@xiggi
I used to not really understand what you meant by Lake Wobegon and CC. Now I totally get it.
For some reason, it’s taken this thread for me to see it.
You give too much credit to the idea personal prefs can’t be wrangled.
Try to remember there are multiple reads on each kid past first cut. And common goals in building the class.
NSF, btw, starts with these bullets:
- What are the intellectual merit of the proposed activity?
- What are the broader impacts of the proposed activity?
While the elite is concerned with the class and a vibrant community.
A couple years ago, I tried to see if I could make a set of rules that would predict acceptance decisions for posters in the Stanford RD thread. I found that by using 3 simple rules, I could predict more than 80% of decisions for thread posters correctly. If I also considered URM/hook status, I could increase that rate to predicting more than 90% of decisions. Obviously thread posters on this site are a ridiculously biased sample that is not representative of the full applicant pool, but the point is the decisions didn’t look random at all. Even with only seeing a small portion of the application that did not include LORs or essays, I could predict the vast majority of decisions correctly. The “random”/“lottery” comments I’ve seen on this site usually relate to focusing on stats and putting little emphasis on the rest of the application. With such a method, the Stanford thread decisions I looked at would have appeared to be almost completely random. There was actually a slight negative correlation between most stats and acceptance decision among thread posters.
The bullet points you quote from the National Science Foundation (in #205) are part of the reason that I would expect NSF funding decisions to be more replicable than college admissions decisions, lookingforward. Yet they have some variability, as mentioned above.
To some extent, with the CC pool, what we can tell rather quickly is just how bright a kid may be. The bright kids process advice and understand strategy.
I’d be interested in knowing more, Data, maybe sometime a PM.
@lookingforward and @Data10
I did the same thing as Data did, back in 2004, shortly after I joined CC, and came up with remarkably similar results. For a self-selected pool, it’s not that difficult to see patterns; it really isn’t.
Totally. That’s part of it.
M2l “No one thinks that the Ivy League (and the 8 to 10 Ivy peers such as Northwestern, before Pizzagirl takes my head off) get all of the best students every year.” LOL
Well, not surprisingly, I would posit that, everything remaining the same --meaning same pool, same adcoms, same institutional demands, and same process but a different order in the reading-- that the result should be exceedingly similar. I think the same students who are subject to a final decisions would emerge and that the decisions would be … identical, safe and except for a handful if the adcoms happened to MISS a part of a file and caught the error upon a second read. I do NOT believe they’d change their minds about many students.
Fwiw, I do believe the often repeated statement that a school could have admitted a larger number of students (that 2 or 3 times the number of enrollment) but that does not mean --IMHO-- that they would have preferred to admit the other group.
Again, it all boils down to accepting the randomness or not. I have yet to see a single reason why I should start thinking the process is random. Reinventing the wheel and suggesting different scenarios does not change anything.
Q, there’s s lot of massaging, to get the finalists down to admits, form the final tapestry.
At that point, nearly every kid is a good choice. Yes, there can be differences if you could rerun. But still the same sort of content and variety.
Does that change it into random? No. It’s purposeful.
Quite so, as the British would say.
Purposeful decision-making does not have much in common with “random.”
I know we keep sliding off the orig question.
So, why can’t we tell kids, “You did the best you could. Now it’s in their hands and we don’t know all the factors they will consider this year or what this pool offers?”
And a la PG, ensure they’re thinking rationally. And a la thumper’s earlier advice, make sure they have a broad and happily list of schools they applied to?
The admission process can change over time. Duke has decided to allow applicants to be rejected after 1st read.
They have turned down 1/3rd after the first read when they went to the new process.
http://www.dukechronicle.com/articles/2012/03/28/university-streamlines-admissions#.VWURUc9VhBc
@Data10 Out of curiosity, what were those factors? Did you do a write up? I think it would be interesting to read it.
The obsessive chancers might be interested in learning those rules, too.
Intterestingly for me at least was that my younger son was so much better at this than the older one. Not surprisingly older son did about as well as his stats would suggest while younger son did much better.
Of course, mathmom’s older son could not have done much better–Harvard admit and Carnegie Mellon in computer science, as I recall.
@QuantMech Princeton would be better, obviously. [Here’s my proof.](http://colleges.usnews.rankingsandreviews.com/best-colleges/rankings/national-universities?int=9ff208)
It’s so embarrassing to be the type of person who obsesses over small differences!
Yog, couldn’t pick a worse group to try to press USNews on.
That was the joke.
The basic idea was that the vast majority of CC thread posters were academically qualified. However, the quality of out of classroom activities (ECs, awards, and more) was far more variable. Many posters had nothing more impressive than some clubs and maybe a sports team. Such posters were almost always rejected, even if they had a 4.0 and 2400. Instead the ones who were accepted usually did something out of the classroom that was impressive on at least a regional level, and often higher. I later came across a write up at http://phs.princetonk12.org/guidance/Forms/Betterton%20College%20Planning.pdf , which has a similar procedure, but a bit different thresholds than what I used.