However, as Krueger highlights in the NYT article, for underprivileged students attending a prestigious university does have a significant impact on earnings (even after controlling for ability). So the very students this NPR article is talking about are those who would most benefit from attending a prestigious university.

> So one answer to "why many smart low-income students don't apply to elite schools" may be that these students are really quite smart, and do actually know better what's right for them.

Nope. They would almost invariably do better by going to an elite university. Doing so will provide a significant earnings boost, not to mention that it's usually free. Intelligent low income students will usually get the best aid package from an elite university (usually a grant covering tuition, room, and board), thus making that choice cost effective even in the short run.

We need to encourage more low income students to realize that elite universities are a great option for them. On the flip side, middle income students should wake up to the reality that attending a second rate private university is a terrible economic decision.

For the attended-in-1976 minority cohort, they found a 6.7% higher income in 2007 for those choosing a more selective – but falling to 1.6% (said by the authors to be "indistinguishable from zero") if the selectivity-outlier historically-black colleges were excluded. For the entered-in-1989 minority cohort, the income boost in 2007 was higher – 12-14%, and survived the exclusion of historically-black colleges. (One way to interpret that might be: the income boost becomes smaller the longer you're out of college.)

Perhaps most interesting: if a student's parents had an average of 12-years education (~high school), selective-attendance gave a 5.2% income boost. But if parents had an average of 16-years (~college), selective attendance gave no income benefits.

But none of these boosts among less-privileged subgroups are very large. The paper also mentions some reasons its estimates of the benefits of selectivity (even where negligible) could be too high, including:

* if selective schools are more generous with financial aid – you and the NPR article both suggest this is the case * if students with high unobserved earning potential are more likely to attend selective schools – seems likely to me, when a student is confident they have the skills, ambitions, or personal contacts to earn a lot post-graduation

So sure, include the Dale/Krueger subgroup estimates in the pitch that selective colleges make. But many low-income students might hear, "you'll get roughly a 5-14% income boost"… and still pass, picking less-selective colleges that match their other academic/community/career/family/friend priorities.

I get why the elite universities want more low-income enrollees: it's a key part of their self-image of outreach & uplift, helps justify their immense costs to others, and immunizes them against other criticisms. But for the low-income students themselves, even given a more-than-zero benefit shown by Dale/Krueger, it's not a slam-dunk decision. When such students choose elsewhere, they may be expressing their own smarts and self-knowledge.