Student Sorting in the College Selection Process

February 26, 2021

This 24-minute podcast was released Feb. 26, 2021.

Oksana Leukhina, economist and research officer

“It’s important to graduate, but it’s also important to learn. My model tells me it’s not just about getting the diploma, it’s about how much human capital you end up with,” says Oksana Leukhina, economist and research officer at the Federal Reserve Bank of St. Louis. She talks with Austin Roberts, social media analyst, about the importance of choosing the college that is right for the student.

This podcast episode is based on college quality research by St. Louis Fed research officer Leukhina; Lutz Hendricks, associate professor of economics at the University of North Carolina at Chapel Hill; and Tatyana Koreshkova, professor of economics at Concordia University in Canada.

Additional research on this topic includes a Regional Economist article with Joseph McGillicuddy titled “What’s behind Rising Returns to High-Quality College Education?



Austin Roberts: Welcome to Timely Topics, a podcast series by the St. Louis Fed. I’m Austin Roberts, your host for this episode. And with me today is Oksana Leukhina, economist and research officer here at the Federal Reserve Bank of St. Louis. Oksana, thanks for being here.

Oksana Leukhina: Thank you for having me.

Roberts: Today we’re going to discuss your research area that’s of special interest to me, as someone who’s recently graduated from college, and that is college sorting and the inefficiencies and policy implications that come with it. So, Oksana, as background, could you give us a sense of what college sorting means?

Leukhina: Well, by college sorting, I basically mean the allocation of different types of students across different types of colleges at the time of college entry. So, some students will start their path towards a college degree in a two-year program. Others will start in four-year colleges with varying degrees of selectivity.

Let me tell you a bit about different types of colleges and how we split them into types because that is going to come up later in the conversation, I’m sure. There are about 2,000 four-year colleges in the U.S. and over 1,000 of two-year community colleges. So, we are going to treat community colleges as a separate type of college. And we are actually looking only at community colleges that offer a two-year general education program that you can later transfer somewhere.

And then we will categorize four-year colleges into quality types by first ranking them according to the average SAT score of their freshman class and then splitting these colleges into three groups with equal enrollment. So according to this classification, higher-type colleges are going to be basically more selective, right? And that is going to matter for peer effects.

But they are also going to provide higher-quality instruction, better advising, better tutoring services. And of course they’re going to cost more. And then students also differ along many dimensions that matter for their choice of college and even enrollment choice. And these differences would be, you know, things like family background or how prepared they are for college, you know, how well they did in high school. So, when we talk about sorting of students across colleges, we generally refer to where different types of students end up going to college.

Roberts: So, your current research compared students who entered college roughly in the 1980s and those who enrolled in the early 2000s. What differences did you find between those two groups?

Leukhina: Yeah, so we looked at these two different generations basically 20 years apart. And there are many differences between these two groups. First of all, let me talk about enrollment. We know that during the 1980s, that’s the decade during which we saw an increase in college premium. So naturally college enrollment rates are much higher in 2000. They increased from about 46% in 1980 to 57% in 2000.

And so what are the types of students that enrolled more in 2000 compared to 1980? Well, we can see – and this is a well-known fact – that there is little gain in enrollment for students with family income in the lower half of the income distribution. And there is also little gain in enrollment rates for students with high school GPA in the lower quartile of the high school GPA distribution. These students enrolled at very low rates back in 1980, at less than 30% rate, and they continue to do so. There is also little gain in enrollment for students in the top 25% of the GPA distribution. Those students already enrolled at high rates back in the 1980s, at about 80% rate, to be specific. And they continue to enroll at high rates.

Basically what we saw happen is that students that increased their enrollment rates the most were those in the middle range of the high school GPA distribution and whose parents were relatively well-off. By that, I just mean the parents whose income was in the upper half of the income distribution. So that’s about enrollment.

Now let me tell you also about financial costs. We know that financial costs increased quite a bit for the younger cohort. The direct cost of college more than doubled. It increased from about $3,600 to $8,600 in 2000. So to cover the increase in cost, we found that the younger cohort basically had to rely more heavily on parental transfers. And that is the most important source of those extra resources. They also worked a little more while in college, and they accumulated a little more student debt. So students’ earnings and annual loan amounts went up by about $1,000 each. And then the rest was covered by increased parental transfers.

Now let me talk a little bit about changes in sorting. So we found that there’s evidence for downgrading of the college type, which appears to be related to financial constraints associated with rising tuition costs. We saw that 30% of freshman in the 1980s went to community colleges, while 37% chose to do so in the early 2000s. And what we see is that downgrading of college type happened mainly for lower-performing high school graduates. They are downgrading from the lower-quality four-year schools to two-year schools basically.

And the opposite is happening for students in the upper half of high school GPA distribution. These students are actually upgrading their college type somewhat. Remember how I talked about different types of colleges that we defined? So these students are going to type 3 and 4 instead of type 2. And so this suggests that sorting of students across colleges is getting stronger.

Roberts: And can you tell me about the tradeoffs involved in different types of colleges?

Leukhina: Yes. So I will highlight the main differences that we chose to focus on. In my research, we basically have a model where we simulate students and their choice of college. This is how we chose to model differences in college quality.

First I should point out that I assume that your earnings later in life are going to depend on your human capital. And what do I mean by that? Human capital is basically all the skills and knowledge that you have that help you be productive at your job and help you earn your salary. So your earnings are going to depend on that and also on whether or not you have a college degree. You can think of it as, well, if you have a college diploma, you’re going to have access to certain type of jobs. But on top of it, also what matters is how productive you are.

The way we think about different types of colleges is that the top-quality schools are basically going to be more selective. So the minimum level of, say, expected learning ability that their students must demonstrate is going to be higher. That’s number one.

Number two, they are going to offer a superior human capital accumulation technology. So this is about all the things you will learn that will help you be productive in the labor market. You will have some human capital upon high school graduation. But as you enter colleges, you will be accumulating more human capital, which will later on help you be more productive at your job.

So what we think is going on is that there is more learning to be done, say, per unit of study time in better-quality schools because more resources there are spent on instructional quality. So you have more access to, you know, your professors. You have tutoring resources, etc.

And possibly the upper bound on what can be possibly learned as an undergraduate student is also higher. So what we have in mind is curriculum rigor, the type of course offering – so maybe in a better school there would be a graduate-level program, so you would have access to graduate-level courses if you wanted to take one.

But it is also going to be harder to graduate from a higher-quality school for a fixed study effort.

So in the end, no matter where you go to college, you basically need to accumulate about 120 credits. But you will have to demonstrate more knowledge per course in order to earn those credits in higher-quality schools. And, by the way, this is consistent with what we see in the data for study time. So we basically see that students study about 13 hours per week in two-year schools. And they study about 18 or 19 hours in the lower two types of four-year schools, the way I define them. And they studied 24 hours per week in the top-type four-year schools.

So it’s evident that either learning is just so much more productive there that students choose to study more in better-quality schools, or else learning standards are set to be higher. Or, well, in our model, it’s going to be a combination of both.

And then finally, higher-quality schools are also more expensive, right? So these are the differences. So what happens in our model is that graduates from higher-quality schools end up with higher earnings because they end up graduating with more human capital. And in fact, that’s the only reason why they have different earnings. They may start with more human capital to begin with, but they also study more. They enjoy access to the more productive learning technology, and that’s how they end up with even more human capital in the end. And the diploma itself, that’s worth the same regardless of where you get it from.

Roberts: I think our listeners would be interested to know, does efficient sorting look like perfect sorting? And what keeps us, then, from achieving efficient sorting?

Leukhina: Well, that’s a good question. So efficient sorting doesn’t mean perfect sorting. And I suppose what you mean by perfect sorting is the sorting on some kind of observable characteristics such as, you know, high high school GPA students going to the best—the most selective universities.

But remember that efficiency is not just about human capital accumulation. Efficiency is about maximizing student welfare overall.

What I’m trying to say is that it may be efficient for a student of relatively high learning ability, so the student is very good at learning things. But it may be efficient for this student to enter a slightly less selective school if they already have a lot of human capital to begin with. So they trade off that extra human capital they could get in the best school possible for more leisure time without sacrificing graduation outcomes. And then of course they also save on tuition.

So perfect sorting is not something that our model is going to suggest should happen. So now talk about sources of inefficiency, what kind of outcomes that we worry about. So what we do worry about would be, say, a student who would ultimately like to choose a better school but is prohibited from doing so financially, maybe due to a combination of low family resources and strict borrowing limits. A student like that would choose to start at a lower-quality college because it is cheaper. It allows them to work longer hours while earning a degree at a slower pace.

And this type of outcome leads to inefficiently low study effort. It slows down human capital accumulation for the student and delays entry into the labor market. So these are the kind of outcomes we would like to avoid.

What we have in our models is that there is uncertainty with respect to human capital accumulation. We could also worry about things like lack of information regarding school quality. So you may start college and end up with a roommate who parties all the time, and it would be very difficult for you to study effectively.

So that’s kind of a random shock for you. But it is going to hinder your progress. You may have to reevaluate your situation after one year if you end up not doing as well as you would’ve otherwise. And you may drop out without earning your bachelor’s degree. Or you may try to transfer to a different school, which of course is going to lead to the loss of credits. And it would hinder your progress towards graduation. There’s no insurance against such outcomes. So that would be another source of inefficiency.

Roberts: So I think it’s safe to say a lot of students and parents who are about to face these decisions or already have would want to know common mistakes they could avoid. What would you say are the most common mistakes students make in the college selection process, and why do they make them?

Leukhina: Yeah, so I would say, well, there is a lot of information out there that students have access to. Things are very different, you know, from the time I went to college. And I think that I would warn students about misinterpreting the average statistics that are out there.

For example, we see that six-year degree attainment rates differ dramatically across the types of colleges that we defined, so about 16% for students that start in two-year colleges to as much as 85% for students that start in the top-tier four-year colleges. Does that mean that these are your graduation probabilities? Does it mean that if I go to a two-year school, my graduation probability is going to be 16%, and if I go to the best four-year school, my graduation probability is going to be 85%? I would like to just clarify that it doesn’t mean that.

So these are just average probabilities. And the reason why they differ is because different students attend these different types of institutions. So for a specific person, when they consider, “Well, should I go to this school, to that school?” They should be looking at the experience of students in each institution that look like them. So the same is true about salary differences upon graduation.

So suppose that you are comparing two California colleges. Suppose you are a super high-performing student in high school, and you want to study engineering. And then you read about this college called Harvey Mudd College. And students in that college all major in STEM fields. It’s like an engineering school.

If you Google it, you are going to see that these students earn about $86,000, the students that graduate. Suppose that you get in. And suppose you also get into University of California at Irvine. And you see that students that graduate from University of California at Irvine, they earn about a third less, which is quite a big difference.

Well, does it mean that that is going to be your experience? Does it mean that if you select to go to UC-Irvine, you are going to earn one-third less? Obviously that’s not true because everybody at Harvey Mudd is going to major in engineering, science, technology or math. And as soon as they graduate, they get those type of jobs.

At University of California at Irvine, there are students majoring in all kinds of fields, some of which are less lucrative. So, again, you have to be careful. You have to think about yourself, and you have to look at students who study engineering. That would be a better comparison group for you. So you just have to be careful not to project the average statistics onto your own projected experience.

Roberts: Your research suggests that the single biggest predictor of how well a student will do in college is how well they did in high school. How strict is that prediction, would you say?

Leukhina: In fact, this is kind of one thing that I found so impressive, that your high school GPA is such a strong predictor of what happens to you later on. So, for example, if I look at freshman from the lowest quartile of high school GPA, only 20% of them will graduate six years later. And if I look at freshmen with high school GPA in the top quartile, so in top 25%, as many as 72% of them will graduate within six years. So that’s a huge difference, right?

These effects are seen right away. So even in the first year of college, we see that those top quartile students earn about twice as many credits as those students that were lower-performing students in high school. So yes, that’s an amazingly strong predictor because it basically captures how prepared you are for college, how good you are at studying, how good you are at submitting those essays and following deadlines.

And my interpretation of this is that college preparedness is key for college success. But of course if you just look at something like an SAT score, there will be a lot more variation because it’s just a test score. High school GPA is an outcome of four years of your study history.

Roberts: Now, digging a little deeper here, from a policy standpoint, what can we do to improve the college selection process and therefore ultimately improving graduation rates here in the U.S.? What type of policies might you say could be introduced?

Leukhina: I would say getting them young and making sure that our students are prepared for college, that is key to ensuring the success. And so we basically need to make sure that all kids have access to high-quality public education from kindergarten all the way to grade 12. High school is about getting those general skills such as reading, comprehension skills, organizational skills, you know, just knowing how to study. This is so important because that’s what matters for how well you do later in college.

And I would also say something about two-year schools. Like I said before, graduation rates for students who enter two-year schools are extremely low. They’re less than 20%. Over 90% of students that enter two-year schools actually say that their ultimate goal is to get a bachelor’s degree.

So perhaps we could improve their performance. There are some students who mature later. Maybe they were bad students in high school. Maybe they just don’t care about meeting those deadlines.

Well, if they enter a two-year school, this is their chance to catch up. So perhaps we could do something about their coursework - you know, offer more remedial courses that can actually help them prepare for college. I saw a recent study about advising. So it turns out that connecting to an advisor early on when you enter a two-year college seems to matter for your success later on. So connect with that advisor. Develop a better plan for how you’re going to transfer to a four-year school, perhaps investing into partnership with local four-year schools to ensure that your general education associate degree can be transferred there.

Perhaps I could also mention that simply subsidizing tuition in a four-year school is not something that I would recommend because what you would get is you would get a lot of marginal students. This is what we get in our model. Students that are in the margin between attending college or not attending college, a lot of those students enter and they drop out after year one. So, like I said before, we worry a lot about students who should be going to college but are financially constrained.

But what we could do to improve the outcomes is basically we could condition tuition subsidies on their academic performance. And this way we are ensuring that the students that benefit the most are responding to these type of tuition subsidies and actually benefiting from entry and from graduation.

Roberts: Great. Well, is there anything else you’d like us to know?

Leukhina: I would just like to emphasize once again that I don’t recommend students being caught up in this game of trying to attend the best possible college. There is a variety of colleges out there, and I think finding a school which is a good fit for you where you can be happy, not stressed out, learn happily, and actually progress towards graduation, this is kind of a key to your success.

And just knowing yourself. Knowing what it is that you want is also very important. So if you know that you like to have challenging courses, then maybe you should go to a school that has a graduate program in the field that you want to study. And if you know that you will need extra help studying, then perhaps you should go to a school that offers tutoring services and things like that. So, again, just knowing yourself and not being caught up in that game is something you should consider.

Roberts: Well, all right. Terrific. Thanks for a great conversation today, Oksana. I really appreciate it.

Leukhina: Thank you so much.

Roberts: To learn more about the St. Louis Fed’s research or Oksana’s research, visit online at

Related Topics

Economists and experts talk about their research, topics in the news and issues related to the Fed. Views expressed are not necessarily those of the St. Louis Fed or the Federal Reserve System.

Back to Top