Yeah the thing I like about D&K is that it’s a quasi twin experiment. I believe especially nowadays that the “quality in → quality out” phenomenon is vivid at the top schools. It’s much tougher to suss out if the institutions themselves were a transformational influence in these students’ trajectories.
Again I’ll try to read it, thanks for the link.
This is incorrect. D&K said this and were proven wrong, as aligns with common sense.
Hm I’d love references. I hadn’t heard D&K as being debunked.
Occasionally vetting common sense is good practice to avoid inbreeding ideas (and stale ideas) so I’m glad work is still being done here
ETA: oh and just to clarify: by far right I meant very very high income. That is, Ivy+ grads appear to have better access to wall st/VC/consulting careers. In other realms and income levels, not as much. I thought D&Ks latest paper added a result beyond their original thesis that Ivys helped those from the lower SES more than their state school did. Could that last point be what was refuted?
With respect to their original conclusion, I think it has been repeatedly albeit anecdotally shown that an Einstein level talent can flourish at MIT as well as Rutgers
Is anyone able to post a gift or archived link? I am curious about it too.
Reductio ad Hitlerum
Thanks for the link!
I was hoping for more insight than just ROI of different colleges / majors. Oh well!
The schools may believe otherwise, but I’m going to go out on a limb here and guess that apps will drop at the test-required schools. Which means that, enrollment targets being fairly inelastic, admit rate must increase. (I hope I am wrong, but I suspect Harvard and test-required ilk will see fewer 1300-SAT and 29-ACT kids applying now)
So if you’re Stanford and Princeton, Penn, Chicago, Columbia, etc. – maybe they don’t really care, but they must be aware that they stand to perhaps receive even more apps if they remain TO. Which means their admit rates will continue to drop.
So we’ll be able to guess institutional priority by how these schools react: making the adcoms’ jobs easier by giving them another data point – SAT/ACT – to confirm academic achievement and readiness, or making apps and admit rate the priority.
I think you are correct. Admit rates will likely creep up at test required schools. Wonder if USNews will move the goalposts accordingly in their ranking methodology to keep pace.
I do not think admit rate is a part of the US News ranking. But certainly correct me if I’m wrong.
USNWR doesn’t consider admit rate in their methodology. They do consider SAT scores at colleges for which at least 50% of students submit scores.
In the past USNWR has changed their methodology as needed to insure the colleges readers expect to see at the top at the top (HYPSM) remain at the top, and I expect they will continue to do so. Examples including things like in the past adding a logarithmic adjuster on spending per student to address "statistical outliers” (Caltech). This change occurred at the same time that Caltech would have been ranked #1 without the newly added adjuster.
Currently they don’t consider admit rate, right? So – probably not. Though they might now decrease the importance/weight of the Pell factor.
This is incorrect. Chetty’s recent student went out of their way to say their results aligned with Dale & Krueger, who only reviewed mean income. An example quote is below:
However, once again, we find very small impacts of attending an Ivy-Plus on average
earnings, consistent with the findings of Dale and Krueger (2002)
…
In sum, our findings on mean earnings impacts are fully consistent with prior work
The point of Chetty’s study was to show that the increased chance of admission at Ivy+ colleges for top 1%/0.1% yields an objective advantage on future outcomes, so rather than focus on mean earnings, like Dale and Krueger, instead they focus on the 3 measures below, which did show more significant advantages upon attending Ivy+ colleges.
The reasons why attending an Ivy+ may increase chance of these 3 specific outcomes are not reviewed in much detail, leaving much open to speculation. For example, the authors do not review how much of #1 relates to being a circular definition vs preference for having attended Ivy+. Or how much of #2 relates to a home school bias. Or how much of #3 relates to increased chance of switching from original post-grad career plan to “elite” finance/consulting.
Yes, to me that’s a really weird measure.
The first two categories are indeed strange and circular to me as well. The third was interesting and led me to business forums, chats with Ivy alum friends in finance, and reading discussions here. The consensus in those venues seems to point to more tribal hiring wrt schools for prestige firms in certain careers such as PE/IB/ and so on. If we consider that valid and try to reconcile with D&K, the distribution that makes sense is one that is similar/overlaps between Ivy+ and non-Ivy for most of the range but with a uptick at the far right for the Ivy grads
Such ROI analyses are near useless since they don’t consider the individual student. They assume your price will be either the sticker price or the average net cost after FA, depending on the specific analyses. They don’t consider your actual cost, which may be very different. They assume your income will be the average income of students from the college, regardless of your major or post-grad career plans. Perhaps most importantly, they inherently assume you are the average student at the school, and school name determines income, with no influence from individual characteristics.
For example, if a student who is accepted to an Ivy+ decides to take a full ride scholarship with the honors program at his/her flagship, then the analysis assumes he/she suddenly becomes the average student at the flagship he attends, who likely has a very different academic and background profile than the average student at any Ivy+ college.
This differs from the previously referenced Chetty + Dale & Krueger studies, which instead compare outcomes between students who applied to Ivy+ colleges and either attended or did not. D&K found that applying to an Ivy+ was associated with increased average income, even if the student did not attend. It made relatively little difference (in average income) whether the student attended or not. They key was whether they applied. I expect this occurs because students who apply to Ivy+ colleges are also more likely to apply to and pursue certain types of career paths that are also associated with increased income, Simply looking at average income by college does not capture this type of influence.
CollegeAdmissions_Paper.pdf (opportunityinsights.org)
From right before the snippet you posted (which was about one particular item):
"These findings differ from a well-known set of studies which conclude that attending a highly selective
college in the U.S. has little impact on students’ earnings (Dale et al. 2002, Dale et al. 2014, Mountjoy et al.
2021, Ge et al. 2022). To investigate why our conclusions differ, we replicate the research design used in
those studies by comparing earnings outcomes for students who attend different colleges, controlling for the
set of colleges to which they were admitted."
They also questioned the methodology of D&K to explain the differences between D&K and the Chetty study:
6A further methodological difference is that Dale and Krueger proxy for college quality by the average SAT scores of
admitted students rather than estimating college fixed effects directly. Within the set of highly selective colleges Dale and
Krueger consider, average SAT scores turn out to be weakly associated with post-college earnings (Chetty et al. 2017). Hence,
it is not that these colleges have no impact on earnings, but rather that mean SAT scores are not highly predictive of value-added
within a sample of highly selective colleges.
From right before the snippet you posted, it states that they are referring to the 3 measures I mentioned in my post above above differing, rather than mean income differing. The 3 criteria they emphasized differ, such as chance of working at a firm that employs a lot of Ivy+ grads. D&K reviewed mean income, rather than these 3 measures, so D&K didn’t report advantages to attending Ivy+. Both D&K and Chetty agreed that attending Ivy+ was associated with little change in mean income.
They also questioned the methodology of D&K to explain the differences between D&K and the Chetty study:
6A further methodological difference is that Dale and Krueger proxy for college quality by the average SAT scores of
From right before the snipped you posted, Chetty states the differences in earnings were “statistically insignificant” and reconcile with Dale and Krueger.
“We also find small, statistically insignificant effects of Ivy-Plus attendance on log
earnings (Appendix Table 7), reconciling our findings with those of Dale et al. (2014)”
They say, “Being admitted to an Ivy-Plus college has a small impact on mean income rank.” “The small impacts on mean ranks arise because Ivy-Plus attendance primarily shifts outcomes within the upper tail.”
That matters. If you end up being a design engineer, a teacher, a nurse, a doctor or a professor, you won’t outperform state flagship graduates by any statistically significant amount. However, if you want to purse one of the few career paths that can get you into that fraction of 1% that outearn the rest of the country by a lot, it matters very much that you attend an Ivy-Plus school. This is America. The “upper tail” isn’t fantasyland. It is possible…if you attend the right school.
They didn’t find the upper tail was fantasyland, if you didn’t attend an Ivy+. Instead they found both groups had a small minority in $650k income, but the chance of earning $650k/year at age 33 increased when attending Ivy+. There was not a significant different during the first ~8 years after graduating, but by age 33 the difference became significant.
The reasons for this difference in chance of earning $650k at age 33 were not investigated. As I touched on in my earlier post, one possibility is the connection with “elite” finance. Attending certain Ivy+ colleges may increase chance that the student shifts career plans towards “elite” finance.