9 ms·

One of the ways in which most grading systems are fucked up is that there's more downside than upside. One failure cancels out 3 or 4 successes, no matter how g

One of the ways in which most grading systems are fucked up is that there's more downside than upside. One failure cancels out 3 or 4 successes, no matter how great. If a teacher decides that any late work gets a zero, then being half a day late on one paper nullifies several great, on-time papers. I think it's a vicious cycle. This sort of system is designed to produce the sort of risk-averse rule-followers required by old-style industry: people destined for jobs where creativity isn't required but following orders 90% of the time is failure. But for that system's defense, it's good practice for real life. Humans are vindictive and focused more on stigma, gossip and the aversion to failure than in seeking out excellence and creativity. Hence, they create societies (like our society at a macro level, like most corporations, and like many school systems) that punish failure brutally but don't really value (or even fully trust) excellence, and which actively discourage the risk-taking usually required to excel. (And no, our society's outsized reward for economic fortune is not "rewarding success" since most of those people at the top are effective social climbers or hereditarily-placed parasites.) The "vicious cycle" is closed if one believes that the vindictive, punitive nature of people comes from how they are taught. Or is it innate? I'm really not sure. The way I think grading should be done is an inverse of the current system. Make the problems really, really hard. Some should be so hard that the hardest exam problems go for years without being answered. Make it so hard that 20% is passing and 60% is exceptional-- instead of 70% being passing and 95% being exceptional-- and so that students are used to it (and therefore they don't panic when they take an exam and can't solve half the problems). I argue for this because the reality for the most creative, hardest-to-get-right pursuits is that for 60% of one's efforts to succeed is exceptional. All that said, I think grading is important, but it shouldn't be started at such a young age, and that grades should generally be internal (for the student's and school's benefit) only. Many business schools actually prohibit their students from disclosing GPA or transcripts, only verifying which courses were taken and passed. If I were running a college, full transcripts would be available only to academic graduate departments-- not to future employers or professional schools. I'm not going to do their job (sorting) for them.

zitterbewegung 15y ago
60% of your efforts to succeed are successful? I have gotten published in poetry journals and the real ratio is more like 90% of what you do is meaningless and 10% is publishable. See http://www.youtube.com/watch?v=BI23U7U2aUY&feature=player_embedded http://www.youtube.com/watch?v=BI23U7U2aUY&feature=playe...
notahacker 15y ago
I'm not sure that the majority of students will "get used to" a grading system where given their aptitude they will be unlikely to score more than 30% - particularly when their reward for thorough preparation is likely to be a mere couple of percentage points on top of a dismal looking score. If anything is likely to condition the median person into scraping the minimum necessary requirement it's drawing attention to how far ahead their smarter peers are. Smarter peers who have more outlets for demonstrating their aptitude other than dwarfing the average test scores, and who are probably smart enough to realise their test scores are a poor proxy for real life success anyway. Skewing the scoring still doesn't make exams resemble real life; one doesn't analyse their successes in percentage-points-from-perfection. If we did, most people would have a serious inferiority complex. Interviewers who screen out unexceptional candidates are normal; those who waste insufficiently exceptional candidates time repeatedly proving they're not creative or excellent enough for the job might even be considered sadistic. There's a difference between failing 90% of interviews and staring blankly at the interviewer as I fail to answer 90% of the questions, knowing I'll have to repeat the experience until someone appreciates the 10% I get right. Likewise compensation packages expressed as a percentage of most productive employees compensation won't garner much interest, even though incentive structures within conventional companies is usually skewed heavily in the favour of less productive employees. If I were running a college, I'd probably accept the reality that job sorting is one of the main reasons my students value the course...
pnathan 15y ago
As a former TA... there needs to be a clear distinction in discussions of students. Not everyone is good. Some students are 'bad'. Incompetent. Lousy. Shouldn't be there. Many students are just banging out what they need for that degree. Why are they wanting the degree? Money. Stability. Cause dad said so. Cause it's what ya do. Do they care? No. And then there are the bright students/hackers. They were so interested - it was so very fun working with them. I reckon the breakdown was 10/80/10 or thereabouts. Good teachers work for all those students. There are different strategies to try to get/maintain interest/excellence. Some try fear, others lots of homework, some general disinterest (sink or swim). (Un)fortunately, we want to have ratings on people who we hire. "Is this guy any good? Yeah? How good?" So... like any project, we have a metric, and like humans we are optimizing for the metric instead of success. This last hundred years, the metric is GPA. And, for what its worth, my recollection of the grades we gave is that it was a rough correlation with the student's ability. Anyway, my 2c.
pbiggar 15y ago
So you might be interested to know that universities in Ireland (and I think England, where we inherited our university system from), everyone aims for a 70% grade. That's a "1st class honours" degree, and is basically the best you can come out with. Below that is "2nd class honours, I" (colloquially 2.1, pronounced "two one")). With a 1st or a 2.1, you're considered a high achiever (like a good job, or grad school, might require a 2.1 or better). A 2.1 is between 60-70%. Pass is 40%, and exams are marked harder to make the numbers work. (I believe harder questions are asked than equivalent US universities, but that's just rumour, unless you really do get multiple choice questions in university). It leads to interesting dynamics. In the Arts and Humanities, students are basically marked out of 80, with scores above that reserved for truly great work which is seen relatively rarely. In CS or Mathematics, where you truly can answer perfectly, sometimes you can get close to 100% in a exam, which throws off the marking scheme. For project work though, 70% means "great job", and 80% means you've done something truly excellent. For my undergraduate thesis, a piece of novel research on sorting and branch predictors that I truly poured my heart and soul into for 6 months, I got 94%. The feeling at achieving that grade was truly exhilarating, in a way that it could never be if I people didn't normally get 72% for great work.
- joelhaasnoot 15y ago
  Having attended an American high school but being Dutch, I know that this whole 100% thing is a very american, cultural thing. Here in The Netherlands we've decided as a society that you should aim high, but our grades are out of 10, and only top students actually get 10/10. The American system is built around the fact that if you do everything as required, you get an A+. Here, we reserve 10/10 for truely exceptional , over the top and amazing work. As such, exams are written to have harder questions...
- noonespecial 15y ago
  The American grading system has a severe lack of dynamic headroom. The measurement system saturates at 100% for even mediocre effort and there is no way to measure achievement beyond that. I once argued in a paper at university that we needed a logarithmic grading system like decibels. I got a C.
Perceval 15y ago
there's more downside than upside. One failure cancels out 3 or 4 successes Some should be so hard that the hardest exam problems go for years without being answered. Make it so hard that 20% is passing and 60% is exceptional-- instead of 70% being passing and 95% being exceptional I think what you're getting at here is the question of scale. The students in the linked story quit upon reaching 100%. But obviously they had not learned 100% of the words in English, they had only learned 100% of the words on a very short vocabulary list. One way to keep their motivation might be to scale their accomplishments relative to the total number of words. There are somewhere between 600,000 and 1,000,000 words in English (http://en.wikipedia.org/wiki/English_language#Number_of_words_in_English http://en.wikipedia.org/wiki/English_language#Number_of_word...). The teacher might keep track of the total number of words the students have learned. Basic English is about 850 words, and newspaper literacy is somewhere between 4000–10,000 words. Regardless of how many they learn in class, they will always know how many more there are to learn. You ideas about higher downsides to grading and scaling the 'exceptional' point to a shift from averaging grades based on mastering course content to cumulative grades based on mastering field content. Does that sound about right?

zitterbewegung 15y ago

60% of your efforts to succeed are successful? I have gotten published in poetry journals and the real ratio is more like 90% of what you do is meaningless and 10% is publishable. See http://www.youtube.com/watch?v=BI23U7U2aUY&feature=player_embedded http://www.youtube.com/watch?v=BI23U7U2aUY&feature=playe...

notahacker 15y ago

I'm not sure that the majority of students will "get used to" a grading system where given their aptitude they will be unlikely to score more than 30% - particularly when their reward for thorough preparation is likely to be a mere couple of percentage points on top of a dismal looking score. If anything is likely to condition the median person into scraping the minimum necessary requirement it's drawing attention to how far ahead their smarter peers are. Smarter peers who have more outlets for demonstrating their aptitude other than dwarfing the average test scores, and who are probably smart enough to realise their test scores are a poor proxy for real life success anyway. Skewing the scoring still doesn't make exams resemble real life; one doesn't analyse their successes in percentage-points-from-perfection. If we did, most people would have a serious inferiority complex. Interviewers who screen out unexceptional candidates are normal; those who waste insufficiently exceptional candidates time repeatedly proving they're not creative or excellent enough for the job might even be considered sadistic. There's a difference between failing 90% of interviews and staring blankly at the interviewer as I fail to answer 90% of the questions, knowing I'll have to repeat the experience until someone appreciates the 10% I get right. Likewise compensation packages expressed as a percentage of most productive employees compensation won't garner much interest, even though incentive structures within conventional companies is usually skewed heavily in the favour of less productive employees. If I were running a college, I'd probably accept the reality that job sorting is one of the main reasons my students value the course...

pnathan 15y ago

As a former TA... there needs to be a clear distinction in discussions of students. Not everyone is good. Some students are 'bad'. Incompetent. Lousy. Shouldn't be there. Many students are just banging out what they need for that degree. Why are they wanting the degree? Money. Stability. Cause dad said so. Cause it's what ya do. Do they care? No. And then there are the bright students/hackers. They were so interested - it was so very fun working with them. I reckon the breakdown was 10/80/10 or thereabouts. Good teachers work for all those students. There are different strategies to try to get/maintain interest/excellence. Some try fear, others lots of homework, some general disinterest (sink or swim). (Un)fortunately, we want to have ratings on people who we hire. "Is this guy any good? Yeah? How good?" So... like any project, we have a metric, and like humans we are optimizing for the metric instead of success. This last hundred years, the metric is GPA. And, for what its worth, my recollection of the grades we gave is that it was a rough correlation with the student's ability. Anyway, my 2c.

pbiggar 15y ago

So you might be interested to know that universities in Ireland (and I think England, where we inherited our university system from), everyone aims for a 70% grade. That's a "1st class honours" degree, and is basically the best you can come out with. Below that is "2nd class honours, I" (colloquially 2.1, pronounced "two one")). With a 1st or a 2.1, you're considered a high achiever (like a good job, or grad school, might require a 2.1 or better). A 2.1 is between 60-70%. Pass is 40%, and exams are marked harder to make the numbers work. (I believe harder questions are asked than equivalent US universities, but that's just rumour, unless you really do get multiple choice questions in university). It leads to interesting dynamics. In the Arts and Humanities, students are basically marked out of 80, with scores above that reserved for truly great work which is seen relatively rarely. In CS or Mathematics, where you truly can answer perfectly, sometimes you can get close to 100% in a exam, which throws off the marking scheme. For project work though, 70% means "great job", and 80% means you've done something truly excellent. For my undergraduate thesis, a piece of novel research on sorting and branch predictors that I truly poured my heart and soul into for 6 months, I got 94%. The feeling at achieving that grade was truly exhilarating, in a way that it could never be if I people didn't normally get 72% for great work.

joelhaasnoot 15y ago
Having attended an American high school but being Dutch, I know that this whole 100% thing is a very american, cultural thing. Here in The Netherlands we've decided as a society that you should aim high, but our grades are out of 10, and only top students actually get 10/10. The American system is built around the fact that if you do everything as required, you get an A+. Here, we reserve 10/10 for truely exceptional , over the top and amazing work. As such, exams are written to have harder questions...
noonespecial 15y ago
The American grading system has a severe lack of dynamic headroom. The measurement system saturates at 100% for even mediocre effort and there is no way to measure achievement beyond that. I once argued in a paper at university that we needed a logarithmic grading system like decibels. I got a C.

Perceval 15y ago

there's more downside than upside. One failure cancels out 3 or 4 successes Some should be so hard that the hardest exam problems go for years without being answered. Make it so hard that 20% is passing and 60% is exceptional-- instead of 70% being passing and 95% being exceptional I think what you're getting at here is the question of scale. The students in the linked story quit upon reaching 100%. But obviously they had not learned 100% of the words in English, they had only learned 100% of the words on a very short vocabulary list. One way to keep their motivation might be to scale their accomplishments relative to the total number of words. There are somewhere between 600,000 and 1,000,000 words in English (http://en.wikipedia.org/wiki/English_language#Number_of_words_in_English http://en.wikipedia.org/wiki/English_language#Number_of_word...). The teacher might keep track of the total number of words the students have learned. Basic English is about 850 words, and newspaper literacy is somewhere between 4000–10,000 words. Regardless of how many they learn in class, they will always know how many more there are to learn. You ideas about higher downsides to grading and scaling the 'exceptional' point to a shift from averaging grades based on mastering course content to cumulative grades based on mastering field content. Does that sound about right?