed to finish the case they were working on when the time limit was reached), but they were told that they could finish earlier if they wished. The time participants spent working on these cases was used to assess how successfully participants translated their intention into behavior. (emphasis added)

The last sentence is bizarre. Using time spent as a measure of success is not the way I normally grade my law students. One might have instead assessed how many of the cases the subjects were able to “solve.” Good students often turn in exams early.

The authors did find that students whose level of commitment had been noticed worked fewer minutes than those whose level of commitment had gone unnoticed. But even here there are unexplained discrepancies between the TED talk description and the article. Sivers claims the difference was 45 minutes for the non-announcing group and 33 minutes for the announcing group. But the article reports 45.65 minutes for the unnoticed group versus 41.52 minutes for the group whose level of commitment was noticed. This four minute differential was still statistically significant but not nearly as dramatic as the 12 minute differential reported by Sivers. (And for the life of me, I cannot find in the article any evidence that subjects were asked after the study about how close they felt they were to achieving their goals.)

(I emailed both Sivers and Gollwitzer to try to find out if there is somehow another experiment that reports the kinds of information contained in the TED talk, and Sivers very nicely sent me several attachments upon which he relied. But I have still not been able to identify an experiment that comes close to matching the claims of his talk. I have some concern that Sivers is simply referring to another experiment — even though he links to the same paper in his blog post on the subject and emailed me the paper as one of his attachments. At a minimum, my difficulty in easily finding the underlying study suggests that TED.com would be well-advised to start including related hyperlinks to the background material supporting the talk. And if the numbers of the Gollwitzer study are in fact misstated, it suggests that TED should start fact-checking the claims of their speakers.)

I’m most concerned that the Gollwitzer study is not well-designed to support Sivers’s claim that we shouldn’t tell our friends our goal so that they hold us to it. In the Gollwitzer experiment, there was no “telling,” there were “no friends” and there was no subsequent “holding us to it.” For goal-telling to be effective, it is crucial that your friends subsequently learn whether or not you succeeded. It is the potential embarrassment of not following through combined with the potential pride in reaching your goal that provides the key ingredient.

Justin Wolfers definitely understood this when he publicly committed on this blog to running the Stockholm Marathon:

I’m going to publicly declare my major fitness goal on this blog, and rely on this blog’s readers to ridicule me if I fail. So, here goes: this summer, I’m going to be visiting the IIES at Stockholm University, and on the last day of my visit, I’m planning on running the Stockholm Marathon. And I hope that you, dear reader, will keep me honest. It would be embarrassing to fail publicly, and I suspect it would be embarrassing enough that today’s public statement of my running goals will keep my future self pretty darn motivated.

It shouldn’t surprise you that his goal-telling as commitment worked. Justin finished it in 4 hours and 23 minutes.

The Gollwitzer experiments might suggest that telling people about goals may not be productive if the audience is never going to find out whether you succeed. So if you tell a friend who is about to leave for Antarctica that you plan to say a prayer every evening before you go to sleep, you may be less likely to follow through than if you had kept your prayer goal private. But if you tell your priest and 5 members of the vestry that you are committing to attending church, you are much more likely to find yourself in the pew next Sunday. The mere possibility that they will notice that you are not following through on your commitment is sufficient to add the extra layer of accountability.

The idea of making a financial commitment to go to church is not a hypothetical. On New Year’s 2009, Justin Noble, a public health student in Toronto, designed a commitment contract to go to church once a week. As I say in my book, Justin had successfully used stickK before to get to work on time. To make sure his commitment to go to church stuck, he backed it up with all three accountability devices that the site has to offer. He put down $600, and he signed up his girlfriend as both a supporter and a referee to make sure he followed through. The contract seemed to work like a charm. Every week, Justin faithfully reported to us that he had succeeded. But when I called Justin to ask him about his experience with stickK, he confessed that he had “cheated” in some of his reports to stickK. Even though the commitment at least indirectly implicates a higher all-seeing referee, Justin (with his girlfriend/referee’s complicity) was willing to send in false reports.

Justin’s story also underscores the importance of designating multiple supporters for your commitment. It’s easier to conspire with your girlfriend than with an entire congregation. I was thinking of Justin when I said a church-going commitment is more likely to succeed if you designate several members of the vestry as supporters.

I used to think that having a referee was the most important compliment to a financial stake. But the stories I’ve uncovered of referees letting their friends slide make me think that, contra to Sivers, having multiple supporters may be as important. When it comes to commitments, there’s safety in numbers.

Indeed, on the same New Year’s morning when Justin was committing to go to church, Andy Mayer, an executive at a Fortune 100 company who focuses on consumer behavior, was signing up with stickK to lose weight. Andy put $1,500 at risk and committed to lose a pound each week for 20 weeks. He also composed an email to send to his friends and family telling them about his weight loss commitment. “I had no trouble signing up for the contract. That was the easy part. The hardest part for me was telling people. . . . That e-mail sat open for several hours. It is one thing to enter into a contract where you and your spouse know what you are going to do, but I knew in my head that the social norming part – the social comparison part – the voyeuristic part was what was gonna make a difference.” Just before dinner, he sent the email and, true to form, Andy came through with flying colors. I caught up with him five months later when he was still riding the high that comes with successfully completing a difficult commitment – in Andy’s case losing just about exactly 10 percent of his body weight.

Success stories like Andy’s are one reason I think it is so dangerous for Sivers to say that it is better to keep goals secret. As Andy recently wrote in his blog:

I told my parents, sisters, in-laws and co-workers about the commitment. I invited all of them to register at Stickk.com and watch my weight loss. Telling them was a lot harder than putting the $1,500 at risk. Ian talks about this social aspect a bit in the book (see page 183), but I can’t underscore it enough. Getting fined and getting embarrassed socially for missing a commitment is a powerful one-two combo.

My commitment actually took one step more than Ian documents in the book. About three weeks into the diet, I presented on loss aversion and other aspects of behavioral economics at a meeting of my company’s top 120 executives. During the Q&A session after the presentation, a colleague publicly “outed” me and my commitment in front of all these folks as an example. In retrospect, I couldn’t have asked for a better incentive.

Andy had a lot more on the line then just money. If he falsely reported to stickK that he was losing weight, his friends would probably call him on it. The possibility of that embarrassment kept Andy honest and on the path toward success.

Now, in some cases, making a commitment public may undermine some of the benefits of success. It might feel like less of an accomplishment if it is publicly known that you had to use extrinsic encouragement to succeed. In Carrots and Sticks, I also write about a woman who used a stickK contract to commit to call her grandmother. It might make her calls less meaningful if grandma learned that the calls were made in part to avoid losing money. People might at times accept a lower probability of success in return for a larger feeling of accomplishment if they do succeed. But that rationale for privacy is very different than Sivers’s claim that publicity reduces the probability of achieving the goal.

One of the more candid moments in Jim Gray’s “The Decision” interview with Lebron James was when Gray out of the blue asked:

Are you still a nail biter?

Notice how the question aggressively assumes that at one time James was biting his nails. The lawyer in me was hoping James would respond “Objection! Assumes a fact not in evidence.” But instead James seemed to own up to the habit, at least in the past. James answered:

I have a little bit. Not of late.

Sivers would have you think that if James wanted to stop chewing his nails, he would be best advised to keep that goal to himself. But I say to you he would be more likely to reach his goal if he publicly announced his goal to stop biting and challenged the press and public to give him grief if he failed.