It is still qualitatively significant, as why would anyone work hard to get into a highly selective college is a most selective will do? Why not just parhtay all the time in high school? Other researchers should chase this one.
Thank you to @Data10 for the link and to the authors for their detailed presentation. To me the results are:
- With all the factors considered (table 3, model 4), college selectivity makes no difference, but major is important, female gets paid less and so does a poor kid. But, this is huge: all the factors explain only about 18% of pay difference
- A poor kid with a degree from a low selectivity college gets paid less but not so from a high selectivity college. However, family income contributes only about 1% to pay difference (table 4, model 4).
So, what not studied in the paper and not focused on by students and parents here has 80+% influence on 10 year post graduation pay.
Oops it’s @NeoDymium who linked the manuscript in post #752 with the above good question for the authors. Though unemployment cannot be included in the model to analyze salary, but being 30% it could be a separate study to show who weren’t working 10 years out. Nevertheless the current study stands with the focus on who were employed. Thanks NeoDymium for sharing.
I’m rather careful to say not-employed rather than unemployed - for all I know they may have just been between jobs, wealthy enough to retire, still in school, doing unpaid charity work, or not willing to follow up on the survey. My issue is that it’s an obvious giant hole that their research didn’t address, among many others, and some of these holes in their methodology are so gaping and noticeable that I feel it prudent to simply call it bad science instead of trying to see what their (obviously non-representative) numbers seem to suggest. Incidentally that’s not to say that the effect of Ivy/elite schools is not important, just that they have done nothing to prove it to be so and the “Ivy have an advantage even after controlling for everything” shtick that some in here have taken as gospel is… based on sloppy science.
The linked paper in #752 does NOT show college selectivity has any effect. However, I wouldn’t be surprised in a refined study where school selectivity is well-defined with fewer college majors included the authors find Ivy/elite schools are better as claimed in the OP. What sticks out the most in the study is that how much one makes 10 years after graduation has 80+% to do with something not examined.
I’d like to see a study like this for college admissions that may contribute 80+% influences to things other than test scores, GPA, EC etc.
If you look at the actual stats, you would find that they have some selectivity terms that are significant and some SES and selectivity interaction terms that are significant. But in words, they more or less make the claims as reported in the article here.
Again, the task of predicting a person’s future from a few stats is not so easy. A lot of people want some “scientific proof” that elite schools were worth attending but the factors that tell you whether or not they were are highly personalized and I have my doubts that these statistics could ever properly address them. The original statement - “even after controlling for everything we usually consider, Ivy schools provide a premium” - simply is not demonstrated here.
If we’re going to bring up Krueger, we might as well bring up his second study, which controlled for which schools the student applied to instead of was admitted to. Based on the second study, students who applied to elite colleges had higher than average earnings, regardless of whether or not they were admitted.
The 2002 Krueger study did find a statistically significant difference in earnings, at least for their exact matched applicants model. Those who went to the less prestigious school earned significantly more(statistically) than those who went to the more prestigious school.
When you actually dig into the 2002 study, the rankings by school SAT don’t always fall out the way most people would expect. Yale is worse than Bryn Mawr. Swarthmore beats out Stanford, Princeton, and Columbia. Below the absolute top tier(HYPMS), the rankings become very subjective, at least within some very broadly defined selectivity group. Many of the small non-Ivy liberal arts colleges in the study(ex. Oberlin) may be great schools with high SAT scores, but probably lack national name recognition that a lower ranked state flagship (ex. U. Michigan) might have.
- Harvard and MIT weren't part of the Krueger study and Princeton and Stanford were hardly the highest ranked schools in the study. Your implied ranking of schools is different than the study's ranking of schools.
- Even some of the lower ranked private schools like Emory and Denison get the vast majority of their students from out of state, so the argument that students who only apply to lower ranked schools all live in the immediate area of those schools doesn't seem very plausible. In some cases ex. Penn State vs UPenn, you have high ranked and low ranked schools in the same state, so geography seems a less than satisfying explanation even for the in state residents.
I almost hate to wade into this because I don’t have an opinion on this issue but I just came across something in this discussion that bothered me.
A poster suggested early on in this discussion that the SAT stats for entering students for Harvard and University of Michigan were virtually indistinguishable. That didn’t sound quite right to me so I looked at it more carefully. I searched the common data sets for both schools and looked at the 2014-2015 cohort in order to avoid any inconsistencies between the old and new SAT scoring. (That being said, the data for 2015-2016 is not appreciably different.) Although the 25th and 75th percentiles showed differences between the two schools (M-620-720 CR 660-760 math and H 700-800 CR 710-800 math), what was really telling was the distribution. Where only 33% of incoming Michigan freshmen scored between 700-800 on CR and 60% did so on math, 78.47% of incoming Harvard freshmen scored at that rate on CR and 81% did so on math. That is a different group of students and no matter what conclusions we draw, I think it’s important to be honest that the cohort at one school is different from the cohort at the other.
I also want to be clear that I think it’s more than possible to get a wonderful education at a non-“elite” school, that there are exceptional students to be found at all schools, and that the non-elite school may well be a better experience and a better path into the future. From where I sit, it looks to me that success comes down to the drive, talent, personality, commitment, and willingness to work hard and the name of the school is much less important.
Michigan’s frosh class is about four times the size of Harvard’s. The top quarter of the class at Michigan, based on the above SAT scores, appears more than competitive in SAT scores than Harvard’s class. While certainly SAT scores are nowhere near a complete assessment of a student’s academic ability (despite the often over-reliance on such by many posters here), it appears that Michigan includes a cohort that is similar in size to Harvard that is SAT score competitive to Harvard, even though it is mixed into many other students with somewhat lower SAT scores.
So a student with Harvard level SAT scores should be able to find as many or more SAT score peers at Michigan as at Harvard.
Does a university want its graduates become members of Congress, or the wealthiest member of Congress, self made? As of today that person is Darrell Issa, a graduate of a directional directional university I’ve never heard of, definitely nowhere near Harvard.
I have seem to have heard that line of argument before.
771 (and 773) yes, but they are diluted. For every 4 people you meet at M, only 1 will be your peer, while at H virtually every person you meet will be your peer. If I had H-level scores, H would seem more exciting, to me at least.
Yes, but in UofM you will not lose the ability to talk to your plumber
I know plenty of U of M grads who can’t talk to their plumber, and plenty of Harvard grads who can (and one Harvard grad who is in fact a plumber.)
This is a silly tangent.
Thank you @colfac92. Never did I suggest that there weren’t talented peers at Michigan or for that matter most other schools. I was only trying to correct an earlier statement that the cohort was the same or close to the same at both schools. Regardless of the actual numbers of high SAT scorers admitted to Michigan, if you want to get an accurate sense of the make up of the two schools, it’s more important to look at the proportions and the ** total ** cohort at each place.
Can I have the number of the Harvard plumber?
Because we can’t get a decent plumber no matter what we try. 
The Harvard plumber can talk Kant, Hegel, and septic tanks with equal depth and enthusiasm. Best $500 I ever spent on a service call!