 Welcome. I know some folks are still filing in. I'm glad to see we have such a good turnout today on this somewhat rainy day. My name's Brian Jacob. I'm a faculty member at the Ford School and director of the Center for Local State and Urban Policy. This is one of our regular public lectures and education policy topics. Today, we are delighted to have Mike McPherson, the president of the Spencer Foundation, here to speak with us about the undermatch phenomenon in college choice. And he will tell you lots about what that is and why we should care about it. I have known Mike since I was a graduate student and reading some of his early work in the economics of education. But when he was coming here, I knew I would have to introduce him. I had the opportunity to look back more in his bio. And it really is remarkable how much he has done over the course of his career from being a faculty member at Williams and then going on to the presidency of McAllister College and now serving as the president of the Spencer Foundation. And what actually, I think it mazes me even kind of more than his earlier academic achievements is the fact that now as the president of the Spencer Foundation, he still finds the time to do serious academic research and not just publish articles but publish entire books. And one of his latest books is Crossing the Finish Line, completing college at America's Public Universities. I think he'll be drawing on some of that material today. And so he is someone with a really unique perspective both in an academic, but someone who interacts with lots of researchers and lots of policymakers and practitioners through the work of the Spencer Foundation. So I'm delighted that he is here to join us. In one minute I will turn it over to Mike to get started. I would like to thank the Ford School for hosting. I'd like to recognize support provided by Charles and Susan Gessner. And I'd like to thank Bonnie Roberts and Tom Cook and some of the other folks who've put together the logistics for this event. And so with that I will turn it over to Mike to tell us about the college under now. Well, thank you very much, Brian. It's always a real pleasure to visit Ann Arbor and see so many friends and so many interesting people at the University of Michigan. Before I get started, I did want to say, I know some of you may be thinking that you'd like to slip out a little early because there's a big calculus exam tomorrow. So just so you know, in the last five minutes I'm going to give you the answers to two of the problems in tomorrow's calculus exam. So you may want to stick around for that. So I want to talk today about a phenomenon that became a big focus of our attention when we were writing Crossing the Finish Line and we came to call it undermatch. And I'll explain more about what that is. It's not something that we were by any means the first to discover, but it actually was not a phenomenon that we had anticipated. And it's always exciting when you're doing empirical work to find something interesting and unexpected along the way. A lot of work was done in this area earlier, very influential work by folks at the University of Chicago, particularly in the Consortium for Chicago School Research. I'll explain to you what this is all about. Let me first describe, let me say what my plan is for this talk. I want to review some of the basics about the book, why we wrote the book, what the evidence is that we came up with and how we interpreted it. I'll comment on a couple of the questions that can be raised about that evidence. Some of you may want to help with that either by answering my questions or more likely by raising more. And then I want to explore what policy implications at different levels might be thought to follow from this analysis. And I hope we can make that part of the afternoon more open-ended. I think actually these are very complex things to think through at the policy level and I would not claim at all that in writing the book where we were mainly trying to describe our findings that we had those thought through and I wouldn't claim that I have them thought through now. But I think they're really interesting and they're a nice case of empirical evidence interacting with empirical kind of economic evidence interacting with broader policy thinking. And perhaps there are other studies that may be suggested, other work that might be suggested as one tries to think the implications through. Anyhow, the way we got into this book was my wonderful mentor Bill Bowen who had been the president of Princeton University and then of the Mellon Foundation and then had pretended to retire got interested in the phenomenon of graduation from public universities in the United States. This was a change of focus for Bill who had done tons of important work on the private university and college world and Mellon had given a lot of its attention to that world. But in the larger scheme of things, we know that public higher education educates the bulk of the students and the issues facing public higher education are ones that are gonna be very important for the future of the country. So Bill recruited me to work with him and also recruited a young man named Matt Chingos who was a graduate student at Harvard University. Actually had just completed his undergraduate career when he started this and was going to begin graduate study in political science at Harvard University after that. Matt is an extraordinarily talented data analyst. The fastest gun in the East I think when it comes to generating results and also a really good analytical thinker. So we formed a team of three people and dug into the data. Now the data was really put together through Bill Bowen's remarkable ability to get people to do what he wants them to do. He is really an amazing man. And what we did was approach flagship public universities, the leading public universities in a number of states to seek all the data they had on students who entered for undergraduate study in the fall of 1999. We did this in 2006 and we looked for everything they had by way of their background data, their admissions file and so on, what kind of data was in that. And their transcript data thereafter. So for these students, we had a very extensive dataset about them. And we succeeded, ultimately Bill succeeded in getting, I think it's something like 22 of these universities to cough up their data. In the age of FERPA, this was a truly remarkable achievement on Bill's part. And those data were then merged with College Board SAT background data, SAT people give an interview about students' backgrounds when you take the exam, similar data from ACT and linked to the National Student Clearinghouse for information on the fate of people who left the university before they graduated. That was quite a lot of data, but it wasn't enough for Bill. So he also persuaded the systems, the public university systems in four states to give us the data on their whole public four year college system. So in addition to those flagships, we wound up with a sample of what used to be called state colleges, kind of a step down in the pecking order of higher education for four states. And in the case of North Carolina, we actually got the data on their HBCUs and on their community colleges. Those are the only community colleges, community college results that we had in the system. Again, same kind of data, individual level data for the students entering in the fall of 1999. So that was a tremendous asset, of course, because it gave us a little better handle on some of what was going on with selection across these different systems. We were interested in the study in trying to figure out the sources of variation in graduation rates. We took, as our objective, to try to explain as well as we could with these data, college completion. We didn't pay principal attention to grade point average in college, or we didn't have data on the students after they graduated, so we weren't looking at later life outcomes. This was really looking at people beginning, following their careers and seeing if they dropped out, or if they completed within six years. We were interested in race and gender effects, in effects that we could infer. Now, of course, I'm using effects here, even though these are observational data, and we need to be careful about causation, but as control variables in our various analyses, we had race and gender and high school test scores and high school grades and family background and family income and lots of things like that, and we proceeded to analyze those data. Motivation for the study is easy enough to describe. This figure shows educational attainment by age 35 for people who were born in various years over a long period of time. I think this is from Golden and Cats, but it's certainly data of the kind that they made familiar, which you see this steady march of improved educational attainment that coincides with the enormous progress of the United States economy over much of that period, and which Golden and Cats and others have argued, there's evidence that those things are causally related. The U.S. led the world in expanding educational opportunity for most of that period. And then a fairly abrupt tailing off of progress in overall educational attainment for people who were born, it's a little hard to see from this angle, but people who were born in something like 1950. And various possible explanations could be offered for that slowdown, drastic slowdown in progress in educational attainment. Maybe everybody who could be educated already was educated, you go through a lot of possibilities. But in fact, we know that over that same period, other countries caught up with and surpassed the United States in these figures, so it's a little hard on that basis to think that we had encountered some natural limit. In any case, it's been a lot of emphasis the United States on wanting to improve educational attainment for reasons both of equity and of economic success. And this is an illustration of what motivated that kind of concern. Another kind of concern and also shedding light on where the opportunities might lie if you want to improve college success is to look at bachelor's degree attainment by socioeconomic status. These are data from NELs, not from our study. And two things come through dramatically in this simple graphic. One is that if you're a first-generation college student and we define that in terms of whether one of your parents graduated from college, overall you have a substantially lower likelihood of graduating from college than if at least one parent graduated. And then within those two categories, your family income background makes a lot of difference. So if you're a first-generation bottom quartile of income, less than one in 10 from that group get a bachelor's degree. If you're a child of college-educated parents in the top income quartile, seven times as large a percentage get a bachelor's degree. Now, most standards of equity or fairness would say there's something very disturbing about these relationships. But also if what you're thinking about is where are you gonna get more college graduates from? It seems obvious that the place to look is the lower bars. You could squeeze out more college graduates from that group that's already getting 68%, but that kind of looks like an uphill battle at that point. And in that region, maybe that's about the number you're gonna reach. But if you could find a way to change the results for people who currently have very low probabilities of graduation, obviously you could do a lot more. We wanted to find out what might seem feasible in that regard. Now, as we did this work, we looked at a number of different factors that seem to be at play. And this is, as any of you who have read it or carried it around know, this is a big book, even though the appendices were put up on the web. So I'm not gonna try to summarize the findings of the book. But one thing started to leap out at us as we continued to do the work. And this was what we came to call the undermatch phenomenon. And this graphic highlights one of the major findings from our evidence in the book. We had, we developed a way to define what we met by an undermatched student. And the idea here was that we could infer quite a bit about what kind of credentials it took to have a high probability of admission to say a flagship state university. And so we could simulate for students in our study the probability, we could identify students who had a very good probability of getting into one of these more demanding selective places if they apply. And we could look at what happened for those students who had those qualifications and did that, and what happened to students who had those qualifications and didn't do that. So here we focus on, over here on the left, the students who went to a less selective institution, in particular category B in our selectivity scheme, even though their credentials indicated that they very likely could have gone to a more selective institution. And we looked at graduation rates for students who had similar credentials. This is very important. These are students who had similar high school grades and similar test scores. In general, you expect to find highly selective institutions have higher graduation rates because they're picking out students who are already more likely to graduate. But in terms of what we could observe, we controlled for that. And then we did the comparison. For four year graduation rates, which is the darker bar, students who under matched had less than a 50% chance of completing the degree. Students who went ahead to their match school, that is who were capable of getting into a selective A and did go to a selective A had about a 60% chance of successfully completing. And if you took six years, you added about 22 points to each of those so the difference in those outcomes was maintained. That's a significant difference for sure in the data set of the size that we had. And it's a little bit counterintuitive. You know, I must say that we've talked about this stuff so much that now it seems not surprising at all. But if you said to yourself, look, I don't actually care how good the food is, how good the education is, anything about what my life is gonna be like, what I care about is am I gonna graduate? Should I go to the hardest place I can get into? Or the easiest place I can get into? Well, turns out you should go to the hardest place you can get into. It doesn't actually work out that playing an easier opponent in tennis terms, in this case, actually leads to a better result. Now, we could have a lot of speculation about why that seems to be true. And of course, I'm stating it more dramatically than the evidence really supplies. But this suggests that a strategy of looking for an easy way to pick up a BA doesn't turn out to be a highly effective strategy, quite the opposite. Now that in itself is an interesting phenomenon and it suggests that if you're interested in college completion, you should encourage people to challenge themselves by going to the more selective place if they can get in. And I should say in most of the cases of undermatch, these are not students who applied, they're not students for sure who applied and were turned down. They're also not students who applied, were accepted and didn't go. Most of these are students who never applied to a more selective place, right? And that's a, they never tested those waters. Now, another very important fact that we discovered how likely were students from different income groups to undermatch. And this shows that the percentage of students who undermatched based on their family income on the left and on their parents' education on the right. And there is clearly a pretty dramatic relationship to family income in the extent of undermatch. So more than half of students from the bottom quartile in income, even if they have more than half students from the bottom quartile undermatch, and only about a quarter from the top quartile undermatch. And something analogous is going on with the level of parental education. So this suggests that in addition to the general concern that people may be making choices that don't maximize their opportunity is a concern that those choices are concentrated among the less affluent students. So this seemed to us, as it did to our colleagues in Chicago, Melissa Roderick and colleagues at the University of Chicago, a very important thing to focus on. And Melissa and company have done a series of studies related to this phenomenon for students coming out of the Chicago Public Schools, which is really very interesting and they were an important stimulus to our end advisor to our work. I wanna emphasize one point that I think is sometimes missed here. We often find this phenomenon talked about in terms of students failing to go to highly selective institutions. So if you think there is a, for example, a group of what I call the snooties, which are sort of the 30 or so most selective universities and liberal arts colleges in the United States, they prefer to call themselves Kofi, the consortium on financing higher education. And if they had let McAllister in, I might have also adopted that name. And you think about, well, and people at Williams wring their hands about how can we recruit more of these students who could actually succeed if they came here. But we found in our data set that this was a very widespread phenomenon at different levels of selectivity. And when David Leonhardt from the New York Times was working on the piece that wonderfully came out on the front page of the business section of the New York Times on the day the book came out, anybody who wants to get attention to a book, give it to Leonhardt two months in advance. We found, as you could see here, consistently as you look at students with different levels of qualification and graduation rates, that if you have a high level of qualifications and you go to a very low selectivity place, you pay a substantial price in terms of likelihood of college completion. And if you have a low level of qualifications, you also pay a substantial price in terms of college completion. So we're not simply talking here about moving people from Roxbury Community College to Harvard University. We're talking about moving people from Chicago State to UIC, from UIC to Urbana-Champaign, from Urbana-Champaign to Northwestern. Those moves seem, there seems to be consistency in this phenomenon of undermatch. We were focused mainly on less selective and more selective four-year colleges, but in North Carolina we had data on people who entered community colleges as well. And looking at the same outcome of bachelor's degree in six years, we examined the likelihood that you would achieve a bachelor's degree in six years if you started at a community college. Now, community colleges have a lot of different categories of students included in their population. And this was a culled group of students that we looked at. These were all students who had taken the SAT, who were attending full-time and who indicated that their aim was to get a bachelor's degree. So a lot of the heterogeneity that you normally encounter in trying to figure out community college data was accounted for. And we used a propensity score matching to identify students whose characteristics were those that most typically would have started at a community college in our data, which were the ones over here, which meant low test scores, low family income, low high school grades, and people who were quite unlikely to start at a community college based on the typical background of community college students. And we found a really big difference in the probability of getting a bachelor's degree within six years if you started by going full-time at a community college. And data like this have been analyzed in several other states and similar kinds of results have been found. And there's also some work that's more difficult to do with national data that suggests something similar. Community college people really, really hate this graph and this message. And the conversation I've had with folks is often that's not true here. That that may be true in North Carolina, but that's not true here. And they believe that. And what I do is ask them to send us the data. And nobody has, and I don't think this is willful on their part, they are influenced greatly by the stories of the people who succeeded, I think. And they are influenced greatly by the fact that this population we're looking at is a relatively unusual one in community colleges and many, many people have very good reasons why they don't finish in six years. And we don't think this is trash in community colleges. I think that's a very important point. One thing that's really important is that in our data, if you entered a four-year institution after completing an associate's degree to community college, your chances of success were very similar to people who had started by entering at a four-year institution. The issue wasn't that people came out of community colleges with associate's degrees and were ill-educated and couldn't cut it in the data that we were able to look at. The issue was they didn't come out with an associate's degree. It was attrition within that population. And I think the people who are doing this everyday are overly influenced by the success stories and excuse the people who disappeared. That would be my story. And of course it's quite possible that, some states do things rather differently and it's quite possible that some of them really have much greater success and it would be great to document that because that would be a real opportunity to learn. But as far as we could see this is a phenomenon of a similar kind to the undermatch within the four-year, the population of four-year institutions. So that's it for basically the summary of the evidence that we developed in our book. Other work has been done since then. I know Jeff Smith here has been doing work on mismatch and so on. I don't, we haven't done any other work on this. After writing a book like this with Bill Bowen and having a full-time job, I'm still in rehab. So we haven't tackled the next chapter. And there are lots and lots of good questions that no doubt partly owing to the limits of our own skills and imagination and partly owing to the limits of the kind of data we have, there were a lot of questions that certainly could benefit from much further research. And we'll talk about some of those as we go along. But I wanna think about what may be going on in producing this phenomenon. And a lot of this is speculative. I mean it's all speculative in terms of the evidence that we have within the confines of our book. But we certainly read a lot of stuff at the time we were doing the book. We talked to people at several of the campuses where we were working. And we've looked at other people's evidence from interviewing students. And there's also evidence that we now have because of an outfit, because of an effort in the city of Chicago to try a random control study to affect the outcome here by giving intense college counseling to students who seem to have the promise based on their junior year records of getting to a more selective place. And so through that avenue in the pilot study of that we've learned more about how students are experiencing this college selection process. So if you think about what leads people to undermatch, one obvious question people may ask themselves is can I really afford to go to Northwestern if I'm a poor kid in Chicago? Can I really afford to go to DePaul? Which is a very different question as I'll say in a second from Northwestern. And to some extent for many schools the answer is no, you really can't afford it. There is a real problem about price and adequate aid and how much borrowing you're really wise to do to go to the more selective place. On the other hand, I'm pretty persuaded that a lot of people are poorly informed about what it would really cost to go to a subset of the more selective places that actually have quite a bit of money. And I wanna emphasize the subset. I was talking to somebody this afternoon in my office who recently visited Albion College in Michigan and they don't have the money to finance a lot of low income students. But McAllister certainly would finance any student who was admitted and many of the most selective schools are in that situation. But I think a lot of people don't know and the system is terribly complex. I think another non-trivial factor is parents are scared of putting a choice like this in front of their kids because they're afraid they won't be able to finance it. And they don't want their daughter to fall in love with a place and then tell her we can't pay for it. And I think they therefore are inclined not to enter the game in the first place, not to apply and well let's find out, right? Let's see if we can afford it. But if you find out then as a parent you're in a box. Like how much am I gonna stretch to make this happen? I think that's one of the things that goes on. Does it really matter where I go? I think one of the reasons you see relatively low levels of undermatch among college graduate families and affluent families is they're well informed about labor markets for more highly educated workers. And they know it matters. They certainly know it matters if you're gonna go on to law school, medical school, business school, et cetera. And it matters in a lot of other careers as well and it matters in networking. It matters in a lot of different ways where you get your degree, let alone the fact that you're more likely to get it. But for families with a first kid going to college for the first time, going to college is the achievement. And the idea that going to college is just a stepping stone to the real achievement. Will I fit in? This is a question for the parents as well as for the student. We don't understand at a place like the University of Michigan and certainly at places like Williams and McAllister, what a foreign country a selective college is to a lot of first generation college students. I always notice, I'm challenged in terms of my sense of direction personally, that I get lost a lot on college campuses. And college campuses remind me of small New England towns that everybody knows where the business school building is. And if you don't know, you don't belong here. That's certainly the way I felt the whole time I lived in Williamstown, Massachusetts. If you need a sign, we don't want you. And just think the physical map of these campuses is terribly confusing. The conceptual map of a college catalog that's got, I don't know, at Michigan, probably hundreds of thousands of courses or something. I mean, we assume people know how to navigate that pretty well. And that's wrong. We assume when they landed O'Hare, they're gonna know how you get to Northwestern. And that's wrong. We just don't comprehend that. And I feel like as a college president, I was way too dense about that for too long. And that's a real problem. There's also the problem, as my friend Morty used to say when he was president of Williams, that in some parts of the country, there are lots of people who want their son or daughter to have a Williams degree, but they don't want a Williams education. Because a Williams education means challenging a lot of the values that they grew up with. And that's the scary thing. Finally, in my list, and in a second, I'm gonna invite you to contribute to it. Why can't our daughter, our son, just stay home with us? For a lot of families, there are very pragmatic, as well as emotional reasons why there's a need to have that son or daughter at home. For families who are not native English speakers, sometimes that child is their communication with the Anglo world. And that's important to them. Sometimes the child is an important source of financial support, and that's important to them. And sometimes the child is an important source of childcare and other kinds of activities, and the list goes on. Plus, certainly in many families, the emotional pain of that separation is great. I remember dropping my son off at Wesley and getting back in the car and crying for 10 minutes before I could drive because the idea of that separation was so strong. But I mean, I've been in this world, I understood this was the right thing to do, this was gonna work out well. If it's not in your background, it's a really different thing. So I think there are lots of reasons why this may be a choice that's not optimal from an educational and economic point of view and where the right kind of intervention might make a difference. Let me, before I go to the next slide, which is gonna raise some other kinds of questions, I would be happy, and I'd like to do this from here out to just invite your observations about this question of what is it that might be producing the undermatch or reactions to the story that I've told. Let me say that on the next slide, I'm gonna talk about a question of whether this is causal. So you're all eager to ask that, but you might hold off on that particular question. But are there observations, anybody would care to share about this? What would make it, and the questions here are really frantic. What makes it particularly likely that people from low income or disadvantaged backgrounds are more likely to undermatch? Yes. I talked about intensive college counseling programs, and I think one of the big things is that students maybe don't even know these colleges exist. If you're talking about students in public school in Harlem, they've never heard of Williams College. They don't know anyone who's been there. So how are they gonna know to apply unless somebody directs them that way? Yeah, so yeah, I think that's a valid observation. I think in general, college counselors are really guidance counselors who do college counseling part time. They're very stretched. There's about one for every 500 students in American high schools. And they're not very knowledgeable. And they have a lot of other responsibilities. And they certainly don't have the kind of time or background to sit down and work in depth with an individual student about these kinds of questions. And I wanna emphasize it's not their fault. I mean, they're asked to do an impossible job. Yes. You kind of mentioned at the micro level about not only the family, but is there maybe also some play of feeling that they're leading behind their community and kind of the roots. And if they go to this new world, what does that say about how they can act as a community? I'm sure that's right. I think in Chicago, I think Melissa and company found that one of the attractions of the local community colleges, that's where their friends are going. So they can stay together. That's, way back in 1968, Sandy Jenks and David Reisman wrote a book called The Academic Revolution. And one of the things they emphasized was that a lot more people were going away from home to go to school. And they really were seeing the beginning of the national market for higher education. They actually suggested in that book that maybe colleges should give a year's credit just for living away from home. And forget it, they're not doing it at Michigan. You're not gonna get that. But it's an interesting thought that part of what makes it easy to stay home may also be a thing that interrupts the possibility of certain kinds of new experiences that would result in you growing. But may also alienate you from things that you value tremendously. Yes. Yeah. And also sometimes perceptions of a degree of difficulty when you don't have peers or older siblings or someone that's gone to a more elite school that it'll be too hard. Whereas a lot of people who've been to very elite schools will almost say the opposite of on a congealed level, like the hardest thing about this place is getting into it. And the work is not necessarily exponential degrees more difficult than something slightly lower to people who are afraid they won't succeed. Yeah, I suspect that that's true. I think also it's true that many people would have no way of anticipating this, but that there are a lot of positive aspects to the peer effect of having, being surrounded by students who have acquired a certain amount of diligence at least until sometime on Thursday. And then they get it back on Monday. And also engage with their fellows in that kind of interaction. I mean, peer effects of one kind or another seem likely to be important to the underlying phenomenon that you are more likely to graduate if you go to a more selective place. But they may also, you may really misestimate that when you're deciding where to go. Yeah. I think there's a growing population of students from citizenship or legal status will dampen their aspirations because they can't get public benefits to financially or just making it possible, especially if work prospects seem different. Yeah, no, I think the whole set of issues that surrounds the opportunities facing the undocumented people in work and in school are certainly a factor for a large part of these populations. Let me, I'm sure people have more ideas. Let me push on and we can come back to any of this. I've told the story on the basis that our descriptive observational analysis in fact is capturing a real causal phenomenon that the people who undermatch really are like the people who matched and if they made the other choice they would have results like the people who matched. But as in almost everything involving education that people study, you have to worry about selection effects, self-selection in this case because we're talking mainly about people who didn't apply, not people who got turned down, so one story you can tell and it certainly doesn't have zero validity is that you know something about yourself that doesn't appear in your high school transcript or in your SAT score, that there's a reason, private information to you relative to the researcher that you can predict that you wouldn't have this typical experience if you went to the more selective place. And therefore there's a rational element in your judgment not to try it because you're not gonna have the experience that those folks had. That's gotta be the case to some extent for folks in this group. There are a couple of reasons why I'm pretty strongly inclined to believe that that's not the bulk of this story. One reason is that big difference in income groups. If there is a set of things that we're missing about people's personal qualities or something we can't get at, I haven't thought of a good story maybe you will about why that should be unequally distributed across income classes. That people with lower income are better at predicting the effect on them of going to a place which is like nothing they've ever seen than people with higher incomes who know more about those kinds of places. That seems surprising if it's true. The other thing is the effects are pretty big and pretty consistent across a lot of different choice sets and it doesn't seem like the same quality would turn up everywhere if there is some hidden set of qualities. That's not a, none of these are knockdown arguments. God knows they're not and of course what you wanna do is test this stuff empirically but that seems like a factor. And then I think something that's not really based on quantitative information. If you look at the story of how folks go about making these choices and particularly one of the papers in the consortium on Chicago School Research series followed students who went to one of Chicago's seven selective colleges, selective high schools. These are the high schools like Walter Payton and six others where you don't get in unless you have high test scores and high grades and good assessments from your teachers and you have good counseling and you get the right schools on your radar screen and you intend to apply. And then in too many cases, you drift away. I mean, you read this stuff, the only word you say is you drift away. Oh, God, I missed that. I just wasn't thinking about it. I'll be fine, I can just do this thing that all my friends are doing. It's not that important. My mom is so excited that I'm gonna graduate and as long as I go someplace, everything's gonna be fine. This doesn't sound, I mean, a lot of these, the experience doesn't sound like people who have some grounded reason for thinking that their experience is gonna be different. We also know from a variety of sources that including sort of personal testimony but also quantitative information that the information lacks are very big. That wonderful study that Bridget Long and Eric Bettinger did working with H&R Block to see what the effect is not on college match, but on college application simply by providing people who came in to get their taxes done with the opportunity to get their financial aid form filled out at the same time. And it had really substantial, significant effects on the probability of college going. So there's just a lot of reason to think we're in an information poor environment and that that would have a much bigger impact on low income people than on people from more affluent communities. But you know, if you really wanna be sure, you want the modern biblical blessing of an RCT. And there are a couple of such studies underway now. Caroline Hoxby and Sarah Turner are working on very low cost, very widespread intervention which just provides targeted, adapted to the individual information in a letter. And there are a couple of variations on this. They're trying to see if they can move students' college choices by doing that. We are tangentially involved, those of us who wrote this book, in MDRC's effort in Chicago to try a very different high cost, intensive intervention to move students' choices. Now, if we move their choices, we then have the raw material for a test of whether the results turn out to be like the results you get from the observational data. Do these folks who made a different choice, we know on average, assuming that the experiments have a positive significant outcome, we know that they did a significant number of them chose to go to a different place than they would have otherwise because they're matched at random to people who didn't get that treatment. And we can see if in fact they turn out to be more likely to succeed in college. So, you know, I'll be around, I hope, in 2017 or 2018, check back. And we'll let you know. There also may be more ingenious things to do with these data than we thought of, or there may be other data. Again, I know Jeff Smith, and I talked to him and a graduate student who is right over here, Nora, right, are working on these problems. And what they're doing sounds like really interesting. So I think they're very good questions. I'm gonna go ahead to talk about policy implications on the assumption that there's a real causal effect here. But before I do that, if anybody wants to raise any questions about this causal analysis, I'd welcome. Yes. Just in kind of separating the mechanisms versus taking the easy way out versus kind of an information thing, did you happen to see if there was any sort of patterns of sort of taking the easier approach in high school like kids with really high test scores taking fewer AP exams or kind of, maybe not having less frequency of taking calculus or anything else that makes it look like a pattern of the easy. Thing is, we were matching people on those characteristics, right? So if- On course taking? Well, not down to the level of course taking, but on the level of, well, you're right, with GPA you could have gotten the GPA with an easier load. No, we didn't do that. That's a very interesting thing to think about. Did, were there signals in other parts of their behavior that would be consistent with them being less diligent or something of that kind? I mean, I think actually one of the important, again, not a new finding, but one of the important things we saw in this study is that high school GPA is far better at predicting college completion than test scores are. And again, we don't have any way of testing exactly why that's true, but I think it's the Woody Allen effect. 80% of life is just showing up. And the way you get a good GPA in high school is you just do the work. And that's a very good predictor of whether you're gonna show up in college and do the work. Kind of follow up to that. I work in a K-12 environment and one of the things that troubles me greatly is that we have in this country about a 30% high school dropout rate. And obviously in pockets it's much greater than that and obviously with populations that are more at risk it's even higher. And there are lots of politics and issues around the public policy that we've implemented in a K-12 environment that I think have impacted that in terms of why kids aren't showing up and why kids are dropping out. But I'm wondering about this whole under-match phenomenon in relationships. Those kids are really at risk. So it's not like I'm deciding to go to Northwestern or out in college, but like can I even get out of the gate and it kind of goes back to your comment about are these students that are becoming disengaged with education and not connecting to the dots about what does this mean for my life disengaging at such an early time that then this sort of idea would then translate and how could we intervene with that? I know that's a big question. It is, yeah, it is a big question and there are various ways in which you may think the junior or senior year or high school is too late for a lot of folks and a lot of folks are already in the process of dropping out at that point. There's a lot of talk and interest in early intervention programs to get college on people's minds when they're in eighth grade or ninth grade. Evidence is not real encouraging about our success so far in finding ways of doing that and it's related partly to the complexity of the whole process because it would be really nice if you could tell people when their kids are 10 that this is what's gonna cost you if you get a mistake with it, but we can't. I co-led with the person who's now my wife project on Rethinking Student Aid and we had this idea that I really like that should start putting money into savings accounts for kids when they're like six and just do it automatically based on their AGI, the family's AGI and make the money available only for post-secondary education. So not a thing where you can change your mind and take it out, pay a 10% penalty and get a boat. I think that actually could really change the way people think about this stuff if you did it on a serious scale. That would however require money and right now that's the scene today in the cards. Yes? If you just find under matching in terms of whether or not a student applies to a given school, how has the changing more efficient methods of applying to schools such as the common application which is seen, which now has bothered 450,000 students so has that had any effect in reducing the reduction? One would think so. We don't have any evidence about the effect of that. You know, it's still true that the, I think it's still true that the modal number of applications a student makes is one and they get in. That is non-selective places dominate in terms of just the sheer numbers and in some ways getting that up to three may be like the thing to do. I think things that smooth the path, not having a quirky eccentric essay question on your, does Michigan have one? Help, what really seems likely to be effective is to have an automatic waiver on the cost of taking the SAT. Now that's not an issue in Illinois because the ACT is required of everybody, but an automatic waiver of the application fee as well. That's actually more important. That it's just, the way it is now, colleges say you send in this form, you report something on your income and then we waive the fee. But if you could just give high school guidance counselors forms and say we trust you, give this form to anybody that you think it deserves it and the fee is waived. I mean that, that just seems super simple. In Northwestern automatically waives the fee if you're from a Chicago public school because 86% of Chicago public school students are classified as low income. But they could also be waiving the fee for everybody who applies from LA Unified or everybody who applies from another city with a high low income population. I can't believe that it's a big revenue source for Northwestern University. But I do think these simplifications are steps in the right direction. Brian, you wanted to raise a question. I was just thinking about the causes. I was struck by the fact that the under-match, while there was a gradient in family income or parental education, even among the top income quartile of those with parents when they graduate school, I think I remember 30%, there was some huge number of folks in that category who were still under-matchable. And I would have to think that... Oh, yeah, I know what you're talking about, yeah. It kind of, the reasons underlying the under-match, for that population, maybe very different than the reasons underlying the other population. I never just did it. No, it would be, yeah, that would be interesting to think about. In my years at Spencer, looking at lots of research proposals and so on, one thing that strikes me is that there's a very strong tendency not to study questions like these for the upper income population, or I think there are huge variations in how successful affluent school districts are with their students. And even though what we really care about, I hope, and I certainly care about, is the low income and disadvantaged populations that doesn't mean you can't learn from studying other things. And I think questions like that could really illuminate the underlying phenomenon. So I now wanna talk about assuming that we've got a causal thing that's a significant fraction of what we're seeing by way of under-match. What kinds of implications might we wanna draw as people who are interested in public policy from these facts? And I have more questions than answers, really. So implications for action. This first isn't necessarily what you would call public policy. And these are cases in which, one way of putting it is you can treat the, it's like the small country model in international economics. You can take the conditions in the world as given and independent of your actions. So partial equilibrium analysis is legitimate and all that stuff. If you have a niece or a nephew who is facing this kind of dilemma and has a boyfriend, girlfriend, whatever going to some crappy place and is gonna follow that person, just ask her if it's really, really serious. Russ Whitehurst, when we presented this thing at Brookings as part of our kickoff for the book, pointed out that he went, his college choice was dictated by where his girlfriend went. And I was, I like Russ and I know Russ and I actually think he's a very good looking guy. But I was tempted to say, Russ, if I were you and I found a woman who wanted to spend time with me, I would definitely stick with it. But. But. But I think more generally at the level of the individual student, making sure that that student, if it's somebody you know, and you know there's not some particular reason why this would be a mistake, to consider seriously extending their range of applications. It just seems like kind of a no-brainer that I certainly, I get all the time as many of us do, I'm sure, get requests for advice on things like this. Does it really make a difference and stuff like that? And yeah, yeah, it does. Now for high schools and for school districts, they're very difficult choices to make in this regard. You do have limited resources. You have in, let's say in the city of Chicago, except for these selective high schools, you're gonna have relatively few students even on a fairly generous definition of selectivity who could benefit by more intensive counseling at this stage. When we were working on the, thinking about the MDRC pilot for this intervention in Chicago, and there are about 115,000 students in the Chicago Public High Schools, 400,000 students all together. And we tried to guess what was a reasonable ACT cutoff for people who might be amenable to this kind of treatment and to get anything like a reasonable number of people in a single high school. We had to keep lowering that thing from what we expected. Half of the students in a typical Chicago high school are not gonna graduate from high school. Of the ones who do graduate, a lot of them are unlikely to do more than get a certificate of some kind from a community college. Does it make sense to devote your scarce resources to improving the placement of the more promising among the students in the school? That is not a trivial question. And we actually talked about that with Ron Huerman when he was the CEO of the Chicago Public Schools. And Ron's judgment for what it's worth was the effect on how teachers and students would feel about a significantly larger fraction of students coming out of regular Chicago high schools into strong places and getting bachelor's degrees, even though the absolute number was a very small fraction, would be great. And that investing resources in that way would have a kind of halo effect on the system that was worth spending money on. Now in fact, Ron wasn't gonna have to spend money on it because we were funding this thing. But this thing's not worth anything unless ultimately people will spend money on it if it works. And that's, they're very difficult choices when you're involved in a high school with a lot of disadvantaged students and big equity choices involved. And the same kind of point goes for school districts. And obviously it would be good to improve the training and effectiveness of guidance counseling in general. Is it a good use of resources to devote special attention to information for guidance counselors who will say specialize in students with better prospects? It's not clear you really have to do that. Some of what you need could actually be automated. You could, as Chicago does now, generate a match list for every student based on their records and they give that to the guidance counselor. So there may be good ways to do this at relatively low cost. And certainly Sarah and Caroline are working on very low cost interventions. But there are policy dilemmas at these levels. And then one of the questions I talked about with somebody during the day to day, I had, by the way, I had a great time today. There were a lot of really wonderful folks here, graduate students, master students, PhD students, even some of the faculty who are really fun and rewarding to talk with. Colleges face a kind of double dilemma when they think about this. They all, in principle, wanna enroll more well qualified, low income and low income students and students of color and first generation college students. In practice, one dilemma they face is they can't afford them. They then either loan them up, which is a very scary thing to do. I'm not one of the people who thinks it's not smart to borrow to go to college. I mean, this is an investment in yourself. It's a reasonable thing to do. But it's one thing to borrow $30,000 over four years, which is about what you can do with federal loans. It's another thing to borrow 80 or 90 and most of that money will be private loans, which will not have the kind of favorable terms that and the deferments and income based repayment and so on. That's a terrible choice in a lot of cases. And the other dilemma that they face is if they wanna avoid doing that, they may become need aware. And in fact, about the time I left McAllister, we became need aware because we didn't think we could sustain being the college we wanted to be and guarantee that we would admit completely without regard to ability to pay. That's a very tough dilemma. But another much more pragmatic point about the college piece of this is colleges usually go out and try to recruit students for themselves, for their own college. So, or maybe you'll have a little group like eight of the best is an example of that. And they will take a tour to the city of Chicago and they will indeed hit some of these schools and they will try to isolate the small group of students whom they should talk with, try to get help at the school to do that, so on. But the fact is that the probability that you're gonna get students who are A, gonna decide to go to a selective school and then B, go to yours. That's a tough combination. This is an area where collective action by colleges just seems to make a lot of sense. Just a lot of sense. And if they have any kind of public spirit at all, collective action not to say, can we find the students who could get into McAllister? Because there's gonna be very few in a, or Williams in an urban school system. But can we raise the aspirations of students more generally? Can we invest some money in getting some students and some counselors on our campus, on campuses like ours, doesn't have to be ours. But to make this world less foreign and less strange. Can we find representatives among our alumni body with whom these students are more likely to identify? And can we do that in some collective way so we can do it? That just seems like such a waste of money to try to do this one place at a time. So it's an obvious area for some kind of cooperation. I also think it would be super easy to get foundations to fund this. Not the Spencer Foundation, but other foundations. National implications. So stepping out of the microeconomic world into thinking about system level effects. It seems to me if you think about this from an equity point of view, it just seems to me to be, on any plausible view of fairness, that gap in the undermatch rage just seems like something to try to address. Intuitively, it seems extremely unfair. Here you have somebody who has had the level of achievement in high school and put up the scores that are needed to actually qualify for this kind of experience. And to the degree that the choice not to do it is something that we can influence, that we can, you know, through better information, and through making it cheaper that we can do something about, it's really hard to see an equity case we're not doing it. And I think if you think about then the policy implications of trying to address that, there may be many, and I would love, and I'll stop after I've said these two things to get your reactions to that. One policy implication is, for whatever resources you have, you really should try to target them on people who will benefit from them in the sense that it will influence their choices. And to do that means being willing to introduce one way or another net price differentials between people who have more ability to pay and less ability to pay. You know, there is a fair amount of evidence and there's a little bit of it in our book that keeping the price down for higher income people, top part of the income distribution, certainly makes them feel good, but it doesn't seem to influence very much their choice about whether to go to college or their likelihood of graduating. But for people lower down in the income distribution, improving their financial opportunities seems to have effects on their behavior. So a concrete policy implication is when states use limited resources to keep the tuition relatively low for all students, as high as it is, it's still relatively low rather than having somewhat higher tuition and somewhat more need-based or income-based even better student aid. Seems like a mistake. And the federal government in the last three years has actually increased by over two thirds the amount of money going into the Pell Grant program. And that has been a significant buffer against the declines in state financing of higher education that has been occurring. I also wanna underscore that the tax expenditure on college tuition, which is partly deduction and partly credit, more than doubled in one year. And one quarter of that money goes to people whose family incomes are between $100 and $180,000 a year. And it's an interesting feature of our political economy that there is an intense debate about what can be done to manage the cost of Pell. And I've been a participant in that debate. Been in that debate. As far as I know, absolutely no debate about what can be done to manage the cost of the tax credit. So there. The other thing I think that actually is really important and in other lives I've done quite a bit of work on this is so have other people at Michigan most notably, Sue Danarski. Making the system simpler. There is a real problem that people sometimes just can't afford to go to the place that would be best for them. But there is also a real problem that people don't know what the actual price would be at the place that would be best for them. And we have designed a system that seems virtually, it's almost as if it were intended to keep that information secret, which I don't actually think is the case. But Sue's work shows that when you, this is mostly about whether or not you go to college, not about college choice. But I think the same principles would extend to college choice. When you have something really simple and clear that lets you know the price is gonna be low, it has a significant effect on the probability that you will enroll. She did this as many of you will know, looking at the Georgia Hope Tax Credit and other similar tax credits. And looking at when the Social Security System ended the survivor benefit that meant college was cheaper for you if one of your parents had died than it would be if that parent was in the Social Security System. That meant a sudden increase in what had been a very simple, increase in price to people who had been receiving a very simple, well understood benefit. And she did difference and differences analysis and found significant effects from these changes. So Sue has pushed for getting the FAFSA, which is now 108 questions or something out of the back of a postcard. And in our Rethinking Student Aid thing, we recommended the somewhat simpler step of abandoning the FAFSA altogether and just having treasury base Pell Grants on AGI. And we've made progress. There's been actually a lot of progress. Particularly the tax people, the IRS is talking with the, and cooperating with the Department of Education people. So on the fairness side, let me see if people would wanna make any contributions or ask any questions about that side of things. Maybe there are other things we should be doing as well as pricing aid and information. Ah, so the information problem strikes me in here that I was feeling really, really hard. I've been a part of it. So those of us who were in the business actually know about the whole range of schools that we had thought about sending our kids to you and that's sending our kids to you, et cetera. And that's an enormous amount of information. That we actually accept, right? As, oh, everybody knows that. And in fact, almost nobody knows. And so the question, ask the question of 1,000 high school seniors, juniors here, what's the best school you could get into or even ask the question, rank the following 50 schools? And a very large number of people will have that. I think no idea about match, under match, whatever it is you're talking about. And so just trying to imagine getting sort of systematic, reliable, well-believe information that when a lot of kids who read your paper, read the book, right, to have some notion, what's the best school I can do? How do I fight, how do I know? In fact, what's the best school? I'm not so bad at school. All that kid has to do is write a one-page letter to Bill Bowen that demonstrates that he actually read the book and Bill will get him into Princeton. Once again, you and I know that. No, I understand. I think one important thing, and this is a very important thing, is not to be too ambitious about this. There is a tendency to think that the key thing in the under-match problem is getting people in the Northwestern. I keep going to that example of getting people out of state people in Michigan, maybe in state people these days in Michigan. But in fact, you can be a lot less ambitious and still make a real difference. And I think in a lot of cases, at least for most people, if you simply gave them guidance on public options in your state, you know, and you didn't, I mean the information overload thing is a real problem if you push very far. But that much, that's a fairly short list usually. You can have categories in that list. And I think that kind of information is more feasible to think about. And the other thing is information about how the system works. That is the peculiarity that you're actually not gonna find out the price until next April. But you don't have to decide until after that. With these counselors that have been working with Chicago Public Schools, who have come in and talked to us about what their work is. These are amazing young people who themselves solved this conundrum and came from backgrounds similar to the students they're working with. And man, you hear about every day for a week, I was on the phone with this kid or he was in my office every day during that season and I to just understand what his choices were and how to make it work. We can't do that generally, I recognize that. And it is a really, really hard problem but if you could simplify some of those choices down for most students, I think you might be able to get somewhere. But of course it's not only information, it's information with all kinds of emotional weight attached to it but it makes it even worse. Yeah. One more question, maybe someone can pause. Oh, are we done? Well, we're technically, this is, I'm just getting started. Sorry I kept wandering along, I was just having such a great time. Thank you all so much.