share interactive transcript request transcript/captions live captions
download
|
MyPlaylist
LAURENCE STEINBERG: So I want to talk today about a program of work that I've been involved in for the last 15 years or so on adolescent risk-taking. And I want to start by putting this in a public health context.
Adolescence typically is a very healthy period in the lifespan. There is not a great deal of illness or disease. And the major contributors to morbidity and mortality in adolescence are behavioral. And so just a few examples of this. Adolescents commit more crimes than adults. This is a very robust age pattern called the age crime curve, showing that crime peaks in the late adolescent years. You find this in any industrialized country in the world, and it's been very stable over time in the United States.
Adolescents, as you probably know, have more car crashes than adults. And it's been documented that this is not simply due to inexperience. If you compare 16-year-old drivers and 20-year-old drivers with the same level of experience-- that is, beginning drivers in either case-- you find that there are more car crashes among adolescents. And as you see, it decreases linearly with age.
One thing that's really interesting to me is that adolescents drown more often than people of other ages, which is very surprising given how strong adolescents are. So these are data from the CDC on unintentional drownings. And this, I think, suggests that probably the cause of the drownings is reckless behavior or bad decisions about where one is swimming rather than poor endurance or muscular weakness. Adolescents attempt suicide more often. Here are data, again, CDC data, on what are called non-fatal self-inflicted injuries, again showing a peak.
So most of these trends peak at around the same ages. And we can make a long list of things where risky and reckless behavior is more typical during adolescence than it is during adulthood. And as you probably know, adolescence is the time during which most people initiate experimentation with tobacco, alcohol, and illicit drugs. I think we're so accustomed to that as being the time when people first start using that that we don't think about the fact that it doesn't have to be adolescence. It could be at some other point in development.
So when I was here at Cornell and an undergraduate investor, we were taught a number of things about why adolescents made poor decisions, and they've turned out not to hold up very well to empirical study, in part because of the work that Val and her colleagues have done. Adolescents are not illogical. At least, they're no more illogical than adults are. By the time people are 15 or 16, their performance on tests of logical reasoning are comparable to those seen in samples of adults. They don't underestimate risk. If you give people questionnaires, asking them how likely various risky things are to happen, adolescents, just as adults do, overestimate risk. But they overestimate it even more than adults do. So that can't explain their behavior.
We were taught that adolescents suffer from delusions of invulnerability or the personal fable. But studies have shown that we all suffer from delusions of invulnerability, and adolescents aren't more likely than adults are to suffer from them. Studies of information processing show that adolescents don't think poorly. They might make bad decisions, but they don't think badly. And their performance on various measures of information processing also improves until about age 15 or 16, and then it plateaus, and it doesn't really improve very much after that.
And they're not unaware of the dangers of risky behavior. Virtually all adolescents in industrialized countries receive some kind of health education in which they are told that they shouldn't smoke, drink while-- drive while they've been drinking, or have unprotected sex. And these questionnaires show that they are very much aware of the dangers associated with these, but they do these behaviors anyway.
So these are some data from a study I'm going to talk about later. This is a composite measure of different basic information processing abilities. And what we find here is what you see in many other studies, which is that there's a trajectory of improvement during the pre-adolescent, early adolescent years, but that it plateaus at around 15 or 16. So in this sense at least, adolescents are as smart as adults are by the time they're 15 or 16 years old. But remember, that peak in risk-taking occurs after that. It's around 16 or 17. So it can't be because they're not that smart.
And so one of the questions that has been driving this research for a long time is a question that perplexes parents, I think, who look at their kids who are able to do calculus in school and perform really well in difficult classes on history or biology or literature, or whatever, and yet in the real world, they do a lot of stupid things. And so we're trying to understand why that is. Because if we could figure that out, we could design better prevention programs that might reduce adolescent risk-taking and improve public health. So here is one of the latest stupid things that adolescents do. This is a poster taken from the New York City subway system, reminding people, mostly adolescents, that it is a bad idea to hold onto the outside of subway cars as they move from station to station.
So one thing about kids that we've discovered is that if you look at measures of psychosocial rather than cognitive development-- and this is a composite of things like impulse control, risk perception, resistance to peer pressure-- that you see a very different developmental trajectory than you do if you look at measures of cognitive development. And so you see steady improvement in psychosocial development, or maturation in the psychosocial realm, that continues well into the 20s.
And this has led us to posit that there is a kind of maturity gap, and that adolescents mature intellectually before they mature socially and emotionally. So along the vertical axis here is the percentage of individuals in a given age group who score at the median level of the adults in this sample, or of the oldest age group in this sample. And what you see is that by the time you get to age 16 or 17, 45% or so of teenagers are scoring at the median of the 26 to 30-year-olds on measures of cognitive development, but not on measures of psychosocial development. And that still is improving later through adolescence and young adulthood.
So we've posed an alternative view of adolescent risk-taking than the conventional one, which is that it's because of cognitive deficiencies or a lack of information. We don't think it's due to cognitive deficits. We think that it's the product of psychosocial rather than intellectual immaturity. We think it's normative for adolescents to take risks, and we think adolescent risk-taking is evolutionarily adaptive. So if you think for a moment about adolescence in other species, I mean, all mammals go through puberty. And even though we might want to call what they go through-- we might not want to call it adolescence, we can model certain aspects of human development by looking at this period in other mammalian species.
And what happens when juveniles in other species go through puberty is that they leave their natal environment and go out in search of mates. And this is very dangerous. This is dangerous because there are competitors and there are predators. And we believe that there is some hardwired inclination for individuals to be more risk-tolerant during a stage in which it's important to be able to take risks, and that this has been conserved across species.
I dug around in the literature a little bit to ask, when are people most fertile? And it turns out that the peak fecundity among females is about five years post-menarche. And in an industrialized society where menarche occurs at around age 12 or so, this would make the peak age of fecundity around 17, which is exactly the same as the peak age in most forms of risk-taking.
So I think we can tell a story here. Some of you might think of it as just-so story. But I think we can tell a story here that links the increase in risky behavior during adolescence to something that has been conserved and that has some evolutionarily significant adaptive active value.
So the questions that we've been asking in this program of research are, why is there an increase in risk-taking between childhood and adolescence, and why is there a decline between adolescence and adulthood? Some theories are good at addressing the first of these questions, and some are good at addressing the second, but there aren't a lot of theoretical models that address both of them. And then after I've finished showing you some data on those questions, I want to talk about, why is it that adolescent risk-taking is especially likely to occur in groups?
So we work from a dual systems model of adolescent risk-taking. It is a model that we developed in our lab at Temple, but it is not dissimilar from several other models that are out there in the literature that all kind of were arrived at at around the same time. And so we have been interested in two brain systems that mature during adolescence. One of them we call an incentive processing system. Some people might call it a Pavlovian system. It's linked to the valuation and prediction of rewards and punishments. It also happens to be very important for the processing of emotional and social information.
And for those of you who have a little neuroscience background, the most important regions for this are the ventral striatum, which is part of the limbic system, and the orbital frontal cortex. And what do we know about the maturation of this system during adolescence? It undergoes major changes in early and mid-adolescence around the time of puberty. We now know that some of the changes that take place in this system are actually directly due to the impact of gonadal hormones on the brain. Some of them seem to be coincident with puberty, and some of them seem to be both due to puberty and due to the organizational effects of hormones during prenatal and perinatal development.
And most importantly for our model is this increase in activity involving the neurotransmitter dopamine in this pathway between the limbic system and primarily the ventral striatum and the prefrontal cortex, and primarily the orbitofrontal and ventromedial area of the prefrontal cortex. So there is more dopaminergic activity in this part of the brain during adolescence than at any other point in human development. And this is important, because the ventral striatum is an area of the brain that's highly significant for our experience of reward or pleasure. And so one of the take-home messages here, which I'm sure some of you will be very unhappy to hear, is that nothing will ever feel as good to you for the rest of your life--
[LAUGHTER]
--as it did when you were an adolescent. And that's because there's much more dopamine activity in reward centers of the brain during this time than during other periods. And these changes in dopaminergic activity show an inverted U-shaped pattern with age, with a kind of increased proliferation of dopamine receptors during the period from before puberty to about mid-adolescence, and then a decline in dopamine receptors after that.
And these changes, we think, have some behavioral manifestations. They make individuals more attentive to rewards, because if rewards feel so much better, you're going to be paying more attention to where they are. They make individuals more likely to engage in sensation seeking and novelty seeking, because they're willing to do things to get these rewards. It's also the case that changes in the system lead to greater and easier emotional arousal. So there is some evidence that although adolescents aren't moodier than adults in the sense that their moods don't change more often, they have higher highs and lower lows than adults, in studies that look at-- experience sampling of emotional states during the course of the day.
And adolescents show increased attentiveness to social information, to the social cues put out by others, to their standing in a social hierarchy, to what other people think of them, and so on. And I think this helps us understand things like their susceptibility to peer influence, that we now are beginning to understand some of the neural underpinnings of that.
Now this model posits that adolescent risk-taking is the product of changes in two different systems. Remember, it's a dual systems model. And so the second system is what we refer to as a cognitive control system. This system is the system of the brain that's important for working memory, for logical reasoning, for planning, and for regulating impulses. And this is localized primarily in the lateral areas of the prefrontal cortex, so the dorsolateral prefrontal cortex, as well as the areas of the parietal cortex and the cingulate.
Now this system also changes during adolescence, but it changes in different ways, and along a different timetable than the incentive processing system. So we know that it matures gradually from pre-adolescence on, and continues to mature into the early and mid-20s. It has been hypothesized to be independent of puberty. There are some new studies that have come out in the last couple of years that suggest that it may not be completely independent of puberty, but we believe it's less dependent on puberty than changes in the incentive processing system, and actually a little bit more experience-dependent, which I think has important implications for intervention that I'll talk about later on.
And we have a pretty good understanding of the structural and functional changes that take place in regions of the cognitive control system during adolescence that involve both synaptic pruning and mylination. And importantly-- and I think over the next decade, you're going to start to see much more attention, not so much on the frontal lobe, which is what gets the popular attention now, but on connections between the frontal lobe and other brain regions. And we think that a lot of improvements in cognitive control during adolescence are not due just to changes in the structure and functioning of the prefrontal region, but in connections between brain regions. And scientists have been developing new ways of measuring both structural and functional connectivity between brain regions, and it's showing some very interesting patterns.
I should say as a detour for those of you who study neuroscience that I think this is one of the reasons that some of the findings in this literature are inconsistent as to whether what we see in this part of the brain during adolescence is more activation or less activation. And I don't think it's a matter of more or less. I think it's a matter of connectivity between this region and other brain regions. And these changes are thought to result in improvements in things like impulse control, coordination of emotion and cognition due primarily to better connectivity between limbic areas and cortical areas, more foresight, and more planning ahead.
But if there is a serious take-home message from this talk, it's that the differential timing of the developments of these systems is very important to understanding adolescent risk-taking, because the excitation of the incentive processing system occurs very early in adolescence, around the time of puberty and shortly thereafter. But the maturation of cognitive control is protracted and very gradual, and it's not complete until late adolescence or early adulthood. And so the metaphor that we've used in describing this is that the accelerator is activated before a good braking system is in place. And some people have talked about starting the engines without a skilled driver behind the wheel.
So to just summarize this introduction and to lay the groundwork for the research I'm going to talk about, we believe that reward seeking and self-regulation are subserved by different brain systems, and they mature along different timetables during adolescence. For those of you who do research in this area, there's a really important methodological implication that I want to just be explicit about. A lot of our measures that we use of sensation seeking and impulsivity are conflated. Sensation seeking and impulsivity are not the same thing. In fact, they're very, very different things.
And so if you're skeptical, just do the thought experiment of, imagine waiting in a really long line for a very long time to ride a really scary roller coaster, all right? So that's impulse control, self-regulation in the pursuit of sensation seeking. And these are not the same thing at all. And we now have a pretty good idea of the fact that they're undergirded by different neuro systems, as well as are conceptually distinct, and that they develop along different timetables.
What this means, though, is that middle adolescence should be a very vulnerable period for risky behavior, because this is the time when you have the greatest imbalance between an easily aroused incentive processing system and a still immature cognitive control system. I won't have time to talk about it in today's lecture, but middle adolescence is also a very vulnerable period for all kinds of psychiatric disorder. If you were to plot out the average age of onset for most major types of psychiatric problems, the average for most of these problems is somewhere in the middle adolescent years. And that includes most mood disorders, it includes eating disorders, it includes substance abuse disorders, it includes a wide range of disorders. And we believe and have written about the fact that we think that this vulnerability to mental health problems is also related to this maturational imbalance between the incentive processing and cognitive control systems.
And I'm going to argue later on today that this particular vulnerability is exacerbated by the presence of peers for the following reason. Peers have extremely high reward value during adolescence. And they further exaggerate this imbalance by adding extra activation to the inventive processing system because they have such reward value.
So our first study of this was done with funding from the MacArthur Foundation. And it was done mainly in the pursuit of our understanding of the determinants of adolescent criminal responsibility. As [INAUDIBLE] mentioned, we were very active in several of the Supreme Court cases involving whether adolescents are less responsible for their behavior than adults. But this research program is also informative about risky behavior as well.
So we wanted to look at age differences in capacities affecting judgment and decision making. And we had five data collection sites with a diverse, ethnically and socioeconomically, sample of 935 individuals between 10 and 30. That's a really wide age range for a cross-sectional study, and it was a challenge to come up with measures that seemed to be appropriate for people as young as 10 and as early as 30. And our measures involved a battery of computerized performance tests, as well as some standardized self-report measures.
So I just want to talk about these constructs today. Given the theoretical framework that I set up, we wanted to measure reward processing-- and I'll talk about these specific measures as I show you some findings-- and also to talk about cognitive control. And within each set, we have some behavioral measures and some self-report measures.
So we are in the field now with funding from the Jacobs Foundation. Yes, there was a lot of money associated with that prize. Unfortunately, or fortunately for the field, but unfortunately for me, I wasn't allowed to put it in my personal bank account. It had to be spent on research. But this money was invested in a cross-cultural replication of this.
Because as I began presenting findings from this program of research, a lot of people said, well, yeah. This is American adolescence. It's because of the way we raise kids in the United States. And so we've taken this into the field, and as you can see, some very, very different parts of the world. We will ultimately have samples of about 500 people in each of these countries. We have now-- we're now up to about 4,000 people in total, and the data come in every week from these different sites. And I'm going to show you what we see from peeking at some of the early-- at the early results. So I'm just going to show you the results from the six countries in which, at this point in time, we had enough people in each age cell to be able to talk about age differences.
And let's start by talking about reward processing, and start with the most simple version of this, which is simply to ask people about sensation seeking. So the results displayed here are the results of a composite measure. There's a sample item at the top of the slide. I sometimes like to do things that are a little frightening, and what you see is the predicted, inverted, U-shaped curve with an increase from pre-adolescence into mid-adolescence, and then a decline thereafter.
And here's what we're seeing so far in the cross-national sample. It There's a significant curvilinear effect, but the peak is later. We're not sure if this is going to hold, or why it is. But we are finding this increase and then decline. Now it could be that in some of these countries, especially the poorer countries, puberty occurs later. And if that's true, it should shift the peak in sensation seeking to a later point in development.
It's also the case that, as we're finding, that there are many larger cross-cultural differences in the way we treat people who are 18 and older than the way that we treat people who are 10 to 17. In most of these countries, people who are 10 to 17 go to school. They pretty much follow a schedule of adolescence that doesn't look that different than what you'd see in the United States. But after age 18, there are big cultural differences in what people do with their lives. And so we might expect to see more cultural variability at that age than before.
If we measure this by looking at something we call risk preference, which is, how do you compare the benefits of a risky activity with the risks, and we've asked this about a series of risky activities, we see a similar curvilinear pattern here. Again, increasing until middle adolescence, and then declining afterwards.
I don't have a slide to show you this, but if I just look at a question asking, how risky is this activity, we don't see age differences. We don't see very big ones. It depends on the particular item. But we see consistent age differences in risk preference. That is, in the extent to which they associate the risk with a reward. And here's what we're seeing in the cross-national sample. Again, an increase and then a decline. The peak is somewhat later. The decline is not as sharp as it is in the American sample.
Our behavioral measure of reward processing, one of them is the Iowa gambling task. Some of you may be familiar with it. Imagine that I present you with four decks of cards, face down, and I tell you that two of these decks are good decks, and two of these decks are bad decks. And when you turn over a card, you get information about whether you've won or lost, and how much you've won or lost. And you can portray this in terms of money or points or candy, depending upon the age of the sample that you're testing. And we administer this in three sets of 40 trials. And we tell you that what we want you to do is to turn over cards in a way that's going to maximize your winnings on this.
So you're shown a screen like this, and you're asked if you want to play or pass. This is a variation on the original development of the task, but it's an important variation, because it allows us to separate out sensitivity to good cards and sensitivity to bad cards, because we can measure whether people pass or not. And if you play, the card turns over, and you're given some feedback about what the outcome has been, and there's a running total at the bottom of the screen. If you pass, you don't learn anything at all about that deck.
Now the bad decks have larger gains, but they also have larger losses. And the ratio of losses to gains is calibrated so that if you keep pulling from those bad decks, you will lose over time. The good decks have much smaller gains, but they also have smaller losses. But the ratio is calibrated so that if you keep pulling from the good decks, you will win over time.
How many of you like to gamble? I do. So people who like to gamble like that feeling of hitting that blackjack, right? Even though you know when you're playing that you're going to lose if you stay at the table long enough, because that's how it's calibrated so that you will lose. But the feeling of winning every once in a while feels so good that you do it anyway.
And so when individuals perseverate in choosing the bad decks, what that indicates, or at least what it's thought to indicate, is that their decision making is disproportionately influenced by this potential for a really big reward. And so what we do in the Iowa gambling task is we track people's deck choice over time. And as I mentioned before, we can separate out choosing good decks from avoiding bad decks because of that pass or play option in here.
So if we do that, we see something really interesting. So this is, along the vertical axis, the difference between the number of cards chosen in the last trial versus the first trial. And lines that are above the midline there indicate an increase, and lines that are below the midline indicate a decrease. And the red cards indicate good decks, and the yellow cards indicate bed decks.
And what you see here, if you just focus on the red bars for a moment, is a pattern that we've now seen in sensation seeking and in risk preference, and it's this inverted U-shaped curve. And we see that as a measure of reward sensitivity. And the pattern of the yellow bars is a measure of people's avoidance over time of bad decks.
Now if you want to look at age differences here, let me suggest that you start at the right side of the chart, and look at the behavior of the oldest groups. They show a change in approaching rewarding decks over time that's about the same as their change in avoiding bad decks, right? I mean, the red bars go up, and the yellow bars go down.
Now go to the left side of the chart and look at the little kids, right? The only change in their behavior over time is in approaching the good cards. In fact, the youngest age group shows no change at all in their avoidance of the bad cards. So we think that this indicates a heightened reward sensitivity during adolescence relative to adults. It has important implications for how we discipline kids, I think. I think that younger adolescents in particular are much more reward-sensitive than they are punishment-sensitive And here's what we're seeing in the cross-national sample. Not exactly the same pattern, but pretty similar if you look at the youngest groups and the oldest groups, and you look at the general shape of the red bars and the general shape of the yellow bars.
A second reward processing task we use is a temporal discounting task or a delayed discounting task. So imagine that I've asked you if you'd rather have $200 today or $1,000 in six months. Depending upon how you answer it, I'll give you a second offer, which we adjust depending upon your answer. So if you were to say, I'd rather have the $1,000 in six months, I would say, well, how about $600 today? And then if you said, no, I'd rather have the $600, then I say, well, how about $400 today? And so we kind of titrate you until you get to a point in which the subjective value of the immediate reward is equivalent to the subjective value of the delayed reward.
And in behavioral economics, this is referred to as your indifference point. And everybody has an indifference point. And what we do is we drive you to that indifference point by offering successive choices here. And when people have a lower indifference point, it indicates a stronger need for short-term gratification, because you're willing to accept a smaller reward in order to get it sooner, OK? Right? If your indifference point is lower than mine, that means that you're more interested in getting it sooner than I am, because you're willing to take less in order to get it immediately.
So here's data on the average indifference points in the immediate reward versus $1,000 in a year. And what you see is that the younger individuals are much more likely to take a smaller reward in order to get it sooner. And there's a statistic that is used in temporal discounting, which is called your discount rate, which reflects really how steeply you discount the future, how much you prefer immediate rewards. And a higher discount rate indicates a stronger preference for immediate rewards.
And so if we graph this this way, what we see is that there's a big change in discounting between age 13 and age 16 or so, right? That's what that line means there, that the 13-year-olds are much more focused on the immediate reward than the 16-year-olds are. And here's what we see in the cross-national sample, that during the exact same period of development, individuals are changing in terms of their relative preference for immediate or delayed rewards.
Now let's shift and talk about the other system in our model, the cognitive control system. And again, I want to start with the simplest version. So this is from a standardized impulsivity measure. This is a composite based on a number of items. The sample I've given you is one that's reverse scored. And what we see here is a gradual increase in impulse control over time that continues beyond middle adolescence and into the young adult years.
We're not seeing the same thing in this cross-national sample. The linear effect, in fact, is not even significant so far. Now we'll see if this holds up. But this is consistent with our suspicion that cognitive control or impulse control may be more susceptible to environmental influence than reward processing is. That's just a speculation at this point. We're not sure. But it's a speculation that we're going to follow up.
Our behavioral measure of self-regulation or cognitive control is a task called the Tower of London. In the Tower of London, you're given a configuration of three colored balls and three pegs of different sizes, and you're asked to rearrange these balls onto the pegs by moving them onto the pegs, and then back and forth among the pegs so that you've changed from the start configuration to the goal configuration. And you're asked to do that in as few moves as possible.
And the problems that we give people range in difficulty from very easy problems that can be solved in three moves to very difficult problems that take seven moves to solve, and everything in between. And one of the measures that we use to look at impulsivity in our sample is the amount of time people take before making their first move. Because if you make the wrong first move, then you're going to have to undo it, and you're not going to be able to solve the problem in as few moves as possible.
I didn't mention before, important to know, when people come into the lab to do this, we tell them, you're going to be paid $35 to do this test battery. And if you do really well, you'll get $50. All right? But we don't define what really well means. But we're just trying to get them to try their best on these tasks.
And so here's what we see when we look at time to first move. The red bars indicate the easy problems and the green bars indicate the difficult problems. So what you see here is that there is a very nice linear increase in how long people wait before approaching a difficult problem with age, and no increase in how long they wait for the easy problems. And in fact, if you look at the youngest age groups, they don't wait any longer when a problem is really hard than they do when a problem is really easy. But look at the adults. There's a big difference between how long they wait between easy and difficult problems.
And here's what we're seeing in the cross-national sample. Again, no age differences in how long people wait for the easy problems, but a significant increase in how long people wait with age before approaching the difficult problems. Again, not exactly the same pattern, but I think it tells a very similar story about improvement and impulse control in this respect over time.
We've also been interested in another aspect of cognitive control, which is resistance to peer influence. And here, we find that it increases linearly, just like self-reported impulse control does, improving with age. And in the cross-national sample, we're seeing something very similar to that. Again, not a steady, but an increase in resistance of peer influence self-reported over time between early adolescence and young adulthood.
So here's the data from Paige Harden. And this is an analysis of self-report data from the children of the National Longitudinal the Study of Youth, showing self-reported sensation seeking and impulsivity-- I'm actually graphing impulsivity here, not impulse control-- in this very large national sample of individuals. And you know, she reports pretty much what we were finding in our data. Sensation seeking follows an inverted U-shaped curve. Impulsivity follows a linear pattern. This period of heightened vulnerability is this middle adolescent period where sensation seeking is relatively high, and so is impulsivity. Or in our data, it would look like sensation seeking very high, and impose control relatively low.
So the way that we think about this now is that we can talk about this at different levels of analysis. So with the neurobiological analysis, we're interested in brain systems involved in reward sensitivity and brain systems involved in cognitive control. And these have what we might call downstream psychological manifestations in things like sensation seeking and things like self-regulation. And these then have behavioral manifestations, and we're interested in risk taking, but you could use this to study other kinds of behaviors as well.
But notice that at the bottom, that risk taking takes place in a social context. And so even though I think that the underlying neural and psychological processes are very similar among kids around the world, I don't think that we'll see that in their manifestations of these processes in actual risk taking. So as an obvious example, I doubt very much that there will be as much binge drinking in Jordan as there is in Italy. But I do think that we're going to see very similar patterns of reward sensitivity and cognitive control in those two cultures. By the same token, I think we're going to see a lot of risky sex in Sweden and very little risky sex in China, but I think we're going to see very similar age patterns in the way that reward sensitivity and cognitive control change.
So it's fine for me to say this, having studied here of all places. I'm not a biological determinist. You know, I think we need to take context into account, because context provides both opportunities and barriers to people's behavior. But I think we can learn something by studying the underlying processes that are manifested behaviorally.
So just to summarize this part of the talk, incentive processing changes in important ways at puberty that lead to increases in sensation seeking, but this takes place against a backdrop of still-maturing self-control. And mid-adolescence, then, is a heightened period of sensation seeking and poor self-regulation, and this increases people's vulnerability to risky behavior, but that the real world manifestation of this takes place within a cultural and a historical context. OK?
Now I want to go back and show you, this is a slide that I began with, showing you that individuals mature intellectually before they mature socially and emotionally. And here's what we're seeing in the cross-national sample. Very, very similar pattern. That is, the pace of intellectual development plateaus out at a much earlier age than the pace of social and emotional development. Again, in very different countries that have very different ideas about adolescence, I think.
So what do we do about this? This is an interesting quote from Shakespeare in which he says, "I would that there were no age between 10 and three-and-twenty, or that youth would sleep out the rest; for there's nothing in between but getting wenches with child, wronging the ancientry, stealing and fighting." This idea that adolescence is a period of heightened risk taking is not new. It appears in almost every description of the period in recorded history.
And you know, our work has been criticized by some youth advocates as painting an unflattering portrayal of young people. And I don't see it that way at all. I think it's like saying that infants are deficient because they can't run, right? I mean, there's such a thing as development, and people change over time. And the fact that the adolescent brain is wired in the way that it is is just a feature of this developmental period. It doesn't mean that adolescence is a disability or a mental illness or a deficiency in some way.
So now I want to switch gears and talk about some really exciting new stuff that we've been doing, which we called the peer effect. And so these are the people working on this. This is funded by NIH, both NIAAA and NIDA. And we've recently received some money from the US Army to continue this work. It turns out, we don't think of it this way, but lots of people in the army are adolescents. Lots of people in Iraq and Afghanistan are 18 and 19 years old. And the Army is as interested as we are in trying to understand the dynamics of decision making and risk taking in their population as well.
So as I mentioned before, adolescent risk taking usually occurs in groups. Most experimentation with alcohol and drug use occurs when kids are with their friends. Their risk of a serious automobile accident increases exponentially as a function of how many passengers are in the car if you're a teenage driver. If you're an adult driver, there's no relationship between crashing and having passengers in the car. And if you look at FBI data, you see that relative to adult criminals, adolescent offenders are much more likely to commit their crimes in groups.
So in some senses, adolescent risk taking is a group phenomenon. And I think that most of you can think back to your own teenage years. I know for some of you, that's not that long ago. And I think you probably would agree that you would do really stupid things when you were with your friends that you probably wouldn't do if you were by yourself. At least, most individuals that hear this talk agree with that. And most parents say that their sons and daughters do many more stupid things when they're in groups than they do when they're by themselves. And we sought to try to understand why this is.
Well, one obvious hypothesis is that, well, they do everything more when they're in groups than when they're alone, because they spend so much time with their friends. And we do know from time use data that adolescents do spend more time in groups than adults do. So we decided to look at this experimentally where we could control that.
So we have people come to our lab, and we ask them to bring two same sex, same age friends. Let me just cut ahead and sort of preempt some questions. No, we have not done this yet with other sex, other aged peers, OK? We're getting to that. But so far, all these studies have same age, same sex peers.
We randomly assigned them to take this test battery or a battery like it, the one you just saw, either alone or with your two friends watching. And we do some studies between subjects, and some studies within subjects. In the initial studies we did, the peers were in the room with the person when she was on the computer. We then developed ways of having this done with the peers in an adjacent room, where we could control how much they could communicate with the subject, because we wanted to take it into the MRI environment, and it's really hard to squeeze all those people into the magnet at once.
So in the first study we did, this is a study Margo Gardner and I did a bunch of years ago in which we had a video, risky driving game, and people were randomly assigned to play the game by themselves or with their friends. And what you see here is a big increase in the amount of risky driving when peers are present among adolescents, and a small increase among college undergraduates, and no change at all among the adults. So this then rules out the possibility that it's just because adolescents spend time with their friends, because we've experimentally controlled them.
You've seen these data already. These are bars that are extracted from an earlier slide showing the indifference points of 14 and 15-year-olds and 18 and 21-year-olds. Remember that? And remember that the smaller the indifference point, the stronger the need for short-term gratification. So we did the same study, except we did it with individuals either alone or in groups.
And what you see is that-- this is interesting when talking to undergraduates about this. We can turn 19-year-olds into 14-year-olds by putting them with their friends. So when you take 19-year-olds and you randomly-- this is now-- what's really interesting about this is that this isn't a risk-taking task. This is just, like, your choice between an immediate reward or a delayed reward. Right? And when college undergraduates who are 19, 20 years old are with their friends, it makes them prefer more immediate rewards. So this rules out the possibility that this peer effect is just limited to risk taking, and it goes along with our idea that it has something to do with reward processing.
So we got tired of dealing with the logistics of having people bring two friends with them. And with teenagers, we had to get parental consent for the observers, and all three of them had to show up at the same time, and we had to pay all three of them. So we said, let's see if we can just do a virtual peer. We'll just convince people that there's somebody their age and their sex who's down the hall, watching them.
So one of our really smart undergraduates devised a great script to do this, and it works. The deception works. The individuals are told, there's somebody down the hall that's playing with you. We have this whole prerecorded script that they think they're talking with somebody who's real, but in fact, they're saying, Hi, Gary. You know, I'm Larry. I'm a psychology major. My favorite color is blue, and my favorite movie is The Big Lebowski. And then the tape recorder says, Hi, I'm Gary. My favorite color is blah, blah, blah. So they believe that there's somebody down there, and the experimenter sometimes has to run out of the room and say, oh, something's going on, but the other subject will be right back, and so forth. So they're deceived by this.
And so here's a study of delayed discounting with a virtual peer. And we find the same effect, that individuals discount more steeply when they think that a peer is seeing their performance than when they're by themselves. So this then argues that it's not simply being with your friends. It's being-- so even with somebody that you'd never met before can induce this peer effect in college undergraduates.
So we started thinking about, how can we study this, better understand whether peers are influencing reward preference or cost aversion? So one of our grad students developed this task called the probabilistic gambling task. So we show you a pie that's divided into sections that indicate your chance of winning, your chance of losing, and your chance of no change. And your only decision is whether you want to bet on this, and it's going to spin around, and an arrow's going to point on the piece of the pie.
And we have stimuli that are in various degrees of riskiness, and so we have a low-risk stimulus, and then an ambiguous stimulus, and a high-risk stimulus. And we did a lot of pilot testing to find out what ratios we needed to do to create these different perceptions. And so we show you a stimulus, and we ask you, do you want to bet or not? And then it spins around, and then it lands on one of the areas.
So what we see here is a peer effect on people's gambling choices in this task. So in low-risk wheels, there's no difference between how adolescents-- these are 15 to 17-year-olds-- how adolescents behave when they're alone or with their friends. But the presence of peers increases betting when there's an ambiguous degree of risk, and it increases betting even more in terms of the magnitude when the risk is even higher. So not only do adolescents take more risks with their friends, they do it in riskier situations than when they're with their friends.
One of our driving tasks is a stoplight task. And so this mimics a situation that all of you who drive have been in. Imagine that you're in a car. You're trying to get someplace in a hurry. You approach an intersection, the light turns yellow, and you decide whether you're going to run it or not. We've all been in it. We've all made the bad decision, and we've all made the good decision. So this is what this task does. So you're driving down this. You're going to come to a bunch of intersections. At each one, the light is going to change, and your decision is just whether to brake or not.
And we tell you, there's three things that can happen. You can brake at the intersection, and you lose some time. Oh, that's the motivation in this thing. You're supposed to get through the course in as little time as possible. And you sit there, and the clock is very prominent, as you can see. And it's counting down, counting off the seconds while the light is cycling around back to green again.
You can run the light successfully. You can get through the intersection, and nothing bad happens to you. Or you can crash. And there's a breaking glass sound and squealing tires. And you lose a lot of time when that happens to you.
[LAUGHTER]
Twice as much as if you decided to brake. And actually, in an experiment like this, it's three seconds if you break, it's six seconds if you crash. Six seconds is a really long time to be sitting there, waiting for-- I mean, you all play on computers all the time. Six seconds is a really long time waiting for something to happen on a computer. And it turns out that risky driving on this task is correlated with sensation seeking, but not with impulsivity. So we think of this as a sensation-seeking task, not an impulse control task.
So we took this into the MRI environment, and we replicated the behavioral findings that we saw before. If you compare the blue bars when people are alone with the red bars when their friends are in the next room and can see their performance on a monitor, what you see is this increase in risky driving in the adolescent groups, but not in the adult groups. We didn't see the undergraduate effect in this study. Now this is when people are having their brain scanned. Have any of you been in an MRI? It's noisy, it's uncomfortable, it's unpleasant. And we still get the exact same peer effect that we saw before.
And what's interesting about this is that when we look at brain activity, what we see is that adolescents show heightened activation of incentive processing regions when their peers are watching them, relative to when they're alone. But there are no differences in brain activity in adults. The same pattern, regardless of whether they're alone or being watched by their friends. In addition to that, the degree to which an individual's reward centers are activated is correlated with their likelihood of running the yellow light. So we think that this is a kind of reward pursuing task that we have here.
So we've been developing and fooling around with different kinds of tasks. This is a reward task that you're show in a cue. You're asked, do you think that the arrow that's going to come up after this is up or down? So you have no control. You know that it's a random guess. And then you discover the outcome, whether you've won or lost. And we have three different types of cues, large, small, and none.
And this is just pilot data from the first attempt to use this task. And what we see is that adolescents but not adults show greater striatal activation during the anticipation of the receipt of a large reward when they're with their friends compared to when they're alone. This again confirms our idea that there's something about peers and the reward system that we need to better understand. It's not just about risk taking. There's no risk involved here. You're just guessing whether the thing is going to be higher or lower. And again, it's in the striata.
So what's going on? Well, as I said, we ruled out the fact that adolescents simply do everything more in groups, right? We rule out the fact that they put explicit pressure on each other to take risks, because we find this effect on reward tasks where there's no risk involved. We rule out the fact that adolescents are deliberately and explicitly encouraging each other to behave in certain ways, because we have the studies where the peers are in the other room, and they can't communicate with the adolescent. The adolescent just knows that he or she is being watched.
And so one interpretation that we've gotten a lot is, well, even if their friends can't tell them what to do, the adolescents know what their friends are going to want them to do. And so they're trying to behave in ways where they impress or please their friends. And as I said before, we know there's an increase in sensitivity to social information during adolescence. So this might make sense.
So we said, let's see if we can do this in a group of adolescents that aren't trying to impress or please their friends. So we just finished an experiment in which we did this with mice. This is really cool. So we created peer groups of mice that were raised together. So we took a breed of mouse that is known to consume alcohol willingly. And we weaned them shortly after birth. And then we took one mouse from each of three litters, and we put them in a cage, and they grew up together. So we created this little mouse peer group.
And we know with mice when they go through puberty. It's around 30 days or so. And so we took half of the threesomes, and we tested them when they were adolescents. And we took half of them when they were adults. They were eighty days old. And then within each of those age groups, just like in our human studies, half were tested alone and half were tested with peers. And the test was that you were put into a cage with open access to alcohol, diluted to the level that mice seem to like.
[LAUGHTER]
And you were either put into that cage with your two friends, or you were put into that cage by yourself. So here's what the cages look like. You can see along the top and the left and right side, those little bars, those are spouts, all right? So there are drinking bottles there. And so there's no competition, because there's four spouts, and at most, three mice. So everybody can drink. And there's a little wall between with an archway so the mice can cross back and forth between that, because we wanted to measure their activity as well as their drinking behavior.
And so what we find is really pretty stunning. That adolescent mice drink more when they're with other mice than when they're alone, but there's no difference in drinking behavior among adult mice when they're alone versus with their cage mates. Which is pretty interesting. I was really excited about this. And I said to my neuroscientist collaborator, we've got to get this sent off immediately. Call Science and tell them that the paper is coming. And he said, no. We really need to do this again to be sure. And so we just finished a replication last week, and we're seeing the exact same things. Totally new sample of mice, reared, weaned, put together in exactly the same way.
And what's interesting is that it's not due to differences in activity, and it's not due to differences in touching, and how often they touch the spout. It's due to differences in how long they stay on the bottle when they're drinking. So there are all kinds of things we want to do to follow this up. Some obvious things are to do it with just water, not a rewarding substance. We want to do it with glucose rather than alcohol. We want to look at-- you know, you can do all kinds of variations. And plus, because they're mice, we can look at their brains in ways that we can't look at human brains.
So we think that the heightened reward value of peers during adolescence increases reward sensitivity, and that this accentuates the imbalance between these brain systems, and increases the chances of risky behavior. And we believe now that this impact of peers on reward sensitivity, it may be a hardwired feature of adolescence. I don't have time to show you this today. There are individual differences in susceptibility to the peer effect. Not all people show the effect to the same magnitude.
So let me just conclude by talking about some implications, I think, for policy and practice. So if risk-taking is an inherent feature of adolescence, what is the appropriate public health policy? I think that most people, if they were fair in their assessment, would look at existing health education programs in public schools and say that they don't work. They've been evaluated-- most of them were never evaluated. But the ones that are evaluated don't work. They're very good at increasing people's knowledge and sometimes changing their attitudes, but they don't have very much of an effect on changing their behavior.
And I think the reason for that is that, as I've argued, the problem isn't insufficient knowledge. The problem is something else. And it seems to me that if adolescence is inherently a risky time, then we might think about public policies that protect adolescents from poor decisions. And so what we've argued for is to try to change the context rather than try to change the adolescent.
So some examples of that. The most effective anti-smoking intervention ever done was raising the price of cigarettes. Not telling adolescents they shouldn't smoke. Just making them more expensive to get. There are many studies showing that there's a correlation between the density of retail outlets in which tobacco and alcohol are sold and the extent to which adolescents smoke and drink in those neighborhoods. We can provide better after school programming for kids so that they're not left alone without adult supervision. I think we can think better about how to change this context to make it less likely that kids will do harmful things.
I don't want you to leave here thinking that what the implication of this is is that risk-taking is a bad thing. It's not a bad thing. We want to adolescents to take risks. We don't want them to take risks that are going to hurt themselves or hurt other people. And not only is it not a bad thing, it is, right? It is a feature of adolescence. It's there for a reason, and there's probably not very much we can do about it.
And so we've been asking these policy questions. Given this, given the fact that adolescents are more impulsive, that they're more reward-sensitive, should we view them as criminally responsible for their acts in the same way that we view adults? Should we be granting them the same privileges that we grant adults? How should we use this information to inform adolescents' medical decision making? That is, when we allow them to make autonomous decisions, and when they should have an adult with them when they make a decision.
So if you're interested in any of the papers behind this, you can go to my website at Temple University, and they're all available for download. And I think we have some time for questions and discussion. So thank you very much.
[APPLAUSE]
Yes?
AUDIENCE: Did you ever look at the brain activation where you have the [INAUDIBLE] in a social situation, but there's no incentive motivation [INAUDIBLE]?
LAURENCE STEINBERG: No, we haven't. Well, only in that pilot study with that reward task where it's just-- you're just guessing whether it's up or down. That's the only one. And we get the peer effect on striatal activation.
AUDIENCE: There's a reward [INAUDIBLE]?
LAURENCE STEINBERG: Yes, but they're-- yeah. But there's nothing you can do other than just guess.
AUDIENCE: But you know there's a reward?
LAURENCE STEINBERG: Yeah. You know that there's a reward. Do you have a hypothesis about this?
AUDIENCE: No, I was just wondering the extent to which just being in a social circumstance without any incentive [? has ?] whether [INAUDIBLE] social influence is to reward or what?
LAURENCE STEINBERG: Oh, oh, right. OK. So we can't-- we don't find the peer effect when we do this with response inhibition tasks. So we've run people through like a go, no-go task. And we don't get a peer effect at all. I can't say that it's only limited to reward tasks, but we don't get it on tasks that don't involve reward. Yes?
AUDIENCE: Could you elaborate just a little more on the evolutionary adaptive reason for the risk-taking? Why is it so necessary that they take such risks? Is it the leaving the nest? What's the reason in the context of the life-long survival?
LAURENCE STEINBERG: Right. Well, I think that in the wild for a juvenile animal that's just matured physically to venture off into an unfamiliar situation could be construed as a risky activity. I think there's a different narrative that you could tell that wouldn't be all that different, and we can talk about it as novelty seeking rather than risk-taking, or exploration of the environment. Although, in the mice study, we don't see changes in activity, which if it was exploration, you'd expect to see more back and forth between the two halves of the cage when the peers are watching. So I just think that it's inherently risky for an animal to go off into unfamiliar territory. You don't seem persuaded.
AUDIENCE: Well, I'm just trying to understand it sort of in today's modern context, as apparent to me.
LAURENCE STEINBERG: Oh, oh, oh. Well, I don't think you--
AUDIENCE: I still don't [INAUDIBLE] like really [INAUDIBLE] understandably how necessary.
LAURENCE STEINBERG: Oh, well, it's no longer necessary. But that doesn't mean it's not still in our genes. Right? I mean, so there are lots of things that are conservative evolutionarily that no longer serve a purpose in modern society, and I think that this may be one of them.
AUDIENCE: So can you give some quick snippets of how this boils down to-- one of the things I saw on the description of the talk is sort of parental relationships with risk-taking adolescents. Just how in the day-to-day life could these [INAUDIBLE]?
LAURENCE STEINBERG: Sure. I mean, I think that the more a parent is willing to let her teenager spend time in unstructured, unsupervised activity with peers, the more opportunities there are for that teenager to do things that are potentially risky. So here's a concrete example. I did some work with Allstate on the Teen Safe Driving Program. So all parents know that if their child is leaving the house and is drunk, they shouldn't let their kid take car keys and go out. But they don't know that if their child is leaving the house and getting into a car with three other friends who are in the back seat that they shouldn't let their child do that either. So I think we can help parents figure out how to do a better job of creating a safe context in which adolescents can be adolescents, but not be so exposed to risk.
Now so that's one. A second is, we're doing some research at Temple in which we're looking at whether putting kids through training programs to strengthen their cognitive control will make them more resistant to the peer effect. So the peer effect is due to this maturational imbalance between these systems. So one question is, if we can increase the function of the cognitive control system, will that have an effect on their behavior?
So another thing that parents could do, if this turns out to work out, is that they could provide more opportunities for kids to exercise that muscle, right? And in our schools, especially now with all the focus on testing, we don't do a lot that demands higher order executive functioning from kids. And so we may not be taking advantage of the fact that this is a developmental period where kids could be developing better abilities that help strengthen the self-regulation system. Other questions? Yes?
AUDIENCE: So you said that adolescents are [INAUDIBLE] in the presence of peers. Have you tried to see the same in the presence of adults?
LAURENCE STEINBERG: Well, as I mentioned before, we haven't yet done that. So we don't know whether this is just a social facilitation effect, or whether it's specific to peers. Now that we have the virtual peer paradigm down, we can do that easily.
Our grant from the Army, actually-- so some of you may be interested in this-- is that when soldiers go out on combat missions in the Army, they go out in foursomes. These are called fire teams. And we learned from the Army that when the Army assembles fire teams, it does not even think about the age of the people that it's mixing together. And we said, well, what if it turns out that putting four 19-year-olds together is worse than having three 19-year-olds and a 25-year-old? And they said, that would be important to know. So that's the grant that we got from the Army, to look at different age mixes and to see the impact on the peer effect. Yeah?
AUDIENCE: Have you ever questioned the quality of adolescents' relationships with their parents or their peers, and how that would affect their risk-taking?
LAURENCE STEINBERG: Yeah. Great question. And a perfect context in which to raise it, because some of the groundbreaking research on the relationship between parenting and susceptibility to peer pressure was done here at Cornell in the 1960s by Urie Bronfenbrenner and Ed Devereux. So we know from self-report measures of susceptibility to peer influence that kids who are raised in households where their parents are warm but strict are less susceptible to peer influence than other kids are. We haven't studied that in this paradigm yet.
The other thing we've been interested in is the peer effect and how it changes as a function of the relationships among the peers, right? I mean, we were wondering for a long time about dominance hierarchies, and what if the subject was a leader, and the other three were followers, and so forth. The fact that we can get it with a virtual peer where they don't know anything about that person other than it's the same age and same sex says that maybe we don't need to worry about that as much. And the fact that we get it with mice, you know-- although we don't know what the dominance hierarchy is of the three mice, and we know that animals form them at puberty. Yes?
AUDIENCE: So about the implications of that, I'm a little bothered by that. You say, we're not saying that taking risks is necessarily bad. But it seems like what the implications of this are is preventing kids from taking risks.
LAURENCE STEINBERG: Well, from taking some kinds of risks.
AUDIENCE: So what kinds of risks would you encourage them to take? Because if we prevent them taking risks, I feel like something is lost.
LAURENCE STEINBERG: I'm not saying that we should prevent them from taking risks. I'm saying that we should change the context so that the risks they take aren't going to kill them. So I think that going out for the soccer team, calling somebody for a date that you've been nervous about calling, trying out for a part in the school play, taking a course that's harder than what you normally would take. I think there are lots of risky things that kids can do.
One question that we don't know the answer to is whether this is a kind of hydraulic system in which if we get you to take good risks, are you going to be less likely to take bad ones? And I think that depends on your model of how this system is working. But, no. I hope that that's not the impression I gave you, because I don't think that we want to do that. But I do think we want to make it harder for kids to get their hands on substances that are going to hurt them. It may mean taking some privileges away from young people that you wouldn't be happy with. Look, I think that we should raise the driving age to 18, for example.
AUDIENCE: Should we lower the drinking age?
LAURENCE STEINBERG: Should we lower the drinking age.
AUDIENCE: It happens in Rome. Kids can drink at 16.
LAURENCE STEINBERG: Well, yeah. Kids can drink in Italy at 16, but there's a lot of substance abuse in Italy as well. I mean, the idea that just because kids are exposed to alcohol in their families that that makes it not a problem is a myth. So I think that--
AUDIENCE: Right. But there's a difference between binge drinking and only one drink. Between trying marijuana and being dependent on it.
LAURENCE STEINBERG: Absolutely. Absolutely. I actually have become increasingly favorable about a proposal that was made by the president of Duke about having a kind of drinking learner's permit in which between 18 and 21, you would be able to buy beer and wine, but not hard liquor. And if you were ever found intoxicated or if you did something bad while you were intoxicated, there'd be really harsh penalties, and you'd lose your learner's permit as a way of maybe exposing kids to alcohol in a context that doesn't allow them to binge drink. I think there are issues in the implementation, but--
[LAUGHTER]
Yeah. OK. Yes?
AUDIENCE: When you measured how the peer influences changed the mice drinking or the kids driving, have you measured that with the risks of unprotected sex? Because obviously, kids [INAUDIBLE] brag to their friends. But that's like a peer thing.
LAURENCE STEINBERG: I wish you could talk to NIH for us, because we had a really nice grant proposal in which we were going to look at the effect of boyfriends on girls' risk-taking to see whether they would behave in more risky ways when their boyfriends were around than when they weren't around. Now obviously, we weren't-- and then we were going to correlate that with self-reports of sexual risk-taking. But it was turned down twice. And we don't like California's criminal law, because it's three strikes, and you're out. But NIH is two strikes, and you're out. So it's even a little harsher. Yeah?
AUDIENCE: I'm sure you're aware of the recent hazing incidents that we've had on campus.
LAURENCE STEINBERG: I'm not, actually.
AUDIENCE: Well, basically we've had a couple of bad things that have happened from hazing, a death a couple of years ago, and things like that. Have you ever looked at the peer influences in terms of, like, Greek life and fraternities and hazing, and anything like that?
LAURENCE STEINBERG: No, we haven't. I think that others have, yeah. I mean, I'm sure that-- I would be pretty confident in saying that I think that individuals wouldn't engage in hazing by themselves.
[LAUGHTER]
Now-- right. Whether they engage in hazing with other fraternity members, because of the peer factor, wouldn't they do it, because that's what hazing is? It's sort of hard to tease it apart.
AUDIENCE: I'm a little confused about the drinking part. Is the risk-taking about drinking that they're drinking and it's illegal because they're underage, or is it the risk-taking that they're drinking, and then they might drive and be affected, as you said, adversely by the accident? Then how is that connected one of the last things you said in terms of, should children be held to adult standards in culpability? So how does that-- what are we concerned about?
LAURENCE STEINBERG: So there's two questions. One is, what is it about drinking that I think involves risk-taking? And let me just answer that first, and then get to the criminal law question. So if you're referring to the mouse study, I mean, the mouse study was done to look at peer effects on reward seeking, because alcohol is inherently rewarding to the brains of mammals. At least some mammals, in this species of mouse.
So that was just to see-- we could study risky behavior in mice too. There are paradigms to do that. But because our whole theory is about the effect of peers on reward processing, we wanted to see how it effected their reward seeking. In terms of human drinking and risk-taking, you know, clearly there are degrees of drinking and conditions of drinking that don't involve harm, and that don't involve exposure to risk.
But I think we know that-- we know that binge drinking is a problem, and a dangerous problem. And we know that drunk driving is a problem, and a dangerous problem. And we know that a lot of fighting and aggression occurs when people are drunk. We know that a huge amount of unprotected sex occurs when people are drunk. And we know that an awful lot of delinquency occurs when people are drunk or drinking. So you know, I mean, there's nothing inherent about drinking alcohol to a certain level that's risky. But I think we know that it is associated with a lot of risky things. We also know that the impact of alcohol on the adolescent brain isn't so great. I mean--
AUDIENCE: That's what I thought [INAUDIBLE].
LAURENCE STEINBERG: Right. So that's another thing. In terms of the criminal thing, I think a lot of adolescent crime can be seen as sort of a subset of risky and reckless behavior. One of the chief features of adolescent crime is that it tends not to be premeditated. It tends to be impulsive, and it tends to occur in groups. So it shares a lot of characteristics in common with other forms of risk-taking.
And so the idea about diminished culpability is if this is an inherent feature of the way people this age are, that we shouldn't look at them the way that we look at adults. And that the reasonable person standard that we apply when judging the behavior of an adult might not apply, and maybe we need a reasonable adolescent standard in judging the behavior of a teenager who's broken the law. Yes?
AUDIENCE: You mentioned that the Iowa gambling task had to do with reward sensitivity versus loss and that you have a way to do that to separate those out. And I agree with you [INAUDIBLE]. But recently, there were some studies published showing that if you remove the memory demands of that task, and you provide the [INAUDIBLE] information in a more salient way, you had a running total and also a running total, that many of these developmental differences, at least in that pass involving experiential learning seemed to disappear.
LAURENCE STEINBERG: Hm. I haven't seen those papers. I mean, I know that in other versions of the game, they do find the developmental pattern.
AUDIENCE: [INAUDIBLE].
LAURENCE STEINBERG: Yeah, yeah, yeah. Yeah. So I'd love to see those papers. Yeah. But we do know that individuals who have dorsolateral lesions play this game just fine if their ventromedial cortex is intact. So it doesn't seem to be a working memory task.
AUDIENCE: Although, a lot of the models in adult behavior, like the Busemeyer model or some of the other experiential models, they typically incorporate a memory parameter, but also sensitivity to gains and losses. And usually, the memory parameter is significant.
LAURENCE STEINBERG: Yeah. I think--
AUDIENCE: [INAUDIBLE] the consequences, it can be a problem.
LAURENCE STEINBERG: I mean, I think it's fairly safe to say that there is a lot of disagreement in the field about what the Iowa gambling task measures. Yes?
AUDIENCE: I have just one question for the peer effect. On your studies, and so far, we've been discussing the [INAUDIBLE] negative effects of, like, how risky [INAUDIBLE] increases. Are there studies which look at how the influence of a peer can be positive?
LAURENCE STEINBERG: Yes. I mean, not in this paradigm, but we did research, and so did Bronfenbrenner and Devereux here, showing that-- I mean, for instance, there's a lot of evidence that adolescents who have friends that are highly oriented toward achieving in school do better in school than adolescents who have the same ability, but hang around with friends who don't value academic achievement. So I don't think that peers are a negative influence.
AUDIENCE: Yes, but also like nowadays, there are a lot of programs out there that say like, don't drink and drive, or don't text during driving. So would the effect of a peer actually being involved in an awareness program help people [INAUDIBLE] and not get into such risky behaviors better than if, like, an adult would be [INAUDIBLE].
LAURENCE STEINBERG: So you probably couldn't hear it in the back. So the question, which is a good one, is if we're trying to do programs that raise awareness about not drinking and driving or not texting while driving, I assume if you have a peer with you saying, don't do that, you're not supposed to do that, would that help? I don't think we know the answer to that. I mean, I think there are lots of approaches to intervention that think that selling the intervention through peers is going to be more effective than having adults deliver it.
But you know, in talking to kids about this, I think that-- and some of you who are more recent teenagers than I might be able to be more informative here. It seems to me that lots of times, kids are afraid to speak up. When they're in a car, and the person who's driving is driving too fast, or has been drinking, or-- I mean, we hear this a lot from kids. And in some focus groups that Allstate did, it led us to believe that one of the things that we needed to do was to encourage adolescents to speak up when they were scared about something reckless that their friend was doing, because a lot of them don't-- especially boys are afraid to do that.
AUDIENCE: But now there's programs that are bringing in some victims to speak about it.
LAURENCE STEINBERG: That doesn't work.
AUDIENCE: Doesn't work.
LAURENCE STEINBERG: No. The bringing in victims to speak about it, no. That doesn't work.
AUDIENCE: I think that the risk that you're talking about is pretty much on a spectrum. So like, as a teenager sitting in a car with someone who's driving fast, how fast is too fast to speak up? If he's driving 10 miles over the speed limit or 40, that might change a person's willingness to speak up. Or, how drunk is too drunk to be driving? Some person might feel comfortable driving with someone who had a drink at a party rather than four drinks. So I think that there's a spectrum in how people speak up against risk.
LAURENCE STEINBERG: Right. And I think that your comment and this question, to me as a scientist trying to study this, you know, it shows how difficult it is to try to create situations in a controlled experimental environment that have some real world ecological validity. So for us, in order to stop-- in order to not find ourselves in a situation where the adolescents were cheering their friends on and the adults were saying, don't, we had to then develop a script that only allowed them to say their names and hi, right? If we just let them say whatever they want to say, which is what the real world is like, then we won't know what to attribute the developmental differences to. So it's always a sort of-- as most you probably know by now, being in human ecology, that trying to go back and forth between the lab and the real world is not such easy business, you know? Great. But thank you all.
[APPLAUSE]
Dr. Laurence Steinberg, the Distinguished University Professor and Laura H. Carnell Professor of Psychology at Temple University, presents the results of his program of research on the underpinnings of risk-taking in adolescence that is informed by recent advances in developmental neuroscience.
