10 ms·

The way grades work these days (most recent source: a friend who graduated public high school in Palo Alto in 2014), a large fraction of the grade is doing home

The way grades work these days (most recent source: a friend who graduated public high school in Palo Alto in 2014), a large fraction of the grade is doing homework. Lots of homework. Enough that the friend I'm thinking of was often awake into the a.m. doing it. It's not that he had to spend that much time to learn the subject for the test; he probably could have dropped half or more of his homework and still done fine on the tests. The trouble is, if he did that, even if he got perfect scores on the tests, he would have gotten Cs or worse in the overall grade, because of the portion that is homework. This friend decided he wanted to play the game and max out his grades because he wanted official recognition (he did get into Harvard later). Another friend, probably equally intelligent, did drop a bunch of homework, did get Cs, and had a rough time with college admissions (through a string of luck and outside intervention involving setting up a meeting with a Fields Medalist professor at Berkeley who then advocated for him, he managed to get into Berkeley). Both friends are now doing math PhDs. I, myself, also dropped a bunch of homework, also got Cs, and left high school after 10th grade, and never went to college. I subsequently studied programming on my own, and have been employed doing that for nearly four years.

WalterBright 8y ago
Caltech also had an institute policy that if you aced the final, you got an A, even if you never did any of the homework. There were students who did that (google Hal Finney), but alas I did not have that kind of wattage upstairs. I could not do well on the finals unless I mastered every last homework problem, which often involved getting help from my ever-indulgent roommate (thank you, Mark!).
- waterhouse 8y ago
  That would be an improvement. I think, more generally, students should be able to choose to make their grade depend entirely on their test scores. (Knowing all the material on a test is much different from acing it; you have to be thoroughly checking your work multiple times to reliably achieve perfect scores. I've taken a decent number of math contents where the winners were basically determined by "who makes the fewest careless mistakes?". It is good for a person to be able to do that, but by that point it's measuring a different skill. Having a higher ceiling is much better—but teachers like to have the grade be "number correct / number of problems".) If you didn't answer the 10 homework problems designed to make you practice the law of cosines, but you correctly applied the law of cosines on a quiz and a later test, how is it justifiable to take points away? It seems to me that a student's grade should be the best available estimate of their ability in, knowledge of, and perhaps interest in, a subject. If a teacher knowingly gives a C to a student in the above type of situation, that seems to me like lying. And if the tests aren't enough, in total, to make a decent stab at measuring all relevant aspects of the subject matter being taught, then that seems to be a deficiency in the tests. Maybe some subjects need take-home projects because a timed test simply isn't a good format to demonstrate some skills, but grading down for the lack of regular drill-type homework seems indefensible to me. (I wouldn't make all the students switch over to being graded solely on tests, because some probably do like the current system, and because there are reasonable arguments that a high-pressure timed test is especially bad for some students and isn't a good reflection of how well they'd do real work. Similar arguments may be made about coding interviews.) Another, possibly more important, reason to oppose required grading of drill-type homework: It basically forces everyone to follow the same study program. (Well, they can study more on top of it, but they can't follow any study program that doesn't include doing all of those problems.) Given the range of human variation, this single study program is going to be suboptimal for lots of people. That's a problem in itself, but moreover, it gives students no incentive to try out different studying approaches and learn what is best for them. Then, when they go to a more unsupervised environment, where the problems are hard and studying really is important, they may be totally unprepared. Some may rapidly seize upon decent studying methods, others may just fail miserably. (I've heard this is a classic failure mode of smart kids who go to college.)
  - WalterBright 8y ago
    How can you divine that a person has mastered the material when they fail every test you can concoct? > careless mistakes You haven't really learned the material if you keep making careless mistakes. I wouldn't want to hire a person who was so unreliable. I wouldn't want to fly on an airplane with a accomplished pilot who kept making careless mistakes. I wouldn't want the best surgeon in the world to work on me if he was careless. Would you?
    - waterhouse 8y ago
      > You haven't really learned the material if you keep making careless mistakes. In case it's not clear what "careless mistakes" means, this would be something like taking a physics problem where two rigid balls of a given mass and given initial velocity and momentum collide, and correctly carrying out the steps to determine their final velocities, except you forgot to carry a 1 or flipped a sign somewhere along the process. Given that the student showed his work for each step of the process, I think this can be clearly distinguished from, say, a student who didn't know how to apply the laws of conservation of momentum and energy. I'd agree that it makes sense to take a point off, because there is a correlation between mastery and a lower error rate, but I don't think it makes sense to say "well, you're not perfect, therefore we'll take 20% off your grade for not doing homework". If you took a bunch of physics university professors and gave them a quiz with 10 problems of roughly that level, how many of them do you think would get perfect scores? Er, I assumed that the phrase "aced the final" meant "got a perfect score". Was I wrong? If it's more like "got 90+%", then forget everything I said about it. > How can you divine that a person has mastered the material when they fail every test you can concoct? Er... you decide that they probably haven't mastered the material? I haven't talked about anyone who fails tests, just those who make a few math mistakes and don't do most of the homework. > I wouldn't want to hire a person who was so unreliable. I think there's a lot of time in between high school and becoming a pilot or a surgeon. Once you're pretty close to knowing your career path, it certainly makes sense to practice a lot and lower your error rate a great deal—and if, for some reason, you can't do that, then it would make sense for people to refuse to employ you in that field. However, when students are taking classes in 4 completely separate subjects, and there is no way any career would cover more than 2 of them, then intense practicing is very premature—most of the work would be wasted. Also, by the way, pilots and surgeons are two examples of professions where your work is direct and has immediate effect, and mistakes can be catastrophic and hard to fix. In fact, it's hard to think of professions that are more extreme in that regard (drivers and other operators of heavy machinery—but airplanes are some of the biggest and most dangerous machinery out there). There are plenty of professions where all your work can be looked over by a peer before it goes into production / otherwise has a real effect. That is a cost, but if you're good enough in other ways, you may more than compensate for it (and I think it is common for sufficiently large tech companies to require peer review for all changes before they go into the master branch, let alone production).[1] So I would say it is inappropriate to apply the standards of pilots and surgeons to all students of all subjects. [1] Come to think of it, I know two programmers who have some form of dyslexia; one is my coworker and his pull requests are full of typos, the other is a friend and keeps mixing up words. Both of them also spent lots of time studying CS on their own—one focusing on data structures, the other on programming languages—and are quite valuable engineers. They're also fans of programming languages that do advanced compile-time type-checking.

WalterBright 8y ago

Caltech also had an institute policy that if you aced the final, you got an A, even if you never did any of the homework. There were students who did that (google Hal Finney), but alas I did not have that kind of wattage upstairs. I could not do well on the finals unless I mastered every last homework problem, which often involved getting help from my ever-indulgent roommate (thank you, Mark!).

waterhouse 8y ago
That would be an improvement. I think, more generally, students should be able to choose to make their grade depend entirely on their test scores. (Knowing all the material on a test is much different from acing it; you have to be thoroughly checking your work multiple times to reliably achieve perfect scores. I've taken a decent number of math contents where the winners were basically determined by "who makes the fewest careless mistakes?". It is good for a person to be able to do that, but by that point it's measuring a different skill. Having a higher ceiling is much better—but teachers like to have the grade be "number correct / number of problems".) If you didn't answer the 10 homework problems designed to make you practice the law of cosines, but you correctly applied the law of cosines on a quiz and a later test, how is it justifiable to take points away? It seems to me that a student's grade should be the best available estimate of their ability in, knowledge of, and perhaps interest in, a subject. If a teacher knowingly gives a C to a student in the above type of situation, that seems to me like lying. And if the tests aren't enough, in total, to make a decent stab at measuring all relevant aspects of the subject matter being taught, then that seems to be a deficiency in the tests. Maybe some subjects need take-home projects because a timed test simply isn't a good format to demonstrate some skills, but grading down for the lack of regular drill-type homework seems indefensible to me. (I wouldn't make all the students switch over to being graded solely on tests, because some probably do like the current system, and because there are reasonable arguments that a high-pressure timed test is especially bad for some students and isn't a good reflection of how well they'd do real work. Similar arguments may be made about coding interviews.) Another, possibly more important, reason to oppose required grading of drill-type homework: It basically forces everyone to follow the same study program. (Well, they can study more on top of it, but they can't follow any study program that doesn't include doing all of those problems.) Given the range of human variation, this single study program is going to be suboptimal for lots of people. That's a problem in itself, but moreover, it gives students no incentive to try out different studying approaches and learn what is best for them. Then, when they go to a more unsupervised environment, where the problems are hard and studying really is important, they may be totally unprepared. Some may rapidly seize upon decent studying methods, others may just fail miserably. (I've heard this is a classic failure mode of smart kids who go to college.)
- WalterBright 8y ago
  How can you divine that a person has mastered the material when they fail every test you can concoct? > careless mistakes You haven't really learned the material if you keep making careless mistakes. I wouldn't want to hire a person who was so unreliable. I wouldn't want to fly on an airplane with a accomplished pilot who kept making careless mistakes. I wouldn't want the best surgeon in the world to work on me if he was careless. Would you?
  - waterhouse 8y ago
    > You haven't really learned the material if you keep making careless mistakes. In case it's not clear what "careless mistakes" means, this would be something like taking a physics problem where two rigid balls of a given mass and given initial velocity and momentum collide, and correctly carrying out the steps to determine their final velocities, except you forgot to carry a 1 or flipped a sign somewhere along the process. Given that the student showed his work for each step of the process, I think this can be clearly distinguished from, say, a student who didn't know how to apply the laws of conservation of momentum and energy. I'd agree that it makes sense to take a point off, because there is a correlation between mastery and a lower error rate, but I don't think it makes sense to say "well, you're not perfect, therefore we'll take 20% off your grade for not doing homework". If you took a bunch of physics university professors and gave them a quiz with 10 problems of roughly that level, how many of them do you think would get perfect scores? Er, I assumed that the phrase "aced the final" meant "got a perfect score". Was I wrong? If it's more like "got 90+%", then forget everything I said about it. > How can you divine that a person has mastered the material when they fail every test you can concoct? Er... you decide that they probably haven't mastered the material? I haven't talked about anyone who fails tests, just those who make a few math mistakes and don't do most of the homework. > I wouldn't want to hire a person who was so unreliable. I think there's a lot of time in between high school and becoming a pilot or a surgeon. Once you're pretty close to knowing your career path, it certainly makes sense to practice a lot and lower your error rate a great deal—and if, for some reason, you can't do that, then it would make sense for people to refuse to employ you in that field. However, when students are taking classes in 4 completely separate subjects, and there is no way any career would cover more than 2 of them, then intense practicing is very premature—most of the work would be wasted. Also, by the way, pilots and surgeons are two examples of professions where your work is direct and has immediate effect, and mistakes can be catastrophic and hard to fix. In fact, it's hard to think of professions that are more extreme in that regard (drivers and other operators of heavy machinery—but airplanes are some of the biggest and most dangerous machinery out there). There are plenty of professions where all your work can be looked over by a peer before it goes into production / otherwise has a real effect. That is a cost, but if you're good enough in other ways, you may more than compensate for it (and I think it is common for sufficiently large tech companies to require peer review for all changes before they go into the master branch, let alone production).[1] So I would say it is inappropriate to apply the standards of pilots and surgeons to all students of all subjects. [1] Come to think of it, I know two programmers who have some form of dyslexia; one is my coworker and his pull requests are full of typos, the other is a friend and keeps mixing up words. Both of them also spent lots of time studying CS on their own—one focusing on data structures, the other on programming languages—and are quite valuable engineers. They're also fans of programming languages that do advanced compile-time type-checking.