Paul Tough writes about how college in the US has gone from being a tool for upward mobility to a shield against downward mobility, and at the same time, colleges are employing survival strategies that further divide the diploma-haves and diploma-have-nots. Essentially, college is becoming more like a caste system, where the promise of "breaking out" of your caste via a degree from a selective college is less and less likely. He outlines a few patterns that explain why this is happening and the effect it has on students.
(Note: for brevity I use "smart rich kids", "smart poor kids", etc. as abbreviations for a more nuanced concept. They're not meant to be flip, just to keep things terse.)
Students who attend highly selective colleges are much more likely to be wealthy as adults, regardless of their initial socioeconomic standing. That is, the experience of going to a selective college can erase many socioeconomic disadvantages, although it is more "life-changing" (in terms of likelihood of climbing the socioeconomic ladder) for poor students. However, capable-but-poor students often aren't going to selective colleges: elite schools are largely (over 2/3) populated by rich students, i.e. by those who will benefit the least from the experience. Why is this and what problems does it cause for poor students?
The vicious cycle of standardized tests and admissions. Caroline Hoxby found that at selective colleges, the median SAT score correlates with the amount those colleges spend per-student to educate them. Elite colleges spend so much per student that they lose money on every student as a result. But attending a selective college also improves your economic prospects, so the colleges are counting on those students to be alumni donors later. At the same time, while the official party line given by high school guidance counselors was that any good college was fine and you should find the "best fit", students were increasingly realizing that selective=$$$, so today's students tend to go to the most selective college that will admit them. To do so, they need high test scores, costly test prep, etc. This further concentrates all the smart rich kids in a shrinking number of elite institutions that must then spend lots of money on them to make sure they will donate later—a vicious cycle.
But Hoxby also found (based on studying ACT/SAT scores) that poor smart kids weren't even applying to selective schools that they were qualified for, and she hypothesized that they just didn't know the ropes about discovering and applying to selective schools (including things like knowing they could get fee waivers if they couldn't afford to apply), or perhaps weren't confident of their success, because they didn't have local role models to give them that information. Hoxby and Sarah Turner devised an intervention: they randomly chose 10,000 poor smart kids (based on SAT/ACT info) and sent each a folder with info about applying to selective schools, sample fee waiver forms, information about financial aid, and more. Students who received these packets were 31% more likely to apply to selective schools and 19% more likely to decide to attend, compared to a control group of similarly poor and smart kids who didn't receive a packet.
What about resources made available free to everyone to improve test scores? Khan Academy has a free "official SAT test prep" program anyone can use, but data scientists from Khan Academy and the College Board found that even though each hour of time spent on the site had an equal effect on score-boosting for all students, privileged (primarily white and Asian) students spent two to three times as long using the site as underprivileged students. UPenn social scientists Angela Duckworth and Katherine Milkman had intended to try various interventions to motivate the latter group to study more and develop better habits, but were unable to get the needed data and collaboration in time to do so; ultimately the study didn't get done.
Fear of failure/fear of fitting in. When smart poor kids do get into an elite college, they often have no peer group because the kids around them have been differently socialized. In particular, the rich kids likely came from professional households where the value of networking (with both peers and superiors, i.e. faculty) is well understood; the poor kids often didn't, and some see fraternizing with faculty or even forming peer study groups as an admission of weakness, and that they should instead just work hard on their own and keep their heads down. This causes 2 problems. First, students learn much more and have better psychological support if they study in groups vs. alone. Second, these students don't develop the networking skills they will need in post-college careers. Even when these students are explicitly invited to join (eg) advising groups, they often don't respond. The situation is exacerbated if the students are from groups historically under-represented at that college or in that major.
Ethnic diversity ≠ socioeconomic diversity. Colleges like to boast that they accept a lot of Pell students, but many game the system by finding the students closest to the Pell threshold. Thus, a student who is qualified and just below the Pell threshold is ten times more likely to be admitted than one who is equally qualified but just above the Pell threshold. The result is that Pell numbers don't correlate with socioeconomic diversity: instead you get a bimodal distribution, with enough "barely Pell" kids to boost those numbers and a lot of rich kids to subsidize them. (Extreme examples: in 2013, only 2% of Princeton students came from the bottom 20% socioeconomically, whereas 17% of them came from top-1% families. In the same year, the median family income of entering students at Trinity College that year was $258,000. In 2018, the median family income at the University of Alabama was higher than that at Bryn Mawr.) Using a similar trick, colleges that admit students of color who immigrated voluntarily from Africa or the Caribbean (who are more likely to have attended private school, come from an intact 2-parent household, and have at least one parent who is a college graduate, compared to students of color descended from slaves) can hit an "ethnic diversity" target without improving socioeconomic diversity.
Lauren Rivera, author of Pedigree, found that "elite" firms when interviewing college grads are effectively screening for things that correlate with parents' socioeconomic standing—that is, for who the student was when they applied to college, not who they were at graduation. The rationale is that being accepted to a selective university is a signal that you're smart and can learn fast and work hard. How well you did in college is less relevant, especially for lucrative but entry-level jobs that don't require deep specialized skills but do require networking and schmoozing. Students from elite schools who did extracurriculars (especially sports, and especially sports with a high barrier to entry, like lacrosse) were judged particularly desirable by this criterion. As a result, for a student who comes to an elite university not from that background, their story/struggle (sometimes the very thing that brought them to the attention of an admissions officer) is not something they can talk about, so they feel even more isolated.
Financial aid. To pay the bills, financial aid is often given to rich kids who don't need it in order to get them to come, because even at a discount, they will pay more of the bill than a poor student, and thus help balance the books. The most desirable students from this perspective are below-average academically but high-income: not so far below average that they drag down US News rankings, but enough below average that your offer is probably the most selective they'll get, so they'll be motivated to come. As a result, the richest colleges, which can best afford to take lots of poor smart kids, take the fewest. As one admissions officer said: “Admissions for us is not a matter of turning down students we’d like to admit. It’s a matter of admitting students we’d like to turn down.” His advice for improving the situation? "quit paying so much attention to the SAT & ACT", especially since test-prep and other $$$ strategies can "inflate" SAT scores to where they are out of whack with a student's actual ability (as demonstrated by their transcript, e.g.). When DePaul University (a less selective school) stopped weighting the SAT so heavily but asked students to voluntarily self-report their score after admission, they found that students who didn't report their scores on the application did just as well as a group as those who did report their scores. That is, low test scores are often a "false alarm". Family income predicts SAT scores; it doesn't predict high school performance.
A bright spot is the UTexas system's "top 10%" rule, in which the top 10% of students from any high school are guaranteed UT admission. Yes, some arrive less prepared for socioeconomic reasons, but support systems are put in place to help them succeed in the same courses as everyone else, rather than sending them to "remedial" classes. Some of the support is psychological: when these students have difficulty, remind them that it's because the material is hard, not because they are "less fit" to be here. (Students from less-selective high schools clearly see that their peers from more-selective high schools are ahead when they get to college, but they take this as a signal that they don't belong there.) Students need to hear that "someone here is looking out for me, someone wants me to succeed," and sometimes the support system has to be "focused and intrusive" to do this effectively.
Upward mobility or a shield against downward mobility? Having a Bachelors still commands a premium, but most of the increase in the gap between degree and nondegree students comes from those who continued to advanced degrees. The relative advantage of a Bachelors-only has not changed since the 1970s. Corollary: we need to invest more in other non-bachelors-degree pathways for those not college-inclined who want to pursue economically secure careers that don't necessarily benefit from a 4-year degree. And college is a terrible investment if you don't complete your degree, as you end up with debt but no premium. For-profit colleges are even worse, and spend twice as much on marketing and recruitment per student as on actual teaching. Yet they exist because "traditional" higher ed is structured to exclude so many students, so for-profit education soaks them up.