Well, the problem with that is that practically nobody actually flunks out of medical school. The most difficult part of medical school - far and away - is simply getting in. But once you’re in, you’re virtually assured of not flunking out. </p>
<p>One rarely flunks out of medical school (graduation rates are close to 95 percent), so once a student receives an admission letter to any medical school, that student is guaranteed a place in the upper middle class.</p>
<p>I mean, surely there’s a less invasive way to standardize GPA, which would preserve adcoms’ access to a long-term diligence metric.</p>
<p>For law school, Berkeley Law used to use a metric that adjusted a student’s GPA based on the average LSAT of applicants from that school. Medical schools could use a similar metric. Medical schools could furthermore either adjust for major (e.g. further specification, but robs you of sample size) or only evaluate the core premedical requirements.</p>
<p>There’s still a couple of ways to game this system (taking easier profs within those requirements; undergrads encouraging low-GPA/high-MCAT students to apply even if they stand no chance at admission) but basically I think it solves a lot of the problems without robbing adcoms of a long-term metric.</p>
<p>I would guess that few, if any, flunk out of “graduate” school, either. Sure, a gazillion young adults never finish their thesis (“ABD”) and quit, but they don’t flunk out.</p>
<p>I remember this article a while back about gpa and mcat. I do see Sakky’s point about how students can “game” the system, but apumic is right on that the gpa should be a reflection of how seriously an applicant has taken his or her undergraduate studies.</p>
<p>I would love to see some empirical data to support all these assumptions about “harder” and “easier” grading at different colleges. It is almost impossible to compare HYPS to most other colleges, since their student bodies are composed almost exclusively of top students. Even at a place as lofty as Berkeley, most of their students would not be admitted to HYPS. Given that the HYPS student body, based on high school acheivement, is equivalent, roughly, to the top, what? quartile? at Berkeley, and even higher at most other places, if there were nationally uniform grading then HYPS should have high grades.</p>
<p>In fact the more selective colleges do tend to have higher grades, which implies that there is some degree of homogeneity of grading standards across colleges. Perhaps there are more A’s at Yale than at most colleges because Yale students are more likely to do A work?</p>
<p>Does someone have data, say a look at med school performance, or USMLE grades, for students from various undergraduate colleges, controlled for MCAT? If you could show that on average B+ students from college X got the same USMLE scores as A students from college, with equivalent MCATs, then you would have evidence that X’s grades were deflated relative to Y. </p>
<p>Without that kind of data, these claims of “hard” and “easy” gradiing schools are not even speculation, they are products of the imagination of the speaker, purely fiction.</p>
<p>But I don’t know whether there was ever any validity to this practice, which as you noted, they have discontinued. Presumably if they could have shown that it reliably predicted academic performance in law school better than the unadjusted grades they would have continued using it.</p>
<p>
</p>
<p>Again, assumes thay this would be desirable adjustments. I don’t know of any data to support this assumption. In fact, there is evidence that, controlling for undergraduate grades, engineers underperform in medical school. If “low grades in engineering classes” were true, then engineers should do better in med school than their undergraduate grades predict, not worse.</p>
<p>People major in things in which they do well. Most people in engineering or hard science are going to get better grades in these courses than in say English. Most English majors are going to do better in literature courses than in physics. The students sort themselves by their aptitudes and it might confound interpretation to pretend this does not occur. At many colleges a GPA minimum is required to progress in a major, and many discourage students who are passing but struggling to stick with a major they find difficult.</p>
<p>Someone who cruises to A’s in chemistry and works like a dog to eek out a respectable B in English freshman year is much more likely to major in chem than in English. Should they get a bump up in adjusted GPA based on an assumed harder grading standard in the absence of any data?</p>
<p>For med school the MCAT is a great equalizer. Since admissions gets both grades (how hard do you work? Do you show up every day prepared for class? Get assignments completed and submitted on time? Even when you are bored? When you are sick? When you are sad because you girlfriend cheated on you? Your football team lost to their big rivals? You lost your book money in a card game?) and MCATs (How much did you learn in your introductory classes which have a high degree of similarity of content across colleges?), they can do these comparisons without making up adjustments.</p>
<p>From a very preliminary look at data among law school applicants, top schools tend to be grade **deflated **when grades are LSAT-controlled. (I think this is why sakky has begun talking about top schools being non-stochastic instead of grade-inflated.)</p>
<p>It turns out I misstated the metric. My fault here.</p>
<p>It was based on law school grades and correlations for undergrads from those particular schools, so the validity was fine. The problem was that that it was hideously politically incorrect. Berkeley discovered that contrary to popular belief, prestigious Ivies and such were actually deflated. So when the list came to light, the UC system (which Berkeley is of course dependent on for funding) was horrified that Berkeley Law was downweighting grades from UC schools. Moreover, Berkeley was also downweighting grades from schools filled with less-affluent and racially diverse students, such as the CSU system. The idea was basically that Harvard students were already privileged; how horrible that Berkeley was further perpetuating this institutional discrimination. It was, essentially, an anti-affirmative action.</p>
<p>It wasn’t that the system was statistically inappropriate. The problem was that it couldn’t survive political uproar.</p>
<p>One could compare HYPS to MIT or, perhaps even better, Caltech. I think we can all agree that, academically speaking, the student bodies at all 6 of those schools are highly comparable, yet the grading schemes are vastly different. </p>
<p>
</p>
<p>I’ve said it before, and I’ll say it again: George W. Bush and John Kerry. Would there really be any dispute that if they had gone to MIT or Caltech, they would have flunked out? </p>
<p>Actually, I’ve begun emphasizing stochasticity at the low end because, frankly, above a certain GPA, additional grade inflation doesn’t really matter, as long as you also have a decent MCAT score and other application features. After all, what really is the difference between a 3.7 and a 3.8? </p>
<p>On the other hand, avoiding failing grades always matters. The best way to avoid a failing grade is to simply attend a school that doesn’t even assign them in the first place.</p>
So did they show that adjusting for grading standards across colleges actually improved prediction of law school performance after controlling for LSAT’s? Or did they just find that the adjustment tended to favor Ivy grads?</p>
<p>
</p>
<p>I agree one COULD. My question is whether anyone HAS. </p>
<p>
</p>
<p>Again, the data supporting this statement is???</p>
<p>
</p>
<p>OK I can’t resist. Since they did not go to Caltech or MIT, no one knows what would have happened. Maybe they would have been more committed students. Maybe not. But unless we have a deity among us, no one can tell us what would have happened under circumstances that never occurred.</p>
<p>As long as we are making untestable predictions, I predict that the 5 students who had the highest graduating GPAs from MIT last year would have flunked out of Michigan State had they attended. Since they did NOT go to MSU, my prediction has exactly as much validity as does the Kerry Bush claim.</p>
<p>Again, is there DATA, as opposed to fiction, supporting the claim that there exists a wide range of grading standards at these colleges? </p>
<p>How about a simple regression of mean gpa at a college vs SAT scores of the student bodies? That would leave a lot of ground uncovered, but it would be more evidence than “because I said so”</p>
<p>Sorry, this confusion is a result of my original misstatement. They didn’t correct for LSATs – I’ve done that analysis privately, but that wasn’t the work under discussion here. They carefully watched how well students from a particular undergrad school did during 1L year, then tried to see if there was correlation with undergrad GPA and what that correlation was.</p>
<p>It turned out that there was a predictable correlation. Harvard consistently 3.5’s did much better than CSU-LB 3.5’s.</p>
<p>Again, apologies for my original misstatement.</p>
<p>1) Grading differences b/w Top 20 schools are exaggerated. I want to see actual data on a school’s average GPA before calling them grade deflated. MIT, Cornell, Swarthmore, etc. were rumored to be grade deflated and yet when the actual numbers were released, they were not grade deflated.</p>
<p>2) I think, even with grade inflation, Top 20 colleges are still harder than state schools. There is limited data available. But my conjecture comes from a couple of things:</p>
<p>1) For schools that release the GPA/MCAT scores of their applicant, you can see that top colleges have a lower GPA but higher MCAT than the national average. For example, if you look at Cornell’s applicants:</p>
<p>The median MCAT of its WORST applicants (2.6-3.4 GPA) is higher than a 30. This is also seen in other top universities where their 3.5 applicants typically average a 30-32 MCAT score.</p>
<p>2) If you search on mdapplicants for applicants with LOW GPA’s (3.3 or below) and HIGH MCAT scores (35 or above), 97 (give or take a few) out of the 209 profiles that come up are of students from Top 25 universities as defined by US News. I didn’t even count the top liberal arts students or exclude profiles that didn’t list a school. </p>
<p>Yet, if you search for applicants with HIGH GPA’s (3.9 or above) but disappointingly low MCAT scores (29 or below), only around 8 (EIGHT) out of the 220 results that came up featured a top 25 university. For some reason, almost all of the 8 were students from UCLA (which is barely at the fringe of the Top 25).</p>
<p>Mdapplicants is self-selected. But, I think the differences between the two populations are quite striking. I think those who follow SDN would agree that most of the high MCAT, low GPA students seem to be from top colleges. </p>
<p>So, what is my point? Don’t go to a Top 25 college, even the ones that Sakky is always touting, expecting an easy ride. This doesn’t mean you should shy away from top colleges. After all, they still have good reputations, generally good advisors, low student/faculty ratios, good opps for volunteering/research/recs, etc. Just don’t expect them to be academically easier.</p>
<p>I would have to agree that many top schools tend to have lower GPAs than their pre-med counterparts elsewhere. My medical school has a very high average MCAT, but a relatively low GPA compared to other schools with similar MCAT scores. My school is also a bit more Ivy/private school heavy than the others.</p>
<p>I think Sakky needs to also remember that back when John Kerry and George Bush attended Yale, the average GPA was a lot lower. Back in the 60’s, the average GPA at Yale was around a C+. Today, the average GPA at Yale is around a 3.5. So, I think Kerry and Bush should be given a little more credit (or give a little less credit to the 3.5 GPA Yale students of today).</p>
<p>I think that the curved premed classes have a lot to do with it as well. Take a curved gen chem class at Cornell vs a curved gen chem
class at X average university. I would bet you the median at X university is easier to obtain that the median at Cornell. You’re comparing your performance to a higher level than that found at X uni. </p>
<p>This is an overexaggerated analogy, but say I’m a Mustang, X is a Toyota camry, and Cornell is a Corvette. Yeah my mustang looks pretty good compared to that Camry, but doesn’t look so hot next to the Vette. Does that mean my car got crappier? No. </p>
<p>I believe bluedevilmike investigated this question and published data before. </p>
<p>
</p>
<p>Come on now, I think we know full well that we have to make reasonable assumptions about what we all think would happen in counterfactuals. Simply because nobody can run a ‘multiverse simulator’ doesn’t mean that we can’t make reasonable assumptions.</p>
<p>Otherwise, I can just as easily predict that if I had an extra inch of height, then today I would be President of the United States, have won every single Nobel Prize in the world, would also be the richest man in world history, would also have won 10 straight Best Actor Oscars, and be married to both Jessica Alba and Scarlett Johansson in a polygamous relationship (which would be legal under my Presidency). After all, since nobody actually knows what would happen if I was an inch taller, I can predict whatever the heck I want, using the exact same untestable ludicrous logic you just did, right? </p>
<p>At the end of the day, I think we can agree that Caltech and MIT grade harder than do HYPS, something that I doubt even the students would HYPS would seriously dispute. Furthermore, since Bush and Kerry barely graduated from Yale, I think it is a reasonable statement to say that they would not have graduated from MIT or Caltech. </p>
<p>
</p>
<p>Again, reference some of bluedevilmike’s old posts. </p>
<p>But all of that is missing the point anyway. See below. </p>
<p>
</p>
<p>And similarly, MIT and Caltech similarly also used to grade lower than they do now. Ben Golub, a strong Caltech proponent, has freely admitted that a far lower percentage of students used to graduate from Caltech than they do today. </p>
<p>
</p>
<p>I’ve said it before, and I’ll say it again: what matters is not grade inflation as defined simply by average grades, but rather the stochasticity of grades. Again, the difference between a 3.7 and a 3.9, given a strong remaining application package, is minimal, for you’re still going to be admitted to some med-school, even if perhaps not a top-ranked one. Heck, even Mike McCullough himself freely admitted - and I agree - that certain Stanford premeds may actually be better off taking their premed coursework at a community college rather than Stanford. </p>
<p>What matters far more than simply earning top grades is avoiding bad grades, and to that, the top non-tech private schools offer far more relaxed environments. As bluebayou pointed out, Stanford doesn’t even assign F grades at all, and while the others may potentially do so, the odds of anybody actually earning one are practically nil.</p>
<p>I’d like to pursue this line of logic further: I would like to see some empirical data to support all of the assumptions that have ever been stated on at least the premed section of CC, if not all of CC. </p>
<p>For example, I’ve seen numerous posters - including you afan - provide a ‘recommended’ path towards being admitted to med-school. But does anybody actually know that these pathways are recommendable? Is there empirical statistical data to prove that those who do follow the recommended steps actually perform better than those who do not, after controlling for levels of motivation and other situational unobservables - which I presume could only be performed through randomization? Not only does such data not exist, it couldn’t exist. Nobody can take a bunch of students and randomly order some of them to follow a certain premed curricular strategy and others not to, and then compare the outcomes. </p>
<p>For example, how many times have I seen posters recommend that premeds should simply major in something they like, and they presumably will do better because their enjoyment of the major will translate into a superior college experience, and probably also a higher GPA. But why? Has anybody actually demonstrated that empirically? Has somebody randomly ordered some students to major in what they like and others to major in something they don’t like, and then compared the outcomes? Since we seem to be in a world of untestable predictions, I could just as easily predict that if everybody had majored in a subject that they loathed, then they would be more successful and happier people.</p>