On Undermatching

About July 28th, I was asked to participate on an August 4th panel at The American Enterprise Institute, a think tank in DC, to react to two papers presented at their invitation-only conference, Matching all students to post secondary opportunities: How college choice is influenced by institutional, state, and federal policy.  With just a week to prepare, I juggled some schedule commitments and agreed to participate.

The premise of the conference was to move away from the traditional discussion of “undermatch,” which was made popular by a University of Chicago report on the Chicago Public Schools.  Since then, many definitions of undermatch have surfaced, but essentially the issue focuses on high-achieving, low-income students who don’t apply to, or don’t attend the most selective institutions they should consider or were admitted to, respectively.  It was hoped we could focus on a larger group, rather than just the very high-ability, very low-income student, to encompass a discussion of the fatter part of the bell curve, and to expand our focus on the undermatch problem.

College access and expanding opportunity are hard things to wrestle with and to get your head around, as I’ve written here and, in case you don’t have a subscription to the Chronicle, which I’ve expanded upon here.  But I’m always happy to engage in the discussion. I think it’s vitally important.

My conference role was to react to two papers, one written by Lindsay Page and Jennifer Iriti of the University of Pittsburgh, and one by Michael Bastedo of the University of Michigan. You can read the Bastedo paper, titled Enrollment management and the low-income student here. The Iriti/Page paper, titled On Undermatch and College Cost: A Case Study of the Pittsburgh Promise is not yet posted on the site.

I thought both papers were very good, although I’m not a researcher or an academic per se (or, per anything, actually). And, in a room where I literally (and I do mean literally) could not swing a dead cat by the tail without hitting an economist, I felt a bit like the guy who lifts his bowl and slurps his soup at a dinner with the Queen of England.

The Page/Iriti paper reports the effects of “The Pittsburgh Promise” on college attendance by students who graduate from the Pittsburgh Public Schools (PPS).  And the effect seems significant, if I may summarize: Students are both more likely to attend college, and appear to be more likely to consider and apply to “reach” schools when they know that as much as $10,000 annually is available to assist them. This is very interesting, of course, and it’s nice to have the research to back our intuition on it; I suspect at some point the costs and the economic benefits of the returns will have to be compared. Additionally, as I pointed out on the panel, this trend is even more compelling in light of the national trends that show college attendance of recent high school graduates is down over the same period.

The Bastedo paper is as good an understanding of enrollment management and the explicit trade-offs as I’ve seen from a non-practitioner. There were two critical points that were spot-on: First, that if we were suddenly able to change behavior among students–that is, if all eligible students applied to college, and especially if they applied to colleges they’re “matched” with–less will change than we think.  Capacity in higher education is generally fixed, and when it comes to enrolling poor students who need a lot of aid, it’s very constrained. It’s thus unlikely the pie is going to get bigger; that is, absent any fundamental shift in philosophy among the most selective institutions, we won’t see larger numbers of low-income or first generation students admitted to these places, as doing so to a much greater extent would not be in those college’s best interest.  Second, I was pleased to read his agreement with my long-standing belief that virtually every single factor in the admissions process is stacked against low-income students.

The format was supposed to be 5-7 minute summaries of the paper by the author, and then 5-7 minute reactions from me and the other panelist, Lesley Turner of the University of Maryland. Time has a way of getting away from people who are excited about their work, of course, and 34 minutes in, I was still waiting to go.  Needing to leave some time for Professor Turner and for questions, I ended up collapsing my remarks a bit in light of the time constraints, and in light of some excellent points made in a prior panel by Awilda Rodriguez of The University of Michigan, who had already pointed out in an earlier panel that many factors that determine whether or not a student matches with the “best” institution were inaccessible to her.

My reactions were naturally colored by my own initial thoughts about the term “undermatching,” as well as a very informal poll of fellow practitioners in college admissions, including the college, the high school, and the independent counseling fields. I started with two anecdotes: First, as the child of two parents who never even attended high school (not uncommon for people born in 1916 and 1917), let alone college, I told the story of my brother, who was being recruited in 1966-67 to play football for Dartmouth, but ended up enrolling at the West Des Moines Institute of Technology to study electronics.  So, undermatching is not a new concept in our family.  (And my brother did fine in life, despite the fact he turned into a Republican, which probably would have happened at Dartmouth anyway, so it may just be a genetic mutation.)

Then, I told the story of a dean of admissions at a prominent, super-selective university who told a conference audience that a) her college defined low-income as under $60,000 (while the median family income in the US is about $55,000 by the way), and b) that despite “spending all their time looking,” they discovered there “just were not enough low-income students in the US who could do the work at her university,” due to university demands like writing papers and working independently, which are apparently unique to this particular institution of higher education.  (The Astors threw a tea and cucumber sandwich party to feed the homeless, and none of them arrived, probably stopped short of the penthouse by the security guard, if you’re looking for a good analogy here.)

I went on to outline my conceptual problems with “undermatch.”  First, our obsession with defining colleges as “selective” and summarizing everything one presumably needs to know in that one term.  I know the Hoxby and Avery study that suggests poor kids are better off going to the most selective institution they can is the flavor of the month in this discussion, and I expect another report to come along at some point debunking it (that’s the way research works, of course). But regardless, I opined that “selective,” “high median scores,” and “high graduation rates,” are essentially the same thing, as I’ve tried to point out on my other blog.  While people with academic pedigrees see these as attractive traits that should be pursued at all costs, it’s just as likely students from low-income, first-generation families, especially those from non-majority populations, see these traits as markers of white, privileged, foreboding, exclusionary, and, of course, unwelcoming.

Further, the Hoxby and Avery study points out that poor students at highly selective colleges do well because their net cost is often very low.  Of course, highly selective colleges can afford to be generous because (think about it) they have fewer very low-income students in the first place. (I feel compelled to point out that Harvard is actually to be commended for this; I think they rise above the rest in their efforts.)

Additionally, something as simple as distance, and the cost of traveling several states away; or the perceived need to work to support the family, can all play into choice formation and final choice. Additionally,lots of anecdotal evidence suggests culture fit is just as much an issue as academic fit in happiness and graduation. Students may stay at a university because they feel their only chance is to tough it out; that does not mean they’re happy there.  Switching costs from an Ivy can be quite high.

In reality, “selective,” “high test scores,” and “high graduation rates,” combine to mean that the admissions offices at these places take virtually no risks in deciding whom they should admit.  Using test scores as a definition of intelligence or college-readiness only exacerbates this problem, as indicated by this chart of ACT scores, broken out by ethnicity and self-reported family income.  As Professor Bastedo recalled his observations of admissions officers who zeroed in on an applicant’s single application flaw and pronounced the student “incapable of doing the work here,” (based on nothing other than a whim), I thought to myself that test scores serve as the easiest, quickest, most quantitative way to justify this bias.

Some research on this topic seems to overlook these realities.  Yet almost unanimously, the practitioners I spoke to mentioned that “match” is only half the battle; that elusive, non-quantifiable quality called, “fit” is the other.  If it were not, I argued, we might as well tell students to go to prom with the most attractive person who offers; to marry the wealthiest person they can, or to buy the car with the most horsepower. We presume “the best” is measured by an input, and further assume that this will make people happy, despite our experience to the contrary.

Overlooked in all this is demonstrated by a single chart in the Page and Iriti paper: Of those whose qualifications matched them to a four-year, about 58% matched appropriately; about 7% undermatched; and about 38% did not attend college at all. The longest-standing problem, Unmatch, unable to attract much of the spotlight after all these years, doesn’t get the attention, while the new, sexier, more interesting problem with a name, Undermatch, gets the press.

Some of the proposed remedies to undermatch and unmatch focus on information channels, some on state and federal policy, some on aid allocation.

Too little, I fear, focuses on the other agent in the matching game: The colleges themselves, and the fundamentally broken admissions process that has remained unchanged in decades, and is still shrouded in mystery, especially at “the best” colleges, and especially to those who need to understand it the most if anything is to change.

I’ve suggested lots of ways to make that happen.  What do you think?

8 thoughts on “On Undermatching

  1. Excellent piece, Jon. The problem with economists is that they often assume a “perfect” scenario and focus on a narrow range of elements to observe. (Like physics teachers who say, “Assume a perfect vacuum” for the sake of an experiment. I’ve just reviewed three books that try to enlarge the field of vision at my blog, so your observations here feel especially on point. My initial enthusiasm for the Overmatching idea faded long ago so I’m glad it’s being reevaluated. Thanks as always for your clear, calm, and rational perspective.


  2. As always, Jon you raise the right and most interesting points. This time… where is the conversation about “UNMATCH”? And, yes, when are we going to seriously engage in active, meaningful conversation about changing the process overall? Of course I believe I know the answer to this second question. It comes down to whose interests won’t be served by that conversation. Thank you Jon!


  3. Thank you, I love your stuff. I teach at a nationally ranked STEM high school and I consult privately as an IEC. All your writing is spot on, and of course, your data is the best. Keep doing what you are doing to shine the light.


  4. Jon,

    Hope all is well and a great article. I just have a few questions and I think there are the same concerns you have which were stated indirectly. My lack of practical experience in the enrollment and admission process may become evident with my questions.

    Based on your comments is the issue they are addressing creating a domino effect on college enrollment management. If the under matched were matched correctly wouldn’t those seat displace current enrolled students? If a college is having a bad yield wouldn’t the focus on the under matched make it worse?

    I also find it a little surprising on how big an issue this is. My daughter is a teacher at a very diverse HS and she works with many of the at risk students. Based on her comments, I find it surprising that these talented student are really that under matched. Is it not just a perception of value that needs to be addressed. As you pointed out people select colleges for various reasons, now the colleges are telling us that students are making mistakes.

    It is amazing to me what people are concern about. We have graduations rates declining and transfer rates increasing and people are worried about this.

    If you get a chance over the next few days give me a call. Would like to catch up.


    610 789 1255


  5. Jon – This is great stuff as usual. I really can relate with your comments on the Bastedo paper. It amazing how many great universities our there can’t afford to support any more low income students, even if they are admissible.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s