Here's a study that I did with Marty in my very first year of graduate school. I had come right out of teaching and I was so eager to get started. Marty did point out at the beginning of our mentorship that he doesn't actually study self-control. In fact we, naively I think, called it self-discipline. And I was just like, call it self-discipline. That's what I used to call it when I taught. I didn't actually recognize the fact that nobody else in psychology was calling it self-discipline. So that was a little bit of a rookie error. And if I could go back in time I think I would call it self-control, just to avoid confusion, right? It's often called a jingle or jangle fallacy when you name your construct something different. So if you are going to go and study purpose, just call it purpose and not something else. Everyone's like, what is she studying, what is this capstone about, right? So we called it self-discipline and we had a measure of self-discipline that was a composite. So it's self-reported questionnaire using the brief self-control scale and also another one. Teachers filled out the same thing about you. Your parents filled it out about you. We had a questionnaire that was called a discounting questionnaire where you had a series of question that we're all like this. Which would you prefer? $2.00 today, or $4.00 in a month? $3.00 today, or $10.00 in a year? Just a series of questions. And you could do a little math, and try to calculate how present-centered versus willing to delay gratification the kid is. And in subsequent studies, we made one of the choices real. We're like, take this seriously because you're actually going to get one of the things, and if you chose early you get the, you know. And then my favorite measure, because Marty told me not to do it. So it's also something to remember as you go into your own graduate research, that advisers give advice, ultimately make the decision. So I came to Marty one day and I said, we should give kids money. We should give them a dollar bill, and it'll be like the marshmallow test. So if they want the dollar bill right now, they can keep it. But they can give it back, and then they get $2 later, right? He was like, I don't think that'll work. And I was like, yeah, I think it's going to work. Well no, I don't think that'll work. >> [LAUGH] >> I remember what you said, it wouldn't work. So I left his office and we did it anyway. So $1.00 today or $2.00 in a week, right? And turns out that about 20% of the kids took the dollar right away put it in their pocket went home. And 80% of the kids or so waited, right? And so we also averaged that decision into this composite score. And one of the most important things I want to teach you about measurement, is that since all measures suck, right, in their own way. What you can do is take a bunch of sucky measures and put them together, and the average of those measures will actually be better than any one measure alone, right? It's like the basic idea of triangulating. So if you're going to hire someone. Yeah you could call one reference or you could call five of them. And you're like, now I see everybody's perspective. That's better information then just calling one person. So that's what we did here. Parents, teachers, self-report two different ways of measuring delayed gratification. We averaged these all together. That gives self-control score, IQ on the dotted line. That is a standard IQ test, right? It was like a standard published test by a company that published IQ tests. So what we found is, that when you measure self-control and IQ in the Fall, and then you wait around until the end of the academic year and you measure final grades. The relationship between self-control and grades is stronger in the sample, than it is for IQ and grades. So sure having a higher IQ, kids doing better. Well look how steep this line is. If your in the top 20% of self-control you're doing actually better than the kids from the top 20% that like you. And so we called this paper something like, Self-Discipline Outdoes Talent. And then there was a colon and then there was a blank part. I don't even remember what the boring part of the title was. And I recently got an email from the editor of Perspectives Inside Science because we published this in Inside Science. And he said you know this article is one of the most cited articles in like whatever, however many decades. So now you get to write about what you've learned since writing this. And I think I'm going to write like, I did this study as a first-year graduate student who made a lot of repeat errors. But the one error I didn't make is that I remembered that the most interesting hypotheses come from your life experience. And I had all that time that I spent with kids. All those observations over all those years and I knew in my bones that there was a good idea here. And so my advice to young scholars is to use all that. And yeah, I could teach statistics, which got the lit review, right? Like have two P-values yeah, anybody can learn that. But nobody has the life that you led and the insights that you bring to the world. Okay so any questions on that? The Correlational, the Longitudinal study? >> What happened to the dollar and the $2.00? >> Well the dollar and the $2.00 thing. First I went back to Marty and I said, look the decision that the kids made and it's a simple measure. I think one of his concerns was, is that it's very coarse, right? It's a yes, no, like do you take it or not? And he had an intuition that, that won't be a very precise measure. And he was right about that, but it's still correlated with all the other measures of self-control. And even I think, in this sample, just that one measure was as predictive of your grades as actually your IQ score. Just knowing that. So, then Marty being a judgement scientist was like, cool, right? >> [LAUGH] >> So Angela. >> Yeah. >> You were predictive a couple of times there, but this is a correlational study. >> It was a longitudinal correlational study. So self-control and IQ were measured in the Fall and the grades were measured at the end of the year. So in that context, it's still legitimate to say that x predicted y. What would not be legitimate is to say that x caused y, right? If it were really cross-sectional like measured it on May 7th, measured the outcome on May 7th, then you have to say they're related, right? >> Right. >> So, that's the language. When you read these articles, if it's predictive they're supposed to say predictive is predicted over time. Related means it's cross-sectional usually and then causal is hard to get to, but if you did an experiment you might start to use language like, led to, determine. >> So predictive is different than causal, for the language that you use. >> Yeah, so for example, many of you, well not really, I think this is a myth, when people were like, my knee hurts it's going to rain, right? >> [LAUGH] >> My knee hurting predicts the rain but it's not causing the rain. >> Right. >> [LAUGH] >> That's the difference.
