GPA - How accurate is this? [PaulB, Hjort others..?]

Just to summarize, this is an admission consultant's website that ranks the various application elements in order of importance. His data is a bit dated, for example he lists average GMAT at top schools at 680, and the middle 80% at 630-730, where the current average is around 710 and the middle 80% is about 660-760. He ranks them in the following order:

1. GMAT
2. Essays
3. Timing
4. GPA
5. Work Experience
6. Recommendations
7. Interview

I can agree that work experience is very important, but as we have discussed before, there's a very good reason why it's difficult to use as an admissions criteria. The problem is that it's extremely difficult to determine what work experience most impressive. Consider the following:
1. 2 years in a regional consulting firm
2. 3 years in management training with a fortune 500
3. 4 years as an entrepreneur
4. 5 years in IT

Which is best? The answer is "who the hell knows"? Other than a few name brand feeder companies, schools have a really difficult time differentiating between work experience of different individuals. I believe that, unless you really have unique work experience, this criteria cannot help your application much.

GPAs have the same problem. Is a 3.0 in Physics better than a 3.3 in English? Is a 3.1 from an Ivy better than a 3.4 from a top state school? Some schools are known to have grade inflation, so how does that figure in? It's just impossible to determine how competitive a major, or a department, or a school is; so GPA has limited value as a comparison tool between applicants. The other problem with GPAs is that for many applicants, they can be 3, 5, 7 or more years old - so their correlation to the candidates current profile may be limited as well.

The problem with recommendations is that most successful candidates have stellar recommendations. This is a natural result because each person selects his own recommenders. A negative (or neutral) recommendation can be very damaging, but a stellar recommendation probably doesn't help a whole lot because they (should be) very common.

I understand that the importance of interviews varies according to the school, but I can understand why they are ranked so low. Interviews are useful to support or contradict other parts of the application but probably don't have much value on their own. In fact, if the rest of your app is not up to snuff, it won't even matter if you are the best interviewer in the world because you'll never get the chance (at most schools).

He sort of rolls some of the work experience related stuff up into #2 - the essays. What he addresses under #5 is really just the amount of work experience, i.e. the importance of having 5 years instead of 3 years. From that standpoint I can buy his "ranking."

Yeah,

He should change 2. to essays/quality of work experience and 5. to number of years of work experience.

"First, most B-school applicants were not planning to apply to graduate programs, so they might not have focused on playing the grade game the way most premed or pre-law students did. Why hold mediocre grades against an applicant who had no intention of applying to grad school five years down the line?"

I found this statement from the website suspect in at least two ways. First, applicants to the top clusters have often been planning to enter business school for years and even selected their immediate post college employment with an eye toward being ready for business school after a few years of work experience. Admissions staff have few qualms about holding these studnets responsible for an undistinuished record. Second, why should admissions staff let students off the hook for mediocre grades when the students were not expecting to go on to graduate school? Individuals with an internal drive to excel (who the top clusters are presumably looking for) would endeavor to succeed in college even if they had no definite plans for further academic study.

To be fair, however, I concede that any model of a complex and subjective process like admissions is likely to be false. I am reminded of Professor Box's statement that (roughly) all models are false but some are still useful.

I agree that it sounds like a gross overgeneralization. But how do you explain that Med-school and Law-school applicants have much higher GPAs. Do you think they're smarter? My college buddies who went those routes spent their college years locked away in the library (especially the Med-school types) but I never felt like they were smarter. They were definitley more driven than the average liberal arts majors.