Something interesting about grades, the Crescendo Education Group, the main consulting group that pushes for and trains teachers on equitable grading, claims that one of the signs that equitable grading is more fair is that it correlates more strongly to testing scores. (See https://crescendoedgroup.org/results/) It seems they're doing something weird to get these correlations, but I thought it was interesting.
It's sort of funny to push for grades that are more like test scores...just use test scores! Though I suppose if you want to make grades that capture long-term information about a student's performance and eliminate things like effort...I can see the appeal of that.
I have always been opposed to standardized tests because they are particularly difficult for those with any type of learning difference. Especially those with dyslexia (which is approximately 1/5 the population regardless of language or culture). Wealthy LD children have a significant advantage because their disabilities are often caught early and they are given appropriate tools for learning.
Something to motivate any of your students who might be struggling with standardized tests: your cousin (who is a professor of Mathematics at an ivy league school) scored in the low 500s on his math SAT!!!
It is interesting Hattie included a study in his #1 influence, "Self Report Grades" in Visible Learning (VL). However, the study was on the reliability of students honestly reporting their grade to College Admissions. The study found a high correlation, but that students overstated their result (who would have guessed?). The high correlation when converted to an effect size yielded one of the largest ES in VL of 3.1. Hattie strangely interpreted this as students self reporting has one of the highest influences on that students achievement. https://www.researchgate.net/publication/240723602_The_Validity_of_Self-Reported_Grade_Point_Averages_Class_Ranks_and_Test_Scores_A_Meta-Analysis_and_Review_of_the_Literature
I've long been aware of the correlation between grades in high school and grades in college. But I don't have the same takeaway.
Kids who care about grades are markedly different from kids who don't. Kids in each group range from not too bright to brilliant. When they go off the college, the kids who care about grades will still care about grades. If they can't get good grades in one subject, they will change majors until they do. Kids who don't care about grades will not make these choices. It doesn't surprise me that kids who don't care about grades might be more likely to quit school.
What this research *doesn't* prove, I think, is that kids with low SAT scores and high grades are smarter than kids with high SAT scores and low grades.
I've never argued that SATs are fairer than grades. What they are, within broad categories, is more accurate about the tester's intelligence. And what we really should want to know about our poorer students is which ones are bright.
You can choose where you take the SAT (so a place with A/C), as well as retake it to make up for a bad day. Also, the things that make HSGPA subjective also make college GPA subjective, which explains why the former is so predictive of the latter and this affirms Bryan Caplan's education-as-signalling thesis. Education needs to be decoupled from employment and there needs to be greater liquidity in labor markets if social mobility is a concern.
1. I think to ask a kid to optimize their SAT-testing location is asking a lot. Also retakes cost money, can be demoralizing, etc. I'm not saying it can't be done -- but it impacts the fairness.
2. I don't think the dual subjectivity makes sense in the way you're saying. For example, ice cream preferences are subjective. If you ask me what flavor I prefer, there's not much correlation with what a random college professor prefers. The fact that both are subjective in fact makes it LESS likely that our tastes will happen to agree.
While individual teachers and professors have different subjective preferences, this doesn't matter, because on a broad level, high school and college are designed to be passable if you work hard (or cheat), but how much you learn is mediated by the effectiveness of the instruction, assigned homework, grading standards, etc. It's not fair that how much you learn in school depends on who you get (out of your control). I think it's fair when how much you learn in school depends on innate characteristics like intelligence (not mediated by the factors listed above) and what you choose to do (in your control). If SAT Math depends on Algebra II, then why not use a truly objective (hence fair, by my definition) test like an IQ test? It's a selection criterion that is not dependent on prior teaching quality, so students can be judged on their own merits.
I don't think that it's more fair to admit students based on innate characteristics like IQ, any more than it would be fair to admit students based on objective measures of their height or attractiveness. It goes without saying that your innate characteristics are very much out of your control, and hence unfair by your own standards. Anyway, one's grades are as a practical matter very much in your control even if there is a real aspect that is up to the teacher.
NBA players need to be a certain height. Models need to have a certain level of attractiveness. Academic study requires a certain level of IQ, although that level has decreased over the centuries and the college -> job link has really changed things. While innate characteristics are out of your control, they're also out of the control of external actors, so I don't think fairness is relevant. On the other hand, school learning is heavily influenced by external actors, so fairness is relevant. Also, I've been talking about learning (instead of grades), so I agree that one's grades are, practically speaking, in your control. But (formal) education is about learning (or it should be, imo), not grades, and I dislike how much it's tied up with life outcomes.
But there's something funny here -- if you want to actually predict how well students will do in theire academic studies you would be better off using grades. So if the reason IQ is "fair" is because it is relevant to performance, then grades should get the same pass.
This conversation has helped to put what my objection is into a clearer focus, so thanks for that. School performance requires intelligence (to be able to get to the starting line) and conscientiousness (to run towards the finish line, but also to jump over the hurdles of poor instructional quality, etc). Grades (which mostly measures conscientiousness) incorporate those hurdles, while IQ doesn't, so the latter is fairer. For grades to be roughly comparable to, or even better than, IQ, there would have to be universal adoption of high quality instructional approaches, such as Siegfried Engelmann's Direct Instruction, Steve Hare's (@sharemath on Twitter) approach to teaching math, etc. This would remove the hurdles and all that's left is getting to the starting line and running towards the finish line. That isn't the world we live in, but it's the one I want to move towards.
Something interesting about grades, the Crescendo Education Group, the main consulting group that pushes for and trains teachers on equitable grading, claims that one of the signs that equitable grading is more fair is that it correlates more strongly to testing scores. (See https://crescendoedgroup.org/results/) It seems they're doing something weird to get these correlations, but I thought it was interesting.
It's sort of funny to push for grades that are more like test scores...just use test scores! Though I suppose if you want to make grades that capture long-term information about a student's performance and eliminate things like effort...I can see the appeal of that.
I have always been opposed to standardized tests because they are particularly difficult for those with any type of learning difference. Especially those with dyslexia (which is approximately 1/5 the population regardless of language or culture). Wealthy LD children have a significant advantage because their disabilities are often caught early and they are given appropriate tools for learning.
Something to motivate any of your students who might be struggling with standardized tests: your cousin (who is a professor of Mathematics at an ivy league school) scored in the low 500s on his math SAT!!!
It is interesting Hattie included a study in his #1 influence, "Self Report Grades" in Visible Learning (VL). However, the study was on the reliability of students honestly reporting their grade to College Admissions. The study found a high correlation, but that students overstated their result (who would have guessed?). The high correlation when converted to an effect size yielded one of the largest ES in VL of 3.1. Hattie strangely interpreted this as students self reporting has one of the highest influences on that students achievement. https://www.researchgate.net/publication/240723602_The_Validity_of_Self-Reported_Grade_Point_Averages_Class_Ranks_and_Test_Scores_A_Meta-Analysis_and_Review_of_the_Literature
I've long been aware of the correlation between grades in high school and grades in college. But I don't have the same takeaway.
Kids who care about grades are markedly different from kids who don't. Kids in each group range from not too bright to brilliant. When they go off the college, the kids who care about grades will still care about grades. If they can't get good grades in one subject, they will change majors until they do. Kids who don't care about grades will not make these choices. It doesn't surprise me that kids who don't care about grades might be more likely to quit school.
What this research *doesn't* prove, I think, is that kids with low SAT scores and high grades are smarter than kids with high SAT scores and low grades.
I've never argued that SATs are fairer than grades. What they are, within broad categories, is more accurate about the tester's intelligence. And what we really should want to know about our poorer students is which ones are bright.
You can choose where you take the SAT (so a place with A/C), as well as retake it to make up for a bad day. Also, the things that make HSGPA subjective also make college GPA subjective, which explains why the former is so predictive of the latter and this affirms Bryan Caplan's education-as-signalling thesis. Education needs to be decoupled from employment and there needs to be greater liquidity in labor markets if social mobility is a concern.
1. I think to ask a kid to optimize their SAT-testing location is asking a lot. Also retakes cost money, can be demoralizing, etc. I'm not saying it can't be done -- but it impacts the fairness.
2. I don't think the dual subjectivity makes sense in the way you're saying. For example, ice cream preferences are subjective. If you ask me what flavor I prefer, there's not much correlation with what a random college professor prefers. The fact that both are subjective in fact makes it LESS likely that our tastes will happen to agree.
While individual teachers and professors have different subjective preferences, this doesn't matter, because on a broad level, high school and college are designed to be passable if you work hard (or cheat), but how much you learn is mediated by the effectiveness of the instruction, assigned homework, grading standards, etc. It's not fair that how much you learn in school depends on who you get (out of your control). I think it's fair when how much you learn in school depends on innate characteristics like intelligence (not mediated by the factors listed above) and what you choose to do (in your control). If SAT Math depends on Algebra II, then why not use a truly objective (hence fair, by my definition) test like an IQ test? It's a selection criterion that is not dependent on prior teaching quality, so students can be judged on their own merits.
I don't think that it's more fair to admit students based on innate characteristics like IQ, any more than it would be fair to admit students based on objective measures of their height or attractiveness. It goes without saying that your innate characteristics are very much out of your control, and hence unfair by your own standards. Anyway, one's grades are as a practical matter very much in your control even if there is a real aspect that is up to the teacher.
NBA players need to be a certain height. Models need to have a certain level of attractiveness. Academic study requires a certain level of IQ, although that level has decreased over the centuries and the college -> job link has really changed things. While innate characteristics are out of your control, they're also out of the control of external actors, so I don't think fairness is relevant. On the other hand, school learning is heavily influenced by external actors, so fairness is relevant. Also, I've been talking about learning (instead of grades), so I agree that one's grades are, practically speaking, in your control. But (formal) education is about learning (or it should be, imo), not grades, and I dislike how much it's tied up with life outcomes.
"Academic study requires a certain level of IQ"
But there's something funny here -- if you want to actually predict how well students will do in theire academic studies you would be better off using grades. So if the reason IQ is "fair" is because it is relevant to performance, then grades should get the same pass.
This conversation has helped to put what my objection is into a clearer focus, so thanks for that. School performance requires intelligence (to be able to get to the starting line) and conscientiousness (to run towards the finish line, but also to jump over the hurdles of poor instructional quality, etc). Grades (which mostly measures conscientiousness) incorporate those hurdles, while IQ doesn't, so the latter is fairer. For grades to be roughly comparable to, or even better than, IQ, there would have to be universal adoption of high quality instructional approaches, such as Siegfried Engelmann's Direct Instruction, Steve Hare's (@sharemath on Twitter) approach to teaching math, etc. This would remove the hurdles and all that's left is getting to the starting line and running towards the finish line. That isn't the world we live in, but it's the one I want to move towards.