What Does Science Suggest About Self-Control Early in Life as a Predictor for Future Prospects?

Science Suggest about Self-Control as Predictor for FutureShutterstock

In the late 1960s, Walter Mischel tempted children with marshmallows: have one now, he offered, or wait to get two. The study is one of the select few that has entered popular consciousness, a success helped by Youtube clips that show children squirming as they try to maintain their resolve to wait. Age makes a big difference. Almost all three year-olds fail, whereas most six year-olds succeed. The interesting period falls in between. With four year-olds, Mischel found wide variation. Some children managed to wait successfully whilst others succumbed.

What made the experiments particularly significant was the ensuing long study. Mischel and his colleagues tracked down the original subjects and found that the ability of a four year-old to hold out—to “delay gratification” in his phrase—is a powerful predictor of later success. The ability to wait at age four is correlated with, amongst other things, higher SAT scores, less use of addictive drugs, and less obesity. Other studies suggest that it is a better predictor of academic success than IQ.

A second group of studies have shed light on the mechanism involved. Willpower, it seems, can be depleted. Test someone’s willpower with one task—resisting the chocolate chip cookies that are supposedly reserved for another experiment—and they will have less to expend on other unrelated tasks—persisting with insoluble puzzles or in holding a handgrip exerciser. In the conception of the experimenters involved, willpower is like a muscle. Exercise tires it in the short-run, though perhaps—here the studies are less clear—it can help to build it up in the long run.

Put the two literatures together and the picture can easily emerge of life as a tug-of-war against temptation: whilst training might help a little, it is those endowed with the largest self-control muscles who are able to pull the hardest. Diagnose a child’s endowment with the marshmallow test, and we have an excellent predictor of how they will do in later life.

However, as the researchers themselves would be the first to admit, the real picture is much more complicated than that. Even if the muscle analogy is right, technique and motivation can be as important as brute strength. More fundamentally, other research suggests that the kind of muscular self-control deployed in the marshmallow task is a last resort. It is much better to structure things so that temptation does not arise, either by engineering the world so that temptations are not encountered, or, more securely, by engineering one’s own mind so that what is encountered does not tempt.

Let’s take these possibilities in turn. Mischel himself found that much of the explanation of success in the marshmallow task came from technique. Those who successfully waited were those who were able to distract themselves. Showing people the rewards for succumbing makes them more likely to succumb; but surprisingly, showing them the rewards for waiting also makes them more likely to succumb. The only effective technique is to show them neither. Subsequent research has shed further light on what happens here. When faced with temptation the children’s judgments tend to shift: they tend to devalue the reward for waiting. Two marshmallows are judged hardly better than one. So if a resolution is to be effective, it had better not be reopened in the face of the temptation, lest a new decision is formed that the reward is not worth the wait. Studies with adult dieters confirm the point. Forming a simple intention not to take dessert is surprisingly effective: faced with the temptation the subjects simply repeat to themselves their resolution and move on. But get them to add a rationale for the resolution—get them to say to themselves “I am not taking dessert because …” and then  fill out their reason—and the effectiveness is lost. Adding the rationale reopens the question, and then it is all too easy to tell a story as to why the resolution should be abandoned, at least for today: one serving of dessert is not going to make all the difference, and beside, it’s been a tough day, and a serving would cheer me up, and tomorrow would be an excellent time to start a diet.

In contrast, if agents can form resolutions that do not require extensive deliberation—ideally they will be simply cued by the circumstances, like refusing dessert when it is offered, or going to the gym when the clock strikes 6.00—then they will be much more likely to stick with them. In fact, experiments indicate that sticking with such resolutions does not deplete agents’ willpower: it does not mean that there is less to be used elsewhere.  And this suggests that sticking to such resolutions isn’t using will power at all. Willpower is needed once deliberation is open. If that can be avoided by a implementation of a prior resolution, especially one that has attained the force of habit, then the use of willpower can be avoided. Agents can avoid temptation if they can restructure their intentions so that tempting options are, for them, out of the question.

How important is this? Clearly there are many cases in which people do have to rely on their willpower: after all, Mischel did find that there were many areas in which those with greater willpower do better. But in focusing on these areas we may be missing even more important areas in which we have learned not to need it.

Aristotle held that there are two ways in which a person can do the right thing. They can be continent, in which case they feel the pull of behaving badly, but they resist. Or they can be truly virtuous, in which case they no longer feel the pull of the bad. As we have seen, the contemporary psychological literature suggests that Aristotle was onto an important distinction here. But in applying it to moral behavior, Aristotle was applying it to an area that is only just being empirically explored.

On Aristotle’s picture, moral education results in the inculcation of certain habits. These are not simply habits of behavior; they are also habits of thought. Agents acquire a moral sensibility which, for the virtuous, renders certain actions unthinkable. To take an example from Bernard Williams: we might discuss various ways to respond to the threat posed by a political opponent, but, for most of us at least, the idea of having them killed is simply off the cards. We don’t contemplate it and then reject it on the grounds that it would be morally wrong. It never gets onto the agenda in the first place, and so it never poses a real temptation. The same goes for more mundane cases. The idea of cheating on an important exam, or on a spouse, is, for many people, simply out of the question. They don’t need willpower to resist, since the temptation is never real.

Contemporary psychological opinion is still divided on the importance of moral education: we do not know how much is learned and how much is innate. But the existence of different moral standards in different societies shows that learning has a role. Central is the idea of moralization. Take the case of vegetarianism. People become vegetarians for different reasons. For some it is primarily a health issue; for others it is moral. Studies suggest that for those who see it as moral issue, the idea of eating meat slowly moves off the agenda. Even if they were initially convinced by highly abstract arguments, meat typically comes to be seen as disgusting, and eating it comes to be unthinkable. Such a shift is far less marked for those who avoid meat for health reasons.

It is early days for research in this area, but perhaps these findings will help to reconcile one of the deepest divisions in accounts of ethics: between those who see moral responses as, at base, emotional, and those who see them as rational. Rather than taking these accounts as rivals, we can see them as complementary. Rational moral conviction can give rise, over time, to an associated emotional response. What better way to keep an option off the table, than to find it disgusting?

In sum then, self-control is a many faceted thing. We know something about how to test children’s ability to resist a temptation that they feel. But we know very little about how they can come to treat an option as unthinkable. Is this a skill that they have early in life that remains largely unchanged? Or is it instead something that can develop over a lifetime? That remains to be seen.

Discussion Questions:

1. Can we show self-control when we’re not even tempted?

2.  Wouldn’t even a virtuous person be tempted if we forced the temptation under their nose strongly enough? Doesn’t that show that options are never really off the table?

3. How plausible is it that our emotional reactions line up with our moral judgments? Don’t we often find that we are emotionally repelled by an action that we theoretically judge to be morally acceptable? Or are emotionally attracted to an action that we theoretically judge unacceptable?

14 Responses

  1. Lisa PZ says:

    If “virtuous” is, by definition, “no longer feeling the pull of the bad,” and if the virtuous may even come to be “disgusted” by the immoral/rejected behavior or object (as in the vegetariansm example), it would seem that just bring the temptation close to the virtuous person would never result in a greater “pull” of temptation. In fact, the closer it is to being “under their nose” the more repelled they might be.

    Maybe a related question is simply: Can this ever happen? Can something that people once found created pleasure actually be transformed into an absent feeling, or even disgust? The case of vegetariansim seems to suggest it is possible. (And my personal experience with sugar also leads me to believe it is possible — although I never thought of it as a “moral” issue really).

    With regard to the second part of the question, about options “never being off the table,” I do wonder if similar processes that underlie the moralization of certain behaviors/acts might be reversed over time. Thus in that respect, (im)moralized options would only “be off the table” until processes took place to bring them back to a point where they are tempting again. Bandura’s work on moral disengagement (where by things that normally do morally disgust us, can be disengaged from so that they no longer do) seems especially relevant to this reverse process.

    • Richard Holton says:

      I didn’t want to suggest that the process of moralization always brings a disgust response. It may bring other emotional responses; and perhaps sometimes options will be seen as ‘off the table’ even when there is no emotional buttressing at all. Plausibly that is typically what happens with sexual fidelity: there I take it there isn’t (normally!) a disgust response to the idea of infidelity even when the possibility is off the table, and perhaps (though this would need some investigation) there is no aversive emotional response to it at all. I think that is why, in such cases, we are concerned when a option is pressed on someone when they wouldn’t have sought it out: that often looks like entrapment.

      Even when we do have aversive emotional reactions, they can coexist with others that attract. Tony Dickinson has a lovely example of being both attracted to, and repelled by, watermelon (he had, without realizing, acquired a conditioned aversion to it as a result of drinking too much after eating it). Similarly, sexual desire can manifest both: consider repressed homosexual desire in a homophobic society. So even when moralization does bring a set of aversive emotional responses to an option, I don’t think that it’s obvious that being confronted with that option won’t sometimes make it tempting again.

      The question of when and where moralization can be ‘undone’ is a really fascinating one. We might indeed think of Bandura’s work in that way, along with work on ‘act identification’: when people do bad things they often try to reconstrue them in a light that makes this less obvious (Baumeister’s book on evil has discussion of lots of cases). Alternatively, we might think that this largely leaves the moralization in place, but provides a way of sidestepping it. I’d love to know more about quite what is going on here.

      • Lisa PZ says:

        I find it fascinating too. I also wasn’t meaning to say that moralization always = disgust. Only that it does seem possible that in some cases at least, bringing the temptation closer could repel rather than attract. But I agree that in other cases full-force temptation might be met with neutral feelings or a compelling “pull” toward them. I guess the question then is, what are the conditions under and mechanisms by which common temptations have or lose power?

        With regard to some of the mechanisms underlying moralization (and possibly normalization), I wonder if some of the same mechanisms that underlie the development of automaticity could be involved (i.e., when something is learned to the point where one no longer needs to think about it, and actually, by that point, thinking about it can actually hinder performance).

        But relating back to self-control, do you have a preferred definition of it? It really does seem like actively resisting a temptation is a different “thing” than not feeling tempted (but both look like self-control). Perhaps in parallel, learning a new song on the violin might feel difficult, but once it is learned it may feel easy and automatic. Successfully “controlling the self” to engage in the target behavior perhaps requires somewhat different strategies in each case.

        • Richard Holton says:

          It does seem very plausible that laying down automatic habits will have some role in building good moral practice. Even here though things are complex. The standard, and effective, treatment for overcoming unwanted habits (tics and the like) is awareness therapy. Once people are aware of what they are doing they find it surprisingly easy to stop. One would hope that with moral behavior things are a little more robust than that: one would hope that simply becoming aware of them isn’t enough to enable us to overcome them.

          I don’t really have a definition of self-control (I don’t really have a definition of any interesting notion!) but I do think of it in a much broader way than I once did, very much along the lines you suggest. I used to think of it just in terms of overcoming temptation, and that is how it is used in much of the literature, both in psychology and philosophy. But now I think that the more interesting idea concerns one’s general ability to get oneself to act in the way that is, or perhaps that one takes to be, right (and spelling that out raises a host of philosophical difficulties.) That includes, amongst many other skills, getting oneself into a state in which one doesn’t feel tempted to behave badly in the first place. We need to disentangle the various phenomena here. Which we end up calling ‘self-control’ then strikes me as of secondary importance.

  2. George Gantz says:

    Thanks for the fascinating article. 

    I was struck by the idea that thinking about consequences (whether the positive or the negative) undermines willpower.  That helps explains the research findings by Eric Schwitchgebel (see http://swedenborgcenterconcord.org/the-moral-behavior-of-ethics-professors/).  He reports that the moral behavior of ethics professors is less moral than other professors.  Apparently if you think too much about morality, it undermines your moral determination.

    More significantly, you have highlighted the important question as to whether willpower, and the virtues associated with it, can be taught.  Based on my own experience as a father, grandfather and volunteer leader in the Boy Scouts, I am convinced the answer is “of course” – but the difficulty is in knowing what methods work and what methods do not.  For example, the self-esteem movement was premised on the idea that affirmation would lead to esteem which would lead to achievement, something which research has now debunked (see, for example: http://psi.sagepub.com/content/4/1/1.short).  Too much praise can encourage narcissism.  Too much self-expression on Facebook can do the same.

    By many accounts, one of the most successful of such teaching methods would seem to be a religious upbringing.  Baumeister and Tierney in their book on Willpower, and Jonathan Haidt in his recent book, The Righteous Mind, (all of whom admit to being secular agnostics) provide compelling evidence that religious people exhibit better self control.  It is easy to see why –  a belief in an all-seeing God and eternal judgment can provide a strong “do not even think about it” motivation.

  3. Richard Holton says:

    It’s a very interesting question whether self-control can be taught. There’s been a lot of work on it (for instance by Angela Duckworth at Penn) but the results have been very mixed. Most of this has focussed on building self-control understood in the Mischel way, rather than on teaching children to get to the state in which they won’t need it.

  4. Meyer1953 says:

    Your essay points out a rather encouraging discovery.  Neither a single marshmallow nor two marshmallows, displayed, encouraged a longer waiting time before succumbing to gratification.  But leaving the decision unruffled by any display, resulted in delay of gratification.

    It seems that intention, left to operate intact, does a better job of fostering self-discipline than merely parading incentives about.  Coupling this with the observation that, those who delayed gratification had more success as adults, suggests that those who are more intentional in their basic characters do more specifically successful things in their lives than those who are more reactive.

    After all, which of the two types would a person want as his or her airline pilot in a “miracle on the Hudson” type crisis?  I suspect that the instant-gratification type would be curled up into a fetal ball, while the intentional pilot would be busy doing the heroic saving of lives.

    There is a vast audience for integrity.

    • Richard Holton says:

      There’s quite a lot of evidence that providing incentives is often not the way to engender the behavior that one wants. In a large series of experiments, Deci and Ryan provided excellent evidence that offering incentives frequently doesn’t result in people developing ‘intrinsic motivation’ for the behaviour that one is trying to encourage. On the contrary: remove the incentives and they are less likely to do it than they were before. Building integrity requires people to internalize certain values, and simple incentive based conditioning is a very blunt way of trying to do that.

  5. wondering14 says:

    Did all the marshmallow children like them equally? Can the prediction affect outcomes?

    So, in avoiding temptation, we should not be rational about it, because we cannot be successfully rational. How is that “simple intention” not to take a dessert, that is effective, formed? If not by reason, is it by following a moral dictate of someone of authority, a skull and bones drawing (to make something unthinkable or disgusting), a TV admonition of some sort? If one doesn’t try to think (rationally) about an action, how does one know if what one is asked to refuse or delay (or accept) is good or not?

    If “simple intention” is effective, then was the “Just Say NO” to sex program of a while back an effective way to delay teenage sex?

    How much can circumstances affect an outcome? Eat this marshmellow and you won’t have any for a month! Or a person has a mother to support, so he can’t eat out once a week to help support his mother.

    • Richard Holton says:

      I don’t think it’s right to say that to avoid temptation we shouldn’t be rational. The point is rather that we have to be rational about when we deploy our thought. For intentions to be effective, they do require serious commitment at the time that they are made, and that requires a lot of thought. Gabriele Oettingen has a large body of work showing that for maximum effectiveness subjects need to contrast the outcome of sticking with the intention with the outcome of not sticking with it; and then they need to think about where temptation is likely to be felt, and what they need to do to resist it. So there is a big place for rationality. It is just that the plan should be reopened at the moment of temptation.

      Premarital sexual abstinence pledges don’t seem to be very effective, at least once you control for the differences between the kinds of people who are likely to make them and the general population. That may well be because they are made in a very abstract way, without much thought as to where temptation is likely to be encountered, and how it should be effectively resisted. One indication of this is that those who have made such pledges are much less likely to use condoms when they do have sex than those who haven’t.

      • wondering14 says:

        Sorry, I remembered wrongly. The “just say no” program was of Nancy Reagan, saying no to drugs, not to sex. You would say, perhaps, that such a program might be as ineffective with drugs as with sex. You note a downside with regard to sex, that of being less likely to use condoms when young people violate their pledge, and do engage in sex. What might be the downside with regard to drugs?

        • Richard Holton says:

          It’s an interesting question whether prior resolutions would help in resisting drugs. Again they would need to be based on a genuine commitment, and not just be the result of pressure to make a pledge; and they would need to involve realistic plans about what to do if drugs are offered. As for analogues to condom use: you don’t want young people to think that once they’ve taken something, the game’s up, so it doesn’t matter what they do. You want them to think that having a few drinks if very different to getting completely drunk; that smoking a joint is very different to smoking many, and that again is very different to taking crack, and so on. The difficult question — and I don’t know of any research on this — is how much they can internalize these ideas whilst still thinking that all drug taking is ‘off the table’.

  6. ianful says:

    A child with self-control indicates a high level soul that is influencing the self. We are more likely to find such a child in a family that exercises a high degree of self-discipline. It follows that these children will succeed better in life. They will be less expectant of immediate or continual rewards and be less distracted and more resilient.

    The advertising fraternity will not like this, but they should be banned from children’s media as it messes with self-control.

    Fasting for a month in Lent or Ramadan is an efficient way of re-establishing self-control for adolescents and adults.

    • Richard Holton says:

      Michel did find that greater self-control was more common in higher income families, but what explains what here is still very unclear. And I don’t know of any research on whether religious practices, like fasting, have any effect, though it would be great to have some research here. As I said in an earlier post, it is very unclear how much self-control can be taught.