Minority Graduation Rates

August 12, 2010 7:54 AM

Using its College Results Online database, The Education Trust has released two reports examining the black-white and Hispanic-white college graduation gap. The worst offenders? Wayne State University in Detroit, where fewer than one in ten African-American students graduate in six years, and CUNY Brooklyn College, where 19% less Hispanic students graduate on-time than whites.

At nearly two-thirds of the colleges and universities in the study, fewer than half the African-American students emerge with a degree. And though the vast majority of Latino students in the study entrust their futures to public colleges and universities, more than 60 percent of the institutions they attend graduate fewer than half their Latino students in six years.

Earlier this year, Ed Trust also released two reports on what some colleges are doing to close the gap.

posted by l33tpolicywonk (28 comments total)

I went to Brooklyn College, I'm white, and I graduated 11 years late.
posted by swift at 8:04 AM on August 12, 2010

Just because Wayne State has a low AA graduation rate doesn't mean that they are themselves causing the low graduation rate. We aren't randomly assigning colleges to people. It's entirely possible that the individuals sorting into Wayne State are already inclined to have lower graduation rates (ie, selection bias).
posted by scunning at 8:11 AM on August 12, 2010 [3 favorites]

scunning: "It's entirely possible that the individuals sorting into Wayne State are already inclined to have lower graduation rates"

The reports highlight universities with similar demographics all of which have low / non-existent achievement gaps. This is a policy / attention problem, not a demographic problem.
posted by l33tpolicywonk at 8:13 AM on August 12, 2010 [1 favorite]

I went to Brooklyn College, I'm white, and I graduated 11 years late.

Well... okay? And obviously plenty of Latino students at Brooklyn College graduated on time. This is about aggregate statistics, not individual anecdotes.
posted by Tomorrowful at 8:17 AM on August 12, 2010

"...19% less fewer Hispanic students graduate..."
posted by mr_crash_davis mark II: Jazz Odyssey at 8:23 AM on August 12, 2010 [4 favorites]

@l33tpolicywonk - Comparing on observable characteristics will let you know that the outcome differences are not driven by those observable attributes in the student and the student's background. But it will not help you deal with the selection problems I had in mind, which is unobservable differences in students that are correlated with the education decision. This is commonly called in policy evaluation "selection on unobservables". For education studies, it's maybe the single most difficult thing in assigning any causality at all. This definitely would be relevant in the endogenous selection of a college.
posted by scunning at 8:23 AM on August 12, 2010 [2 favorites]

scunning: "unobservable differences in students that are correlated with the education decision"

I don't know any of these potential characteristics that don't ultimately tie back to institutional policies, though. Let's say, for the sake of argument, that Wayne State, compared with other institutions of its demographic makeup, is admitting more students who are unprepared to graduate in six years. That's still a big problem - having some college and lots of debt is worse than having no college or having gone to a technical program is worse than having gone to a college that actually cares whether you do well and invests the resources to help you succeed.
posted by l33tpolicywonk at 8:27 AM on August 12, 2010

As a current Brooklyn College student (and, er, also graduating 11 years late) I'm racking my brain trying to figure out some sort of folksplanation for this, but I really can't.
posted by griphus at 8:28 AM on August 12, 2010

@l33tpolicywonk - just read the 8 page executive summary. I don't see anything in there suggesting they even control for observable student characteristics. But I could be missing it - I got bored at parts and started to skim. I only see columns with graduation rates by race, and a race gap calculation. There's no basic multivariate regression to handle the conditional expectation of graduation that I could see. I am entirely confident that there will be a persistent race residual when you regress graduation rates onto race, and that WSU probably still performs badly in that analysis, but it's still a long way home from there to conclude causality. This is just correlation, and having run too many race-related regressions in my short academic life, I know enough about policy evaluation to know I shouldn't even put stock in the regression analysis, let alone simple unconditional mean comparisons which is all this is. We do not know the counterfactual for these students at Wayne State, and that makes saying that WSU is at fault impossible. WSU may very well be doing better than the counterfactual, which might be an even lower graduation rate, but modeling that is extremely complicated.
posted by scunning at 8:30 AM on August 12, 2010

The reports highlight universities with similar demographics all of which have low / non-existent achievement gaps. This is a policy / attention problem, not a demographic problem.

Doesn't saying that assume that people categorized as being of the same race are more similar to each other and disregard that there may be stronger academic similarities grouped across race boundaries? That is - that people considered to be the same race are more homogeneous, academically, than heterogeneous? I guess that would be a question of methodology in the study, how they tried to control for it, but looking through some of the links I'm not seeing details like that.

What's making me think about this is basically the "is Obama half-black or half-white?" question that has been raised since he began his Presidential campaign.
posted by XMLicious at 8:33 AM on August 12, 2010

@l33tpolicywonk wrote "that Wayne State, compared with other institutions of its demographic makeup, is admitting more students who are unprepared to graduate in six years. That's still a big problem - having some college and lots of debt is worse than having no college or having gone to a technical program is worse than having gone to a college that actually cares whether you do well and invests the resources to help you succeed."

I cautiously disagree with this statement, but could be persuaded otherwise. Not to beat a dead horse, but we don't know the counterfactual for these Black WSU students. We need to know what their lives would've been had they not attended WSU or had attended some other school, but we don't know that. It's not even clear what the counterfactual is now that I write that - compared to no college? compared to UNC-Chapel Hill? Compared to community college? What's the comparison being made that indicts WSU as not doing its job?

WSU may not be doing as well as another school, but you're not warranted to draw that conclusion from simply comparing unconditional mean graduation rates by race. Even looking at the Black graduation rates at different schools isn't helpful - the Blacks who attend WSU are different along observable and unobservable attributes than the Black students who attend Berkeley or University of Tennessee. Therefore, because of those differences, you have to try and adjust the estimation with it in mind.

What you could do theoretically is maybe look at ACT score cutoffs. Say that to get into the school marginally better than Wayne State you need an ACT of 20. Why 20 and not 19? There's no reason probably - you have to draw the line somewhere, and at some point, there's a randomness in where the line is drawn. So compare the WSU students with ACT scores of 19 to WSU-prime students with a score of 20. Then try to adjust for the background characteristics if you're still concerned. This is called regression discontinuity and there was a very helpful issue in the Journal of Econometrics a few years ago detailing its development and usage in program evaluation in education and economics. That would require a lot of data, though, because you mainly want to look at the scores of that cutoff, test whether if it is indeed a true discontinuity, and then follow those students only.

I have no dog in the fight as to whether WSU is a great place or a crappy place, but I get a bit testy on causality stuff in education probably. Plus, my new meds for ADD are making me irritable as hell, and so I'm finding fault with just about everything today. :)
posted by scunning at 8:41 AM on August 12, 2010

What's the comparison being made that indicts WSU as not doing its job?

One piece of evidence that Wayne State has a graduation problem that's not confined to race: the school's overall six-year graduation rate is 33.6%.

Regarding selection problems: The point I'm making (or, trying to make) is unlike a K-12 school study, where location-based assignment means that schools have basically the kids from their neighborhood who may or may not be different from kids in the next neighborhood, Wayne State selects 100% of the students who go there on the basis of an application process it has created. Arguably, that comes with an implicit moral obligation to ensure that its students graduate not even in four years, but in six, with an academic credential that allows them to pay off the money they borrowed to attend - an obligation that Wayne State fails to meet two times out of three in general, and nine times out of ten for black students. I'd argue that its failure to enable students to pay off that debt means they should do something dramatic or admit a lot less students.
posted by l33tpolicywonk at 8:58 AM on August 12, 2010

I quickly glanced at the African American data and it would appear to me that there is a lot of regional variation in African America graduation rates. The best performing schools appear to be in the southern states, while the worst performing schools are overwhelming in the North (particularly Michigan, where I live). I wonder what social factors are at play in these regions to create such disparity. I would surmise, that at least in Michigan there is still a huge amount of racism among the both faculty and the students. For example, my understanding about Wayne State is that the student body consists of affluent white students from the suburbs, and black students from Detroit and the immediate area. Many of the white students learn very early on to avoid Detroit, and often commute back home at night. This prevents students from both races from mingling and working together, and probably isolates the African American students. Older faculty from the area probably grew up with same mindset, and remember the race riots and white flight of the previous decades. One example of this continued racism was Grutter vs. Bollinger where a law school applicant took the University of Michigan to the Supreme Court over race being used as a factor in University admissions.

I grew up in Virginia and it seems to me that the class and race divides, and institutional racism present in the North are just not as strong. Furthermore, I think a lot suburbs in Virginia are not as self segregated as they are in the North. Because of this students and faculty of both races have a lot more experience working with and understanding one another. This experience and understanding probably translates into the college setting and reduces the isolation experienced by African American students.
posted by kscottz at 9:05 AM on August 12, 2010 [1 favorite]

One of my nieces is 19% less Hispanic than her half-sister, give or take.
posted by ethnomethodologist at 9:10 AM on August 12, 2010

Having read the above-the-fold part I initially was going to assign blame to society in general rather than any specific offending school. Thinking more (though perhaps not well) I'm not sure that a school should have the mission or obligation to ensure that their students be able to pay back the loans (perhaps likely but not necessarily) incurred in order to attend.
In 1983 I and a bunch of other people were being paid $5.85 to stack 110 lb boxes of salmon in a warren of freezers in Petersburg AK. One of my fellow stackers was a new met friend and had a Master's in Literature from a major Midwestern University. I asked him what the point of his degree was, he responded "It makes my mind a better place to be in my spare time."
So if any of those schools are deceiving their applicants into thinking that they will certainly be able to repay any debt incurred in the process of going to school then, yeah, fuck them. But getting a degree is only one benefit of attending school.

Signed,
28 year (so far) AS student.
posted by vapidave at 10:15 AM on August 12, 2010

There doesn't seem to be any attempt here to control for factors other than race, is there? For instance, if Wayne State's white students are from the suburbs and their black students are from the city, that might cause a large wealth gap. And the black students at Wayne State (or the hispanic students at Brooklyn college) might be poorer in comparison to their white classmates than the black students at Old Dominion (or the hispanic students at John Jay). Other variables like class (besides race) might affect the results comparing which college is "more succesful". This would potentially lead to erroneous conclusions about which programs to boost graduation rates are more succesful.
posted by Jahaza at 10:22 AM on August 12, 2010 [1 favorite]

There doesn't seem to be any attempt here to control for factors other than race

There doesn't seem to be any attempt here to control for fucking anything, because these are press releases, not real research. Any random one of scunning's comments in this thread so far contains an order of magnitude more intelligent reflection on educational research methodology than these two methodologically half-assed and politically ill-considered "reports," whose chief purpose seems to be naming and shaming individual schools in order to grab some quick media attention. "The Education Trust," whatever that is, ought to be ashamed of itself; this kind of cheap, ill-founded criticism can only serve to hurt the public image of higher education, and with methods this flawed there's just no compensating benefit to be found.
posted by RogerB at 10:46 AM on August 12, 2010 [1 favorite]

@Jahaza - These are unconditional averages by school and race, and nothing more. If you believe that every Black person is identical to every other Black person, and every White person is identical to every other White person, in every single imaginable dimension, then unconditional averages will be sufficient to let you know that Wayne State is not doing its job.

But if you think that there are differences -- both observed and unobserved -- and that those differences are correlated with the student's educational choice, then you have no idea what any of these correlations mean. For all we know, there are more disadvantaged Blacks living in closer proximity to Wayne State, and thus the higher drop-out rate is driven by heterogeneity in selection into the school. But that just means Wayne State services a more disadvantaged population, not that Wayne State is dysfunctional. In fact, if it is heterogeneity and selection, then Wayne State may in fact be doing a great job which is masked by the fact that the students they service are already disadvantaged prior to admission. (Think if disadvantage as possibly that the public school system is worse for those Black students who attend Wayne State as a concrete example). What we want to know is what would the student outcomes have been had the person gone to the next best alternative than Wayne State - which is something that studies like this cannot answer.

Studies like this are valuable for two reasons only, neither of which make them useful for policy. They are easy for anyone to understand and they give us someone to blame for poverty. They fit easily into a newspaper headline, and the implied direction of causality makes sense to everyday people who have never had to think about the difficulty in estimating the causal effect of a school on an outcome.

It's entirely possible that UNC-Chapel Hill is actually at the bottom of some list like this, but that that is not visible because the quality of the Black and White students who enter UNC as freshman is so high. It's obviously possible, too, that Wayne State really is the worst one, but you can't know that from this "study".
posted by scunning at 11:22 AM on August 12, 2010

The best performing schools appear to be in the southern states, while the worst performing schools are overwhelming in the North (particularly Michigan, where I live).

Aren't most HBUs in the South? Maybe that influences the data.
posted by atrazine at 11:26 AM on August 12, 2010

atrazine: "Aren't most HBUs in the South? Maybe that influences the data."

For purposes of comparison, HBCUs were explicitly excluded, though the story there is pretty mixed too, and is closely connected with an institution's financial viability.
posted by l33tpolicywonk at 11:34 AM on August 12, 2010

I know what schools can do to fix this. They could make it harder for minorities to get in. They could admit fewer minorities.

I'm not saying that there aren't other factors at play but I think differences in admission standards between the races probably have a lot to do with differences in graduation rates between the races. This doesn't mean that I think this would be a good way to correct the problem in the ways that are described or that there is something wrong with colleges using differing criterea for underrepresented minorities. I'm just saying that it seems like an obvious reason for the gap. If I got into a school where most students were better prepared than I was my chances of gradutating would be slimmer.
posted by I Foody at 1:26 PM on August 12, 2010

It's a bit out-dated, and its sample was necessarily tiny considering the method (just two schools, including one HBCU, IIRC), but even taken just as an anecdote, the ethnography _Educated in Romance_ may offer some insight into this.

The study asked why women who enter college as math/science majors switch to huma/soc/educ majors at a rate much higher than men. The answer turned out to vary according to the students thoughts about what they were doing in college in the first place, which had a very noticeable correlation with ethnicity.

A tiny number of women, both white and non-white, articulated a rationale for college that can be summed up as "acquiring skills from experts." That group had the low rate of major switching, as well as a high graduation rate. Apparently, that's the "right" attitude to have for college success, because it aims at high grades while accepting low grades as a simple indication that a particular skill is hard to acquire and needs more practice.

But most white women articulated a rationale that can be summed up as "I was good at math/science in high school, so I naturally majored in math in college." That group had a high rate of major switching, explained as experiencing serious difficulty in science/math classes as a blow to their self-esteem, but also a high graduation rate, because they still found ego rewards in their college social lives (the eponymous romance) and because, let's face it, huma/soc course work is easier.

And finally, most African-American women articulated a rationale that can be summed up as "I need a college diploma to do what I want after graduation." That group had a low rate of major switching, tolerating low grades very well because grades didn't matter as much as the diploma itself, but also a low rate of graduation, because the effectively binary rationale (diploma or no diploma) was a terrible way to approach the problem of success in college.

I don't know how well this study has held up over time or even how accurately I've summarized it--it's been over 10 years since I read it--but it's the first thing I thought of here.
posted by Monsieur Caution at 4:29 PM on August 12, 2010 [1 favorite]

...and because, let's face it, huma/soc course work is easier.

Can we get through a single goddamn week without this?
posted by griphus at 4:35 PM on August 12, 2010

Can we get through a single goddamn week without this?

Is it really that common here? I have a BA and Ph. D. in anthropology, but also around 100 semester hours of science courses completed over 17 years of continuous university enrollment, and my vote is soc/huma courses are *much* easier. :D
posted by Monsieur Caution at 4:52 PM on August 12, 2010

Speaking of sampling problems, did you ever consider that your personal "vote" doesn't count for that much in a cross-field comparison? People who go on to the Ph.D. in a given subject/field/area usually find it a lot "easier" to get good grades and learn the material there. And yes, the cheap "admit it, science is harder" thing is both common around here and really dumb.

As to Educated in Romance, admittedly I've just now skimmed a couple of chapters of the book, but I don't see anywhere that its discussion of the student culture among a few women at one particular HBCU is claimed to be reflective of all African American student culture at all other institutions (or, indeed, any) — that seems like a pretty enormous unwarranted leap to me. How is this supposed to be a better explanation than "white people go to college like this, black people go to college like this amirite?" In any case, it seems like a shame to start constructing elaborate conjectures about black students supposedly having different cultural expectations in a discussion that's based on wholly illicit conclusions drawn from some think-tank's fake research.
posted by RogerB at 5:25 PM on August 12, 2010

Speaking of sampling problems, did you ever consider that your personal "vote" doesn't count for that much in a cross-field comparison? People who go on to the Ph.D. in a given subject/field/area usually find it a lot "easier" to get good grades and learn the material there.

Of course I thought of that or rather have thought of it while spending more years than most observing the issue directly and, you know, reading research like what's cited above, which asks not why fewer women major in science but why they *switch* and get better grades in soc/huma in spite of having believed they were better at math. Does that come up every week here?

To the extent that you're not enamored of ethnographic method, which almost always involves a tiny sample size, or the comparison of incommensurable studies to provoke new ways of looking at a problem, I guess that too may count as evidence that cultural anthropology, at least, does not really have a very challenging standard for making an argument, because I assure you that's the norm even among professionals and, accordingly, isn't quite as hard for anyone reasonably literate to succeed at.

How is this supposed to be a better explanation than "white people go to college like this, black people go to college like this amirite?"

Well, I guess it does have one thing in its favor: it's based on actually getting out and asking the students what the problem is. Which is certainly harder than googling it. :D
posted by Monsieur Caution at 5:56 PM on August 12, 2010

From what I know of Wayne State, especially their composition program, I would suggest that the school shows up on this list because it is racist.

Well, okay, yes, that's a deliberately incendiary way to say it. The less incendiary way to say it, judging from the department I'm most familiar with, is that they are in love with the results of standardized tests, specifically with how ACT/SAT scores correlate with college-level performance. The composition program, for example, aggressively tracks students with low ACT/SAT scores into remedial classes that don't count for college credit; this increases the time to graduation, with only very questionable benefit to the student. Furthermore, they are one of the few programs I know of that privileges grammar and spelling in the teaching of composition; most current research in the field admits that teaching grammar and spelling in the context of a college-level or near-college-level course for college-age students is, for a variety of reasons, both pointless and futile, and instead focuses on teaching how to use writing as a tool to develop complex ideas. Wayne State serves, for the most part, two demographics: relatively poorer black kids from the city who managed to fight their way through the broken public school system, and relatively wealthier white kids from the suburbs who couldn't get into UM or MSU¹. The remedial classes grade, for the most part, on the ability to use the home dialect of the latter group of students; the former group are, predictably, fucked.

This assessment is based on both my observations of the established practices of the composition program and on talking to neighbors in Detroit who describe Wayne State as racist — largely due to their own observations of the school's tendency across disciplines to track black students into years of remedial classes that don't count as credit for the students, but do make the school money.

If Wayne State reflected the interests of its city, it would be a tremendous asset, not just for Detroit but also for Michigan and for the United States as a whole. As is, it is most charitably described as a missed opportunity.

[1]: Pardon, there's a third demographic: they also heavily recruit foreign students. The funny thing about their overseas recruitment materials is that they somehow never mention what city the school is in....
posted by You Can't Tip a Buick at 9:58 PM on August 12, 2010

I attended Wayne State for graduate school. It is a very odd place. It isn't a HBCU but nearly all undergrads are black. Nearly all the professors and any staff who make enough to actually live on are white. Yet it is forbidden to notice or to comment on race.

My program was nearly 100% white and suburban, and we had a very high graduation rate.

The administration wanted to become a first tier research university, and would ignore anything that didn't get them towards that goal.

Or, in other words, a lot of what You Can't Tip a Buick was saying.
posted by QIbHom at 8:38 AM on August 13, 2010 [1 favorite]