Berkely alum wins Nobel Prize

Colleges and Universities A-Z University of California - Berkeley
81

<p>
[quote]
So what major would this be? The largest majors at Berkeley, MCB, Poly Sci, etc. seem to be able to handle all the students who want to enter the major (i.e. no impaction). And I think we hypothesized before that many of the impacted majors could probably be not impacted, and that the majors could probably hold all the students who want to major in that major if some things were done differently, so again it doesn't become a matter of "transfers taking spots from freshmen admits" but that Berkeley chooses to run things so that the majors cannot hold all the transfers and freshman admits and weeding must take place.

[/quote]
</p>

<p>Like I said, these were made-up numbers. But it was all to illustrate how weeding actually works, and specifically how the transfer system actually INCREASES the problem of weeding. Essentially, freshman-admits have to be weeded more in order for the transfers to be weeded less. You are robbing Peter to pay Paul. No wonder Peter is angry. </p>

<p>
[quote]
But following your logic, if you want to weed out transfers so that not as many freshan admits are weeded, the net weeding will still be around the same. It's just that you weed less freshmen admits and weed more transfers. Which I suppose would seem more fair in that both groups are being weeded.
But a simpler solution seems to simply admit less transfers in the first place? You could simply raise admissions standards and admit less transfers without having to try to weed them, unless I'm missing something here.

[/quote]
</p>

<p>I view this as a multiprong strategy. Obviously the ultimate solution would be to eliminate weeding completely. But that's a long-term goal that won't be achieved anytime soon. So in the interim, you do things to properly manage the impact of weeding. Spreading the impact of weeding fairly would therefore be one such management technique. </p>

<p>Remember, this is not an 'either-or' situation. Multiple solutions can and should be utilized to manage the situation. </p>

<p>
[quote]
Hmm, I don't know about this. To have what you will study for the next two years be based on one test? A student could be sick, late due to extenuating circumstances, not a good test-taker, nervous, etc. etc. There is a reason why the college admissions process isn't based on one SAT score. In fact most people take it more than once because their scores often varies by a significant amount.

[/quote]
</p>

<p>Nobody is saying that it is a perfect solution. There is no perfect solution short of eliminating impaction itself, which I said is a long-term solution. However, like Keynes once said, in the long-term, we are all dead. We have to deal with the situation the way it stands. I believe using a universal test is a good short-to-medium term solution.</p>

<p>Besides, look at it this way. Lots of European and Asian educational systems use big entrance exams to determine whether you can move up a grade or are eligible to enter a particular college. The Indian Institutes of Technology in particular assign admissions spots all based on one test. Even nowadays, many American K-12 systems are beginning to implement tests to determine grade promotion. You don't just automatically move from grade X to grade X+1 in many American schools. You have to pass a test to move up. </p>

<p>You can also look at the various professional licensure requirements, many of which are test-based. Becoming a practicing lawyer doesn't just involve graduating from law school. You have to pass the bar exam. You could have graduated with perfect grades from Yale Law School, yet still fail the bar exam and thus be ineliglbe to practice law. To become a practicing doctor, you have to pass the USMLE. You could have graduated with flying colors from Harvard Medical School, yet still fail the USMLE and thus be ineliglbe to practice medicine. Many other jobs require that people pass a test before being eligible to work or eligible for promotion. </p>

<p>The point is that the use of a test to determine eligible for something is not a radical suggestion, but rather is a mainstream technique that is currently used in a wide variety of contexts. You might say that somebody might get sick for the Berkeley impaction exam. On the other hand, prospective lawyers might get sick for the Bar exam. Yet apparently the legal profession has figured out a way to handle this eventuality. </p>

<p>What I am saying is that using a universal test is MORE fair than just using grades. Is it perfectly fair? Of course not. No system is. But I think it is BETTER. Why? Because it's standardized. You give the same exam to everybody at the same time, so everybody is being judged on the same standard. The problem with grades is that no standard exists. Some classes and some profs are easier than others. That fact then drives people to cherry-pick the easier classes and easier profs in the hopes of getting higher grades.</p>

82

<p>
[quote]
Basing it on one test...that's just too much stress right there (can you imagine the pre-meds fighting for MCB?).

[/quote]
</p>

<p>I would actually argue that the use of one test could be a method of actually making things LESS stressful overall. For example, with an entrance test, departments would no longer have to weed. That's because they would have already gotten rid of all of the students who weren't good enough via the test. </p>

<p>The current structure of MCB basically entails that MCB is a 4-year stress marathon, or a "stress-athon". As an MCB student, you are constantly worried about your grades throughout your entire time at Berkeley. With an entrance exam, I can concentrate much of that stress in just one point in time. If you pass this exam and get to the upper division, you will be able to take classes that have been converted to become less stressful (because they wouldn't have to weed). Hence, I don't think I have actually increased the amount of total stress. I have simply shifted it around in time. </p>

<p>Furthermore, I would actually argue that the total stress might actually be reduced. Think of it this way. With an entrance exam, I would be able to get rid of all weeder classes, because the entrance exam would take the function of the weeders. This gives you a mechanism to inflate grades. The reason why grade inflation is resisted by the faculty today is because they feel that they need to weed. The entrance exams give them that weeding capacity, which would therefore remove the resistance to grade inflation. The entrance exams would also not be part of your transcript. So graduate schools would never see that you flunked entrance exams to certain majors. All they would see are your (higher) grades. </p>

<p>What happens currently is not only that the weeder courses serve to weed people out, but that the weeder grades are then PRINTED on your transcript, just like the grades of any other class. That increases the level of stress of the weeders because you not only know that the class might weed you out, but then future grad schools and employers will be able to SEE your bad weeder grade. Using a confidential entrance exam removes this problem. If you couldn't get into MCB or EE or whatever it is, nobody will know that except you. Hence, I would argue that the overall stress level would actually be REDUCED.</p>

83

<p>Maybe you should go visit some Asian countries and talk to those students whose college admissions is based soley on one test, and ask if they think it's less stressful.</p>

<p>Heck, go on CC right now and ask if students would prefer to be admitted to college based on 4 years of high school, SAT, essays, extra-curriculars, etc., or just the SATs. I bet almost all (except maybe high SATers) would prefer the latter.</p>

<p>Just because the test is one-time doesn't mean the other years will be stress-free. If the process is based on one test that just means students will stress for years based on one test. We already have high schoolers worried about MCAT scores. If med school is solely based on the MCATs you can bet there will be plenty of college students stressing/studying for four years for the MCATs.</p>

<p>Furthermore, one test actually adds MORE pressure. Sure being a pre-med is tough, but if you screw up on one test in one class it's not the end of the world. But with just ONE test, there is no room for error, and the pressure tremendously higher.</p>

<p>Why does Harvard have grade inflation compared to Berkeley? A main reason is that Harvard classes have more projects/homework and less tests compared with Berkeley's classes, which are often 2 midterms and a final. This seems to be the case in many tech/sciences classes, where most Berkeley students receive poor grades. Tests are much more arbitrary than homework/projects and it's easier to do poorly on them, and the fewer number of tests, the more arbitrary the grading is.</p>

<p>Think about it: how many college students would rather have the high school structure with many tests/homework assignments, and how many high schoolers would rather have their classes be based on a midterm and a final? Which is more stressful, studying for a chapter test in high school or a midterm in college?</p>

<p>Sorry, I just don't think it's too arbitrary to base everything on one test. In places like China, they reall have no choice because there are simply too many people. Thus, the college admissions test as a sole factor is really the result of overpopulation and a last resort.</p>

<p>And I simply don't like the idea of weeding transfers in general, because it also holds true for what I said above: it's too arbitrary. We can simply raise standards for the admission of transfers.</p>

<p>I think I see your point, but I don't agree with your methods.</p>

84

<p>
[quote]
Maybe you should go visit some Asian countries and talk to those students whose college admissions is based soley on one test, and ask if they think it's less stressful.</p>

<p>Heck, go on CC right now and ask if students would prefer to be admitted to college based on 4 years of high school, SAT, essays, extra-curriculars, etc., or just the SATs. I bet almost all (except maybe high SATers) would prefer the latter.

[/quote]
</p>

<p>That's not the right question to ask. Remember what we are talking about. We are talking about Berkeley students who want to enter those majors who currently have to undergo a long series of weeders. </p>

<p>Why not ask them what their choice would be out of the following choice? Specifically, the choice would be this. Take a student who currently wants to get into, say, the BA program in CS. He can have the choice of either keeping with the current system, where his admission to the major is contingent upon getting high grades in the CS prereqs, and the CS prereqs would be graded as weeders. Furthermore, even if he does get into the major, the harsh grading doesn't stop, but continues on in the upper division (although less intensely). So he can have that situation. Or he can trade all of the harsh grading for a system where his admission to the major is based on one big test. No weeder-style grading on any of his classes. Just one big test, and if he doesn't score well on this test, he doesn't get into the major.</p>

<p>I don't know what the answer will be, but I strongly suspect that most current and former Berkeley students who survived weeders would take the latter. I know I would have. Not because having a big entrance exam is stress-free, because it clearly is not. But it's a heck of a lot better than having to survive the weeder gauntlet. Trust me. </p>

<p>Your analogy breaks down in another way. The SAT is not really something that you can study for. Yes, there are prep classes and so forth, but the truth is, for the most part, you either know the material, or you don't. However, this is not the case for an entrance exam for a major that will consist of material that you can actually study. After all, if it is possible to study to get good grades in pre-req courses, then it is possible to study your way to a high score on the entrance exam. </p>

<p>
[quote]
Furthermore, one test actually adds MORE pressure. Sure being a pre-med is tough, but if you screw up on one test in one class it's not the end of the world. But with just ONE test, there is no room for error, and the pressure tremendously higher.

[/quote]
</p>

<p>This is also not a valid concern. First off, I never talked about reforming med-school admissions requirements. Med-schools can do whatever they want. Berkeley doesn't have a med-school anyway. </p>

<p>I am simply talking about a structure of determining who should get into impacted majors. Obviously we both agree that the best thing to do is to just get rid of impaction and weeding completely. But presuming that this is not on the table, then the next thing to do is come up with a fair way of figuring out who should be allowed into these impacted majors and who shouldn't be. You were the one who talked about MCB in conjunction with premeds, and I went along for the ride, even though, really, we shouldn't be talking about MCB at all because MCB is not impacted, and I hope never will be. </p>

<p>But assuming that one day if MCB is impacted, how should MCB seats be assigned? One way is to use the current system of weeders. I prefer a second way of just using an entrance exam. Like I said, you either know the material or you don't. Let's face it. Some people who get high grades don't really know the material, either because they deliberately took easy grade-inflated classes, or because they are really good at kissing butt with profs, or whatever it is. Conversely, some people who have low grades actually know the material. If you're going to have impaction, you need to use the most fair process to decide who gets in, and the use of an entrance exam is the fairest way. Is it completely fair? Of course not. No method is. But it's MORE fair than the current setup. </p>

<p>
[quote]
Why does Harvard have grade inflation compared to Berkeley? A main reason is that Harvard classes have more projects/homework and less tests compared with Berkeley's classes, which are often 2 midterms and a final. This seems to be the case in many tech/sciences classes, where most Berkeley students receive poor grades. Tests are much more arbitrary than homework/projects and it's easier to do poorly on them, and the fewer number of tests, the more arbitrary the grading is.

[/quote]
</p>

<p>I completely disagree with your premise. I am not aware of any reason to believe that Harvard classes have any more homeworks and fewer tests than Berkeley classes have. I can name you a slew of Harvard classes that are graded based on exams only, with homeworks perhaps serving only as tie-breakers (and sometimes not even that). </p>

<p>I also disagree that tests are any more arbitrary than homeworks/projects. If anything, they are LESS arbitrary. Tests are performed in a controlled and proctored environment. With homeworks, it is extremely easy and prevalent for people to just copy other people's homeworks. After all, nobody is watching you when you do your homework. In fact, it is actually widespread practice in certain engineering classes for groups of students to simply break each problem set into pieces - i.e. you do #1, Jim does #2, Fred does #3, etc., and then everybody get back to together at the end of the week and just copies all the other problems that they didn't do. This is technically against the rules, but it is a widespread practice nonetheless. Hence, what that really means is that your homework score now becomes a function of how many friends you have and how smart they are, and that is pretty arbitrary. This is further exacerbated by the issue of people getting access to previous student's official solutions (because profs often times won't change the homeworks). Basically, the people who were able to wheedle their way to getting access to old solutions (or friends of these people) got high homework scores, and everybody else got bad scores and got weeded out. In fact, in the more extreme weeder courses, it often becomes a desperate battle of survival. You have to get old solutions from previous classes simply because you know that other people will get them too, and you can't afford to lose out on the curve and get weeded out of the major. </p>

<p>The major problem of projects is the extremely arbitrary nature of the grading of them. That's because most projects tend to consist of different topics, and it is impossible to actually use a completely equitable grading scale when people are presenting projects on different topics. For example, if you have to grade 2 chemical engineering design projects, one based on the manufacturing of beer, and the other based on the manufacturing of a pharmaceutical, and both projects are reasonably well done, but you have to weed one group out, can you really judge that one project deserves a higher grade than the other? It basically becomes a matter of taste - i.e., you just happen to 'like' beer, so you give the beer project a higher grade and you weed out the other group. But clearly you can see the tremendous arbitrariness associated with that. Then it becomes a game of students trying to figure out what kind of project you really like, etc. etc., again, all because nobody wants to be the one that gets weeded out. </p>

<p>An entrance exam is extremely fair because EVERYBODY has to answer the exact same questions. Hence, you control for most variables. It is a basic tenet of statistics that if you want to observe the nature of one variable, you should try to control all of the others. Using homeworks and projects simply introduces a huge number of moving parts into the equation. </p>

<p>Now, I can see a situation where a project could be used where everybody had to do the exact same project. But that would basically be the same thing as an exam, except now the project IS the exam.</p>

85

<p>The bottom line is that I think it's easy to say in the abstract that entrance exams are a bad thing. But weeders are a bad thing too. The question is, which is worse? I am fairly certain that people who have experience with weeders might like to try to use an entrance exam instead. Like I said, the MAJOR advantage of it is that the results of the exam are hidden from outside parties. The only thing this exam will do is decide whether you can get into a certain major. But no outside parties will ever get to see your entrance exam score. </p>

<p>Contrast that with the current situation. Let's say there is a guy who wants to get into the BA CS program. So he takes the CS prereqs, and he gets bad grades in them. Those bad grades exist FOREVER. Anybody who views your transcript will see those bad grades. So not only is this guy weeded out of the CS major, but his chances of going to grad school in any subject are now hurt because of his bad weeder grades. For example, let's say he ends up majoring in English, and later wants to go to law school. Well, his poor grades in those CS weeder classes are going to hurt his chances - EVEN THOUGH HE NEVER ENDED UP MAJORING IN CS. That doesn't matter - the fact that he took those weeder classes and got bad grades in them hurts him forever. What if he does really well in English, so well that he wants to compete for the Rhodes Scholarship? Again, his old bad CS grades are going to hurt his chances. What if he wants to graduate with distinction? That honorarium is determined by his GPA at graduation, which includes his old bad CS weeder grades. So he may do great in English, but the fact that he tried to get into CS and got a bunch of bad weeder grades may pull his GPA below the cutoff. </p>

<p>In other words, he is actually permanently hurt by trying to get into CS and not making it. He would have been better off if he had simply never tried to get into CS at all, and instead had just majored in English from day 1. And that's sad.</p>

<p>Contrast that with my proposed system. In this system, he would take the CS entrance exam, and fail it. So he can't get into CS. But his transcript would be unmarred. The results of his entrance exam would be hidden from prying eyes. He would be able to pursue some other major and some other career with a clean slate.</p>

86

<p>How would you write an entrance exam for a prospective CS major who might be interested in the subject but has no experience whatsoever in programming?</p>

<p>CS 61A used to have an entrance exam, but it's been rid of.</p>

87

<p>Well, obviously the guy better learn some programming then.</p>

<p>What I am proposing is to replace the current weeder status of all of the CS prereqs (i.e. the CS 61 series, CS 70) with a full entrance exam. In other words, all of those classes would be graded as regular classes, not as weeders as they are now (especially CS 61B). But to get into the major, you have to pass an entrance exam. Your grades in those CS classes won't matter - what will matter is your performance on the entrance exam.</p>

<p>I see multiple benefits to doing this. First off, it would allow people who just want to learn some basic CS theory, but who have no intention of actually majoring in CS, to take those courses. I know people who just wanted to learn some basic CS, but didn't take to take those CS courses because they didn't want to subject themselves to weeder grading and possibly fail the course. Seriously, if you just want to take the course to learn the material but you don't actually want to enter the major, then why should you be forced to undergo weeder-style grading? </p>

<p>It would also, like I said, help out those people who aren't able to enter the major, as these people's transcripts would not be marred by rough grading. Like I said, the results of your entrance exam will not be printed on your transcript. So if you fail the entrance exam, you can pursue another major with a clean slate. </p>

<p>Furthermore, I think the biggest advantage is the fairness of an entrance exam. It is the fairest and most reasonable method. It's not perfectly fair, as there is no perfectly method available. But it is probably the closest you can get. After all, like I said, right now, we have the situation where transfer students get to skip over some (sometimes all) weeders. That's clearly not fair. Everybody who is trying to enter an impacted major ought to be judged on the same scale.</p>

<p>Think of this situation. Right now, Berkeley offers multiple tracks of the various CS courses. For example, there are several CS 61B lectures taught by different profs. However, many people who take those courses are trying to get into the BA CS major, the profs who teach those different tracks have to get together and devise a standardized method of grading amongst each other. Otherwise you have the situation where students tend to prefer to take the particular track that is taught by the 'easy' prof who is known for easy grading and easy assignments, and shun the prof who is known for harsh grading and hard assignments. After all, when you're desperately fighting to get into a major, you are obviously going to want every single advantage you can get, and if that means choosing the easy prof, then so be it. So to prevent this, the profs have to standardize their work and their grading so that there is no such thing as an 'easy prof' - hence, that they are all the same.</p>

<p>However, if you're going to need to standardize the courses anyway, you might as well standardize the whole darn system through the use of an entrance exam. Using a single unified exam is the most standardized system you can have, because everybody then takes the same exact test under the same conditions. </p>

<p>Now obviously the best solution of all is to simply get rid of impaction and get rid of weeding entirely. I have always endorsed this as the ultimate solution. But presuming that you have to live with impaction and weeding, you at least need to make it fair.</p>

88

<p>Sounds like a plan. Why don't you actually write to the Cal administration? You have lots of good ideas, but expressing them on a message board intended for prospective students won't be very effective.</p>

89

<p>Virtually all of sakky's proposals would be ignored. As far as undergrad is concerned right now, the Berkeley administration is really only interested in increasing socioeconomic "equity and inclusion."</p>

90

<p>
[quote]
Why not ask them what their choice would be out of the following choice? Specifically, the choice would be this. Take a student who currently wants to get into, say, the BA program in CS. He can have the choice of either keeping with the current system, where his admission to the major is contingent upon getting high grades in the CS prereqs, and the CS prereqs would be graded as weeders. Furthermore, even if he does get into the major, the harsh grading doesn't stop, but continues on in the upper division (although less intensely). So he can have that situation. Or he can trade all of the harsh grading for a system where his admission to the major is based on one big test. No weeder-style grading on any of his classes. Just one big test, and if he doesn't score well on this test, he doesn't get into the major.

[/quote]
</p>

<p>Well, look, how are weeders graded anyway? 25 - 30% of the class fails? I think last year there was a math 1B professor who failed 40% of the class and was put on probation, so that estimate looks right to me. So what would happen if these classes don't serve the purpose of a weeder anymore? Maybe it will go down to 20%, 15%. Even if you replace weeders with a test the grade distribution won't be wildly different, because Berkeley probably isn't going to inflate its grades like HYPSM anyway. (and according to gradeinflation.com, Harvard's average GPA was 3.38 in 1996 and Berkeley's was 3.10 in 1996. I suspect both have risen a little but the difference isn't THAT large). So the grading will be a little easier. For most students either they have the same grade (the top students will still get As, the lowest students still get Fs, the second tier students will still get Bs) or maybe slightly better (B instead of B-). This probably won't have a great effect on most students. And I am still of the opinion that most students would want to be judged based on two years' worth of work than one test. The only students who may prefer the test, are those students who did badly in the major requirements, but let's face it, if they did so poorly they probably wouldn't pass the entrance exam anyway.</p>

<p>
[quote]
This is also not a valid concern. First off, I never talked about reforming med-school admissions requirements. Med-schools can do whatever they want. Berkeley doesn't have a med-school anyway.

[/quote]
</p>

<p>Right. I was only talking about it to show that most pre-med students probably don't want the admissions to be based solely on the MCATs, and that having one test (MCATs) wouldn't really reduce the pressure because if a test is that important students will start to prepare years in advance. So, applying that to this situation, having one entrance exam probably won't reduce the pressure, like you suggested.</p>

<p>
[quote]
I completely disagree with your premise. I am not aware of any reason to believe that Harvard classes have any more homeworks and fewer tests than Berkeley classes have. I can name you a slew of Harvard classes that are graded based on exams only, with homeworks perhaps serving only as tie-breakers (and sometimes not even that).

[/quote]
</p>

<p>Here is a quote from calkidd on an old thread:</p>

<p>
[quote]
Let me put the grade inflation thing into perspective for you Harvard people out there. First off, I've studied at Harvard, MIT and UC Berkeley, so I have experienced the coursework and grading at all three places. To the statement "But in the end, you're competing against the best and brightest kids in the country, and every class that you take has whiz kid who specializes in that subject area" - yeah, that's why you shouldn't expect an A. If an EECs whiz kid at MIT takes a hardcore physical chemistry class, he doesn't expect an A- or even a B+, even after "eons of work." </p>

<p>Second, the way in which grades are obtained: at a public school like Berkeley and to some extent at MIT, grades are determined more by test scores, which is a far more arbitrary way to determine grades than with projects or homework assigments (which constitute a higher proportion of grades at Harvard); after all, Harvard students are going to put in those all nighters to finish the projects right? You're dreaming if you don't think that students at UIUC or Cornell wouldn't do the same thing if they were given the chance.

[/quote]
</p>

<p><a href="http://talk.collegeconfidential.com/showthread.php?t=9848&page=4%5B/url%5D"&gt;http://talk.collegeconfidential.com/showthread.php?t=9848&page=4&lt;/a&gt;&lt;/p>

<p>I've also noticed this. Most of my classes are based on 2 midterms and 1 test, while my friend who is currently attending Harvard has all these problem sets and projects (projects? Don't see too many of those at Berkeley).</p>

<p>
[quote]
I also disagree that tests are any more arbitrary than homeworks/projects. If anything, they are LESS arbitrary. Tests are performed in a controlled and proctored environment. With homeworks, it is extremely easy and prevalent for people to just copy other people's homeworks. After all, nobody is watching you when you do your homework. In fact, it is actually widespread practice in certain engineering classes for groups of students to simply break each problem set into pieces - i.e. you do #1, Jim does #2, Fred does #3, etc., and then everybody get back to together at the end of the week and just copies all the other problems that they didn't do. This is technically against the rules, but it is a widespread practice nonetheless. Hence, what that really means is that your homework score now becomes a function of how many friends you have and how smart they are, and that is pretty arbitrary.

[/quote]
</p>

<p>I don't agree that homework is a function of how many friends you have and how smart they are. You are right to an extent, but most people I know either do their homework, get help from the professor if they don't understand something (and thus learn it), get help from classmates (not copy, but get help), or find homework solutions (but they try to understand it first and copy as a last resort).</p>

<p>But I mean less arbitrary in the sense that if you do the homework, you'll probably get most/all the points. You do the work, you learn the material, and you get the points. With tests, even if you know the material, you may not get a good grade. I'm sure you know how often students tend to make stupid mistakes on tests, or when a test just gives an arbitrary and obscure question (especially in weeders), and when there are 5-7 problems on a test, that really hurts. This seems to be especially true of math courses, in which errors in calculation/steps are all too common.</p>

<p>Another quote from calkidd:</p>

<p>
[quote]
More arbitrary grading leads to "harder" grading in the sense of a bigger distribution of grades; that means that someone who makes careless errors during in class exams but still understands the material might very well end up getting a B- or C+. When your grade is mostly based on projects and hwk assignments, you don't account as much for the arbitrary factors involved with test taking; you further have the opportunity to get help from teachers and TAs.

[/quote]
</p>

<p>
[quote]
The bottom line is that I think it's easy to say in the abstract that entrance exams are a bad thing. But weeders are a bad thing too. The question is, which is worse? I am fairly certain that people who have experience with weeders might like to try to use an entrance exam instead. Like I said, the MAJOR advantage of it is that the results of the exam are hidden from outside parties. The only thing this exam will do is decide whether you can get into a certain major. But no outside parties will ever get to see your entrance exam score.

[/quote]
</p>

<p>That's a tough question to answer. What I do believe is that basing something this important on one entrance exam is a bad thing. If your main concern is hiding bad grades on a transcript, maybe you should propose a change in how transcripts are revealed to outside parties (a la MIT), or if your main concern is how many students are failing, maybe you should propose a makeover in weeders (I like the idea of easing on weeders/tech courses and making some of the "fluff" majors harder). But you can certainly accomplish the two without changing anything regarding how one tests into a major. At least we agree that ideally, every student should be able to get into the major he/she wants to without weeders or entrance exams.</p>

91

<p>
[quote]
Well, look, how are weeders graded anyway? 25 - 30% of the class fails? I think last year there was a math 1B professor who failed 40% of the class and was put on probation, so that estimate looks right to me. So what would happen if these classes don't serve the purpose of a weeder anymore? Maybe it will go down to 20%, 15%.

[/quote]
</p>

<p>Actually, it would probably go down to about zero. I can imagine a scenario where CS classes are graded the way that say, American Studies classes are now, in which practically everybody passes. In fact, I would argue that most humanities and social science classes at Berkeley today are ones in which practically everybody passes. They don't get all A's, obviously, but very few people actually don't pass. </p>

<p>
[quote]
Even if you replace weeders with a test the grade distribution won't be wildly different, because Berkeley probably isn't going to inflate its grades like HYPSM anyway.

[/quote]
</p>

<p>I agree that most likely, nothing will ever happen. But hey, that doesn't mean that you stop trying. It's one thing to try and fail. At least you tried. That's better than never having tried at all. After all, if you don't even try, you are GUARANTEED to fail. </p>

<p>
[quote]
Right. I was only talking about it to show that most pre-med students probably don't want the admissions to be based solely on the MCATs, and that having one test (MCATs) wouldn't really reduce the pressure because if a test is that important students will start to prepare years in advance. So, applying that to this situation, having one entrance exam probably won't reduce the pressure, like you suggested.

[/quote]
</p>

<p>You stated two things there, both of which I disagree with, as well as a tangential third topic of which we apparently disagree. First off, I don't think it's true that prospective med students won't want everything to be based on one test. I am fairly certain that most people who have undergone the weeder premed classes would probably want admissions to be based purely on a single test. This would clearly include anybody who got a B or less, and I think it's fair to say that the majority of weeder grades in the premed courses are B's or less. Obviously the people who got A's in the weeders would want to keep the present system, but only a minority of them get A's. Hence, it would seem that by virtue of statistical definition, the majority of premeds would want to have everything based on one test. </p>

<p>Secondly, I still maintain that stress levels would go down from the present system, not because the using a single test is stress-free (it is in fact, high stress), but because the present system is ALSO high stress, in fact, probably higher stress than the alternative. You talk about people possibly studying for years in advance for the MCAT. Well, I already know people who are already studying for years in advance for their college OChem class. I know people who have already read the entire OChem book, and bought a bunch of those Schaum's practice workbooks and done all of the problems in them - all before they even officially take the OChem class. Heck, I even know some people who 'secretly' audit OChem by going to all of the lectures and taking all of the notes (but not actually registering for the class), in the SEMESTER BEFORE taking OChem officially. I think it's fair to say that that's pretty darn high-stress too. It's understandable stress because these people are doing anything and everything they need to do to get an A in the class. </p>

<p>Again, you can reduce the stress by managing it better. The present setup is a 4-year "stressathon". With my proposed system, you are in charge of managing your stress. If you want to take a semester off, and not study so hard, you can do it as long as you pick up the pace later. You can't do that in the current setup because one subpar semester can knock you out of the running. Hence, you have to be running at full-speed for all 4 years. What's more stressful, running at full speed all the time, or having the freedom to control your pace? For example, you could choose to run half-speed one semester, and then run double-speed in another semester. It would be up to you. You would have greater control over your own output. I think that's clearly a lower stress situation. </p>

<p>But thirdly, and most important of all, it doesn't matter. This is not about 'pleasing the students'. Heck, if this was really about just giving the students what they want, then we should be giving all students guaranteed 4.0's for doing absolutely nothing, and just holding a 4 year beer party for them. This is not a popularity contest here. This is about devising a system that is optimal FOR SOCIETY AT LARGE. What is the socially optimal way to determine who should get admitted to med-school, or to a particular impacted major? I think the optimally fairest way is to use a standard test. It's not perfectly fair, but it's a lot more fair than using grades. </p>

<p>
[quote]
You are right to an extent, but most people I know either do their homework, get help from the professor if they don't understand something (and thus learn it), get help from classmates (not copy, but get help)...

[/quote]
</p>

<p>There it is - you just said it right there. Whether it is straight-up copying or just getting help, at the end of the day, your homework scores are a function of how many friends you have and how smart they are. After all, the more smarter friends you have, the better help you will get. </p>

<p>
[quote]
But I mean less arbitrary in the sense that if you do the homework, you'll probably get most/all the points. You do the work, you learn the material, and you get the points. With tests, even if you know the material, you may not get a good grade. I'm sure you know how often students tend to make stupid mistakes on tests, or when a test just gives an arbitrary and obscure question (especially in weeders), and when there are 5-7 problems on a test, that really hurts. This seems to be especially true of math courses, in which errors in calculation/steps are all too common.

[/quote]
</p>

<p>Wrong, on several counts. First of all, weeder homeworks are notorious for being UNBELIEVABLY DIFFICULT to complete. I can name you plenty of technical classes (especially engineering classes) in which practically nobody gets even most of the homework points. Not even close. Even getting 75% of the homework points will put you in the very top tier in many of these classes. That's just a simple testament as to just how ridiculously difficult some of these homeworks are. </p>

<p>Consonant with that, most people in weeders will actually report that test questions are actually MUCH EASIER than the homeworks, not because the test questions are really that easy (they certainly are not), but because homework questions really are that difficult. There are entire engineering and science classes where just a single weekly problem set for 1 class can easily take somebody more than 10 hours to do. And that's just for 1 class. </p>

<p>
[quote]
Another quote from calkidd:

[/quote]
</p>

<p>Calkidd is wrong. Look, Calkidd is an MIT student in a special graduate program that combines some resources at Harvad and MIT. He is now attributing the easier grading he is getting at his Harvard classes to the project/hw-based classes he is taking now. But this is a false attribution. It is definitely true that graduate classes in general tend to rely less on tests. But this is a general truism of ALL graduate classes at any school. Plenty of Berkeley graduate classes are graded on projects and papers, with relatively few exams. Furthermore, the truth is, graduate level classes at all schools tend to be grade inflated (which incidentally gets back to something I said before, which is that one way to get high grades is to take lots of graduate-level classes, presuming that you can handle the higher workload). </p>

<p>So he has made an error in attribution. He has seen a bunch of Harvard graduate-level courses that are both grade inflated and project-based, and he has attributed the grade inflation to the project-based nature of the courses. He's wrong on 2 counts, #1, the fact is, Harvard grade inflation is universal, regardless of whether a class is project-based or not, and #2, like I said, grade inflation is endemic in grad-level courses in any school.</p>

<p>In fact, let's talk about another aspect of graduate classes. The reason why grad classes tend to be grade inflated is simple - grad students aren't really there to take classes. They are primarily there to do research and write their thesis. For example, if you're talking about PhD students, the life completely revolves around their thesis. Classes hardly matter. What matters is learning what you need to know to complete your thesis. You can get perfect grades in all of your classes, and still not be able to complete your thesis. Furthermore, your job placement is determined almost completely by the quality of your thesis (as well as things like your advisor rec's and other articles that you've published), and has almost nothing to do with your grades. In fact, many people say that any PhD student who actually lists his GPA on his C.V. must be doing so because his research is weak. {It's like reading a real estate listing and seeing a house that "has character" - it must mean that there is something seriously wrong with the house, if that's the best thing you can say is that it has character}. Hence, these students don't really care about their grades. They care about learning what they need to know for their research. Hence, in effect, their research thesis becomes their "exam".</p>

92

<p>
[quote]
What I do believe is that basing something this important on one entrance exam is a bad thing.

[/quote]
</p>

<p>Conversely, I believe that basing something this important on uneven criteria like grades is ALSO a bad thing. The major problem with grades is that grading standards vary from prof to prof and from school to school. One guy at one particular school may know nothing and do nothing, yet still get an A. Another guy at another school may work like a dog and be a genius at the subject, yet still get a C. No satisfactory method exists to account for the fact that some profs grade harder than others, and some schools grade harder than others. </p>

<p>Then you have the issue of students deliberately taking classes on things that they already know, just so they can get a bunch of easy A's. For example, I know people who were completely fluent in a particular foreign language. But they decided to take the intro courses for that language anyway. Why? To get a bunch of easy A's. They learned noting in those courses because they already knew everything. But that's not the point. The point was just to pump up their transcript with a a bunch of easy A's for very little work so that they could look good for grad school. But it's clearly a ridiculous situation. The purpose of a class is to teach you stuff you don't know, not just to give you a bunch of easy A's for things that you already know. </p>

<p>All of that stuff can be washed away with a proper entrance exam. You wouldn't have to worry about different grading standards by different profs, because your grades wouldn't matter anyway. You wouldn't have people cherry-picking classes on things that they already know because, again, your grades wouldn't matter, so there would be no incentive to take classes on things that you already know. All that would matter is what you actually know when the exam rolls around. </p>

<p>Besides, I would say that, if anything, Berkeley students would actually SUPPORT this idea. After all, Berkeley is a somewhat grade-deflated school relative to schools like HYPS. So it seems to me that if anything, Berkeley students would ACTUALLY GAIN from an entrance exam. They would be gaining relative to the grade-inflated schools. Heck, any grade deflated school would gain from such a measure.</p>

93

<p>
[quote]
Actually, it would probably go down to about zero. I can imagine a scenario where CS classes are graded the way that say, American Studies classes are now, in which practically everybody passes. In fact, I would argue that most humanities and social science classes at Berkeley today are ones in which practically everybody passes. They don't get all A's, obviously, but very few people actually don't pass.

[/quote]
</p>

<p>
[quote]
I agree that most likely, nothing will ever happen. But hey, that doesn't mean that you stop trying. It's one thing to try and fail. At least you tried. That's better than never having tried at all. After all, if you don't even try, you are GUARANTEED to fail.

[/quote]
</p>

<p>I doubt it would go down to about zero. You yourself wrote that the CS classes are only a little easier for upper-division classes. But why only a little easier? There is no need to weed in upper-divison classes. The thing is CS is simply a hard major. Whether or not weeders are there, CS classes will probably still be hard. Besides, a standard needs to be maintained. What kind of college is it that only hands out passing grades? What's the point of grading then? What's the point of having Ds and Fs if they are never handed out? It sounded like you are saying we should try to make majors with weeders more like the American Studies class, or the inflated classes at, say, Stanford. I don't think that's an enviable situation to have. I don't think massive grade inflation is what we should be aiming for here.</p>

<p>
[quote]
First off, I don't think it's true that prospective med students won't want everything to be based on one test. I am fairly certain that most people who have undergone the weeder premed classes would probably want admissions to be based purely on a single test. This would clearly include anybody who got a B or less, and I think it's fair to say that the majority of weeder grades in the premed courses are B's or less.

[/quote]
</p>

<p>But you are missing the reason why these students would prefer one test, which you yourself have brought up on many occasions, which is that for some reason Berkeley grades tougher yet students need higher GPAs to get into med school. That's why there is a disproportionate amount of people with "low grades" aka ~3.7 or less. Thus this tendency is not because the one entrance exam is preferable but because the current system is flawed.</p>

<p>
[quote]
Secondly, I still maintain that stress levels would go down from the present system, not because the using a single test is stress-free (it is in fact, high stress), but because the present system is ALSO high stress, in fact, probably higher stress than the alternative. You talk about people possibly studying for years in advance for the MCAT. Well, I already know people who are already studying for years in advance for their college OChem class. I know people who have already read the entire OChem book, and bought a bunch of those Schaum's practice workbooks and done all of the problems in them - all before they even officially take the OChem class. Heck, I even know some people who 'secretly' audit OChem by going to all of the lectures and taking all of the notes (but not actually registering for the class), in the SEMESTER BEFORE taking OChem officially.

[/quote]
</p>

<p>But is that higher stress or lower stress? Look, these students are either going to be studying in these four years for their classes, or for one test. They are still going to be spending massive amounts of time studying for either because there is still so much competition. In fact, I would think that basing it on one test would generate MORE competition. Think about it, if you don't beat the student sitting in front of you in one OChem test, it's not the end of the world. But when the admissions is based on one test, not beating the student in front of you could cost you admission. That's some serious stress. And when the stakes for one test are so high, students may not be able to afford to take a semester or two off.</p>

<p>
[quote]
This is about devising a system that is optimal FOR SOCIETY AT LARGE. What is the socially optimal way to determine who should get admitted to med-school, or to a particular impacted major? I think the optimally fairest way is to use a standard test. It's not perfectly fair, but it's a lot more fair than using grades.

[/quote]
</p>

<p>Yes, and it makes you wonder, how seriously will these students take their classes if admission is based on one test? (still using med school as an analogy) Obviously one test cannot cover nearly as much as four years' worth of tests. So what about all the things that are not on the test, but are still important, and mentioned in class? Will students study those? Do med schools really want these type of candidates? Do society really want these type of doctors, who only know how to study for one test?</p>

<p>Same thing applies for majors. An entrance exam can only last so long, maybe 3-4 hours tops. Can that really cover everything there is to know for a CS major to succeed in upper-division? Or do 2 years' worth of work give you a much better idea of the student's breadth of knowledge?</p>

<p>
[quote]
Wrong, on several counts. First of all, weeder homeworks are notorious for being UNBELIEVABLY DIFFICULT to complete. I can name you plenty of technical classes (especially engineering classes) in which practically nobody gets even most of the homework points. Not even close. Even getting 75% of the homework points will put you in the very top tier in many of these classes. That's just a simple testament as to just how ridiculously difficult some of these homeworks are.

[/quote]
</p>

<p>I can believe that for certain classes (notably engineering) but I don't think it's true for most classes at large. I don't think it's true for Math 1B or UGBA 10. I don't think it's true for non-weeders, in general. In most classes, you'll receive full or nearly full credit from homework. And it's not dependent on your friends. You could have no friends and go to your professor for help. So it depends on how much effort you put into it. You could go to a friend or a professor for help, and take time to learn the material, in which a higher grade seems fair to me, or you could not bother to do it, and receive a lower grade.</p>

94

<p>
[quote]
Calkidd is wrong. Look, Calkidd is an MIT student in a special graduate program that combines some resources at Harvad and MIT. He is now attributing the easier grading he is getting at his Harvard classes to the project/hw-based classes he is taking now. But this is a false attribution. It is definitely true that graduate classes in general tend to rely less on tests. But this is a general truism of ALL graduate classes at any school.

[/quote]
</p>

<p>But he's taken MIT graduate classes and he seems to think that MIT grad does not act this way. So why aren't MIT grad classes based more on project/homework than Harvard grad classes?</p>

<p>Besides, this seems like a logical explanation for why MIT undergrads have more difficulty getting good grades compared to Harvard undergrads. I suppose the test administered at MIT could just be much harder and much more arbitrary for some strange reason, but this seems less likely. At least, I think the former is one of many factors.</p>

<p>
[quote]
In fact, let's talk about another aspect of graduate classes. The reason why grad classes tend to be grade inflated is simple - grad students aren't really there to take classes. They are primarily there to do research and write their thesis. For example, if you're talking about PhD students, the life completely revolves around their thesis. Classes hardly matter. What matters is learning what you need to know to complete your thesis. You can get perfect grades in all of your classes, and still not be able to complete your thesis. Furthermore, your job placement is determined almost completely by the quality of your thesis (as well as things like your advisor rec's and other articles that you've published), and has almost nothing to do with your grades. In fact, many people say that any PhD student who actually lists his GPA on his C.V. must be doing so because his research is weak. {It's like reading a real estate listing and seeing a house that "has character" - it must mean that there is something seriously wrong with the house, if that's the best thing you can say is that it has character}. Hence, these students don't really care about their grades. They care about learning what they need to know for their research. Hence, in effect, their research thesis becomes their "exam".

[/quote]
</p>

<p>So essentially, these grad students are in the situation you propose. Do you think it's more or less stressful for them than for undergrads? Do you think they would rather have their PhD based on their classes or research, or only on the thesis? I know my math professor who almost got a PhD would certainly prefer the former, and I'd guess many PhD candidates who couldn't complete their thesis would agree.</p>

<p>
[quote]
Conversely, I believe that basing something this important on uneven criteria like grades is ALSO a bad thing. The major problem with grades is that grading standards vary from prof to prof and from school to school. One guy at one particular school may know nothing and do nothing, yet still get an A. Another guy at another school may work like a dog and be a genius at the subject, yet still get a C. No satisfactory method exists to account for the fact that some profs grade harder than others, and some schools grade harder than others.

[/quote]
</p>

<p>I strongly agree with this, and I think what should be done is to simply leverage the classes. No more easy classes where no one receives a failing grade, and no weeders with a large portion of the class failing. Instead of treating the symptons, let's get to the root of the problem. If this can be done then there shouldn't be a need for an entrance exam anyway, because those with higher grades really did work harder/are smarter/know the material better.</p>

95

<p>
[quote]
. No more easy classes where no one receives a failing grade, and no weeders with a large portion of the class failing. Instead of treating the symptons, let's get to the root of the problem. If this can be done then there shouldn't be a need for an entrance exam anyway, because those with higher grades really did work harder/are smarter/know the material better.

[/quote]
</p>

<p>What's so bad about a class where nodoby fails? What if everybody DESERVES to pass? Listen to this guy:</p>

<p>
[quote]
"I'm not willing to say that 17% of students should automatically receive A's and B's. But I am willing to say that 25% should get D's and F's if they earned them."

[/quote]
</p>

<p><a href="http://ls.berkeley.edu/undergrad/colloquia/04-11.html%5B/url%5D"&gt;http://ls.berkeley.edu/undergrad/colloquia/04-11.html&lt;/a&gt;&lt;/p>

<p>I'm with him. No, he and I probably disagree about who deserves what, but that's another story.</p>

96

<p>Sakky bashes everyone here so bad it isn't even funny. He's better reading than the Atlantic, and much less biased to boot.</p>

<p>Maybe you guys should study a real subject like physics ala Sakky and you would know how to at least respond with meaningful arguments.</p>

97

<p>
[quote]
Maybe you guys should study a real subject like physics ala Sakky and you would know how to at least respond with meaningful arguments.

[/quote]
</p>

<p>Sigh...:rolleyes:</p>

98

<p>
[quote]
I doubt it would go down to about zero. You yourself wrote that the CS classes are only a little easier for upper-division classes. But why only a little easier? There is no need to weed in upper-divison classes. The thing is CS is simply a hard major. Whether or not weeders are there, CS classes will probably still be hard.

[/quote]
</p>

<p>There is a difference between a class that is 'hard' and a class that fails people. I can tell you that MIT graduate-level engineering classes are extremely hard, in fact, almost ridiculously so. But that doesn't mean that a lot of people fail. In fact, usually nobody fails. It's extremely difficult to get an A, but practically nobody actually FAILS the classes. </p>

<p>Even the Berkeley upper division CS classes rarely actually FAIL anybody. Again, it's very difficult to get an A. But as long as you do the work, you're going to get a passing grade, even if it's a low passing grade like a C-. </p>

<p>To give you another example, medical school is extremely hard. I think there is nobody out there that will attempt to argue that med-school is easy. But at the same time, practically nobody actually flunks out of med-school. </p>

<p>Hence, I am fairly confident that you can make it such that the failure rate of CS lower division classes can go down to zero. </p>

<p>
[quote]
What kind of college is it that only hands out passing grades? What's the point of grading then? What's the point of having Ds and Fs if they are never handed out?

[/quote]
</p>

<p>Seems to me that that is exactly what HYPS has done, and nobody seems to have a problem with it. Hardly anybody ever flunks out of those schools, yet those schools are widely acknowledged to be elite schools. I have always stated that Berkeley should be more like HYPS in terms of the undergraduate program. </p>

<p>
[quote]
But you are missing the reason why these students would prefer one test, which you yourself have brought up on many occasions, which is that for some reason Berkeley grades tougher yet students need higher GPAs to get into med school. That's why there is a disproportionate amount of people with "low grades" aka ~3.7 or less. Thus this tendency is not because the one entrance exam is preferable but because the current system is flawed.

[/quote]
</p>

<p>First off, there are plenty of premeds who would give their right arm to get a 3.7. Plenty of premeds are not even within sight of having a 3.7. </p>

<p>But more importantly, you have to keep in mind why grading is so tough at Berkeley. It's because, frankly, there are a lot of not-so-good students at Berkeley, and Berkeley has to protect the reputation of the school by instituting tough grading. Hence, that's why there are weeders - basically, because certain departments have decided that they don't really want to confer degrees upon these mediocre students. I would love to simply get rid of these students. But I don't think that's political feasible anytime soon. Hence, if you have to live with impaction and weeding, then you should at least do it in a well-managed and fair manner. </p>

<p>But, hey, like I said, if you want to get rid of the bad students, that's an even better solution. Then you could raise the curve, get rid of the weeders, and get rid of impaction. But I don't see any feasible way to get rid of these students. If you can devise one, then I'm all for it. </p>

<p>
[quote]
But is that higher stress or lower stress? Look, these students are either going to be studying in these four years for their classes, or for one test.

[/quote]
</p>

<p>Why 4 years? The entrance exam is for entrance to the major, which most likely means that it will be taken in the beginning of your 2nd, or at latest your 3rd year. Hence, we're really only talking about 1 or 2 years worth of studying. </p>

<p>
[quote]
Think about it, if you don't beat the student sitting in front of you in one OChem test, it's not the end of the world. But when the admissions is based on one test, not beating the student in front of you could cost you admission. That's some serious stress. And when the stakes for one test are so high, students may not be able to afford to take a semester or two off.

[/quote]
</p>

<p>I don't see why it is so. Remember, we're just talking about entrance exam to a particular major. Not getting into a particular major is obviously not good, but it's not like it's the end of the world. Yo ucan transfer to a completely different university. Or you can just do a backup major. </p>

<p>My point is simply this. If we have to live with impaction, then we have to manage impaction fairly. Using an entrance exam is fairer than using grades to determine admission to an impacted major. Obviously, like I said, the best thing to do is simply get rid of impaction. But that would probably inevitably mean also getting rid of the bad students that compelled departments to implement impaction in the first place. </p>

<p>
[quote]
Yes, and it makes you wonder, how seriously will these students take their classes if admission is based on one test? (still using med school as an analogy) Obviously one test cannot cover nearly as much as four years' worth of tests. So what about all the things that are not on the test, but are still important, and mentioned in class? Will students study those? Do med schools really want these type of candidates? Do society really want these type of doctors, who only know how to study for one test?

[/quote]
</p>

<p>You are confusing your analogies here, in a number of ways. First off, an entrance exam is simply an exam that would allow you to get into a particular major, or in your analogy, to get into med-school. It doesn't mean that you will actually get placed as a doctor in the hospital you want, or in the specialty you want. It is like the USMLE exam. You can go through 4 years of med-school, then fail the USMLE and thus not get your license to practice medicine. But simply passing the USMLE doesn't mean that you now get to work in the hospital you want, in the residency that you want. You still have to compete with others to get the best residencies. Passing the USMLE is a necessary but not sufficient condition. </p>

<p>Similarly, passing an entrance exam into the major simply means that you now get to enter the major. It doesn't mean that you will get the job you want when you graduate. All it means is that you are allowed into the major, nothing more, nothing less. Plenty of students from any school, even Harvard, finish their degree yet don't get the job that they really want. Or don't get into the graduate school that they really want. Heck, even completing your PhD dissertation and thus getting your PhD doesn't mean that you will get the academic placement that you want. I 've read about plenty of PhD's who wanted an academic position yet couldn't get placed anywhere and thus were forced to enter industry, which they didn't really want. </p>

<p>All the entrance exam is is a minimum hurdle you have to traverse in order to simply get into the major. But it has to be a fair hurdle that is applied to everybody in a fair manner. It doesn't mean that you can simply pass this hurdle and you now get everything that you want. In particular, companies are still going to interview you for jobs just like they do now, and you may still not get picked. I am not forcing the market to accept people that the market does not want to accept. I am simply proposing a fair process for the way that people can present themselves to the market in the first place. The market still reserves the right to refuse to transact. </p>

<p>
[quote]
I can believe that for certain classes (notably engineering) but I don't think it's true for most classes at large. I don't think it's true for Math 1B or UGBA 10. I don't think it's true for non-weeders, in general. In most classes, you'll receive full or nearly full credit from homework. And it's not dependent on your friends. You could have no friends and go to your professor for help. So it depends on how much effort you put into it. You could go to a friend or a professor for help, and take time to learn the material, in which a higher grade seems fair to me, or you could not bother to do it, and receive a lower grade.

[/quote]
</p>

<p>But even this is not a serious objection. First off, your entire premise is off. It doesn't matter if homework questions are actually easier or harder, because the curve wipes away all aspects of difficulty. I used to actually like extremely difficult homeworks. Why? Because if they were difficult to me, they are probably difficult for everybody else too. Hence, the 'homework curve' will be low. If the homeworks are easy, then that means that high homework scores no longer become a differentiating factor on the curve. </p>

<p>Heck, some of the most brutal weeder classes I took were ones where the homeworks were easy. The questions were easy, so you didn't dare to miss even one single point on the homework because you knew that everybody else was going to do well on the homework. </p>

<p>The notion of going to your prof for help depends on the helpfulness of your prof. Believe me, there are PLENTY of profs out there that are the complete opposite of helpfulness. You go to them for help, and they will berate you for not knowing it yourself, and then they might even go so far as to suggest that you drop the class. Believe me, it happens. That's where you get into the situation where you end up relying on your friends, because it's quite clear that the prof doesn't want to help. In fact, there are some profs that are annoyed at the mere fact of having to interact with undergrads at all, as they are seen as a drain on their research time.</p>

99

<p>
[quote]
But he's taken MIT graduate classes and he seems to think that MIT grad does not act this way. So why aren't MIT grad classes based more on project/homework than Harvard grad classes?

[/quote]
</p>

<p>They are. In fact, that actually points to another weakness in Calkidd's argument. He seems to think that MIT grad classes are less based on projects/hw. Well, I don't know what classes he's taken, but that's not certainly my experience or the experience of anybody that I know. MIT grad-level classes are just as hw/project-based as Harvard grad classes. I don't see any difference. There are entire swaths of MIT graduate classes that have no exams. </p>

<p>
[quote]
So essentially, these grad students are in the situation you propose. Do you think it's more or less stressful for them than for undergrads? Do you think they would rather have their PhD based on their classes or research, or only on the thesis? I know my math professor who almost got a PhD would certainly prefer the former, and I'd guess many PhD candidates who couldn't complete their thesis would agree.

[/quote]
</p>

<p>Again, obviously it's true that anybody who can't traverse a certain hurdle is going to wish that the hurdle didn't exist. And of course a PhD is more stressful than undergrad is. That's why a PhD is considered to be a higher degree. It's SUPPOSED to be more stressful. That's why you never hear of PhD students who are lollygagging around, being extremely lazy, who would rather spend all their time drinking and playing video games and never doing anything at all. But you do see some undergrads doing that. Don't believe me? Go down to where the frathouses are. There are some students there that literally haven't been to class in weeks. </p>

<p>The entire reasoning behind the PhD is that you are a mini-scholar and thus you have to complete an acceptable piece of academic work called the thesis. That is what the PhD is all about. If you don't want to do that, fine, don't do it. Nobody is forcing anybody to get a PhD. Nobody 'needs' a PhD. </p>

<p>However, the PhD still has to be a fair process. You can't just have a rule where some people can get a PhD without doing the thesis, and other people can get it without doing the thesis. Everybody has to do the thesis. That's what makes the process fair. Standard rules apply to everybody. Nobody gets special treatment.</p>

<p>So let's bring this back to what we are talking about here. I am talking about using an entrance exam for impacted/weeded majors. Nobody "needs" to enter an impacted/weeded major at Berkeley. You can major in something unimpacted/unweeded. You can simply choose to go to a completely different school. You don't 'need' to enter Berkeley's impacted/weeded major. But if impaction/weeding has to be utilized, then it has to be utilized fairly. We shouldn't have special treatment for certain people. </p>

<p>As it stands now, transfer students are getting special treatment because they get to skip over some or all weeders. That's not fair. Granted, the transfer students didn't ask for special treatment. But they are getting it anyway. It is that unfairness that I want to stop. If you want to use impaction, if you want to use weeding for a certain major, then it has to be applied fairly to all people in that major. </p>

<p>Otherwise, like I said, don't use impaction, and don't use weeding. I have always stated that that is the optimal solution, if it is feasible. </p>

<p>But the worst solution is the current one, which is to do weeding on SOME people, but not all. That is legitimately unfair. </p>

<p>
[quote]
I strongly agree with this, and I think what should be done is to simply leverage the classes. No more easy classes where no one receives a failing grade, and no weeders with a large portion of the class failing. Instead of treating the symptons, let's get to the root of the problem. If this can be done then there shouldn't be a need for an entrance exam anyway, because those with higher grades really did work harder/are smarter/know the material better.

[/quote]
</p>

<p>Uh, no, the real root of the problem is that there is a long tail end of bad students. Getting rid of them would REALLY be getting rid of the root of the problem. MIT, for example, doesn't have any impacted majors, and doesn't even do weeding to the level that Berkeley does. Why? Because the students are good. You don't have a conspicuous tail end of bad students at MIT the way you do at Berkeley. If MIT did, then MIT would probably implement impaction and Berkeley-style weeding. </p>

<p>But like I said, sadly, I don't see any feasible way to get rid of the bad students. Hence, we will most likely have to live with impaction and weeding. The question then is how to manage it fairly.</p>

100

<p>
[quote]
I've also noticed this. Most of my classes are based on 2 midterms and 1 test, while my friend who is currently attending Harvard has all these problem sets and projects (projects? Don't see too many of those at Berkeley).

[/quote]
</p>

<p>Some of the easiest classes that I have are ones that had just exams. Conversely, some of the most ridiculously difficult classes ever are ones that have had numerous problem sets and projects. How about ChemE140? That class is chock full of problem sets and also has a project. It iis also easily one of the most brutal weeders on campus. How about CS 61B? That class has an endless gauntlet of problem sets and labs, and projects. Let's see, 10 problem sets. 14 labs. 3 projects. Yet that class is a notorious weeder, probably the harshest weeder in all of CS. EECS150 also has numerous homeworks, labs, and an big final project. I don't think you will find a single person who will say that EECS150 has easy grading.</p>

<p>Hence, I would question calkidd's very basic assumptions that a class that has lots of problem sets and projects are easier. </p>

<p><a href="http://www.cs.berkeley.edu/%7Ejrs/61b/%5B/url%5D"&gt;http://www.cs.berkeley.edu/~jrs/61b/&lt;/a>
<a href="http://inst.eecs.berkeley.edu/%7Ecs150/fa06/Syllabus.php%5B/url%5D"&gt;http://inst.eecs.berkeley.edu/~cs150/fa06/Syllabus.php&lt;/a&gt;&lt;/p>