At what point does undergrad reputation hurt you?

[pookie bear]

That is pretty interesting abababba. All of those undergraduate institutions have really good reputations though.

I did a quick search of every top 10 school to see where their students came from. Some of the schools only had job market canidates posted on their website and of the websites that had graduate student websites, only about half of the students had a CV posted. I found 2 students from Virginia and 2 from Case Western. Every other student was either from abroad or from a top university/LAC. Not very scientific or a relevant sample size, but interesting none the less.

Michigan, Wisconsin, UCLA, etc seems to be where students not from ivies, public ivies, and top LAC's start occuring in signifigant numbers.

I disagree jeeves. Suppose you take the pool of total applicants to Harvard and divide them into two sets. Set A is those students from top ivies, public ivies, and top LACs. Set B is everyone not from ivies, public ivies, and LACs. This is pure conjecture, but I would imagine that set B is the larger of the two, I have no idea though.

[pookie bear]

double post

[Golden Rule]

Originally Posted by abababba

I'm at Harvard. My class has people from the following schools (the domestic students):

Yale, Stanford, Berkeley, Harvard, UCLA, Williams, Dartmouth, Columbia

So reputation matters, but there are a few schools in there with great undergrad reputations but without graduate Economics.

Not surprising, but now the question I'm asking is -- what if a Harvard student and a student from a lesser known school ended up with similar quality recommendations (both in terms of the writers' praise and how well-connected the writer is with the schools) and else is equal? My conjecture is that they'd have similar performances in the application process -- that the "Harvard brand" alone wouldn't help all that much, and that concluding "undergrad reputation matters" is confusing causality.

Part of the reason why the top LACs place students well is that tenure positions at these schools are competitive, and these schools all have profs. that came from top programs who are well-connected.

[C152dude]

Originally Posted by YoungEconomist

However, I also believe the process is very random which explains a lot of variation as well.

[pookie bear]

I would say the majority of our LORs will come from graduates of these top programs. However, I doubt they hold much weight if the prof isn't also a well regarded researcher. A well respected researcher is likely to be well connected within the academic community irregardless of where they got their PhD. As such, a LOR from a well respected researcher is likely to be well received by ad coms. However, I think a LOR from some unknown graduate of a top program holds far less weight. I think people confuse a competent researcher with the brand name associated with a profs degree.

[jeeves0923]

Not only does the prof need to be a decent researcher, but having PhD students is also helpful. "jeeves is better than any PhD student I've supervised in 93 years.." might be something a top program would like to hear. It's all about reliable information.

[GymShorts]

Originally Posted by jeeves0923

"jeeves is better than any PhD student I've supervised in 93 years.."

One day I hope to be lucky enough to honestly say that I've been supervising PhD students for 93 years, but before then, I would have to have lived longer than any human in recorded history has thus far.

[YoungEconomist]

Originally Posted by Golden Rule

A lot of the "randomness" in the process is based on who knows who, and it seems random because we typically don't observe it.

I was referring to more actual randomness. I believe there are many times when admissions committees will probably have a significant portion of the pool that are more or less "equal" when all things are considered. In the end, if you have 20 "equal" people still being considered for admissions but only have 5 spots left, what do you do? This is just an opinion though, and you're definitely right that it might be more about unobservables that make it look more random than it actually is.

Note: When I say "equal" I just mean that they rate there chances of success similar although their profiles are different. For example, if one student has 800 QGRE with a 3.7 GPA, and another has 780 QGRE with 4.0 GPA, the adcoms may consider them fairly "equal." Think isoquants.

Originally Posted by pookie bear

Suppose you take the pool of total applicants to Harvard and divide them into two sets. Set A is those students from top ivies, public ivies, and top LACs. Set B is everyone not from ivies, public ivies, and LACs. This is pure conjecture, but I would imagine that set B is the larger of the two, I have no idea though.

For whatever it's worth, I was reading an op-ed by economist Thomas Sowell. He claimed that there are many rarely heard of LACs in the US, which actually send a higher proportion of their students to PhD programs than ivy league schools. His point was that the ivy league universities reputations are mainly because of faculty research, and not neccesarily student value added. By the way, when I say PhD programs I am not just refering to economics.

Originally Posted by GymShorts

One day I hope to be lucky enough to honestly say that I've been supervising PhD students for 93 years, but before then, I would have to have lived longer than any human in recorded history has thus far.

LOL. Unless of course you got your PhD at a very young age (like 7).

[jeeves0923]

Originally Posted by pookie bear

I disagree jeeves. Suppose you take the pool of total applicants to Harvard and divide them into two sets. Set A is those students from top ivies, public ivies, and top LACs. Set B is everyone not from ivies, public ivies, and LACs. This is pure conjecture, but I would imagine that set B is the larger of the two, I have no idea though.

While this might be correct, it doesn't change the fact that Harvard has only seen 3 applications from Virginia Tech in the past 50 years and it has probably seen thousands from Havard, thousands from Yale, thousands from Stanford. You can't equate schools in set B. There is an information deficiency. While your "pile B" might be larger, you can't just assume that an education at Virginia Tech is roughly equal to one at CU-Boulder unless you have evidence from past applications to suggest they are equal. With too small of a sample size from either of those schools, you can't make that significant comparison. Each University is its own set. For that matter, you can't really equate the schools in set A. However, the adcoms don't lack any information about potential grad students from set A.

[Golden Rule]

Originally Posted by pookie bear

I would say the majority of our LORs will come from graduates of these top programs. However, I doubt they hold much weight if the prof isn't also a well regarded researcher. A well respected researcher is likely to be well connected within the academic community irregardless of where they got their PhD. As such, a LOR from a well respected researcher is likely to be well received by ad coms. However, I think a LOR from some unknown graduate of a top program holds far less weight. I think people confuse a competent researcher with the brand name associated with a profs degree.

To clarify the post you're responding to: yes, I agree there are poor graduates from higher-ranked programs, and there are researchers well-respected by everyone. But there are also many researchers, perhaps younger, respected within their field, grad school, and current position, and it'll vary a lot across schools how well their LORs are received. In IO terms, I'm saying there's some vertical differentiation and some horizontal differentation in ad coms' preferences for the recommendations.

Originally Posted by YoungEconomist

I was referring to more actual randomness. I believe there are many times when admissions committees will probably have a significant portion of the pool that are more or less "equal" when all things are considered. In the end, if you have 20 "equal" people still being considered for admissions but only have 5 spots left, what do you do?

Well you have to make a decision, and the decision is ultimately not random. Again, in IO terms, we're talking about horizontal and vertical differentiation. There is some universal agreement in applicant quality (vertical), and then some disagreement (horizontal). For instance, some schools might put more weight on GRE scores and grades than others. Of course we have no way of observing what goes into the heterogeneity in preferences across school, so we can model it as being random, but it's never random. In reality, decisions get made, no two applications are exactly the same, and there's no doubt in my mind that the trust prof's have in LOR writers, which is largely based on social networks between the profs, can be a decisive factor in admissions decisions between very similar candidates -- and I maintain my hypothesis that access to those networks is more important than undergrad brand name.