5 ms·

Dumb question, but are GPAs useless? I mean a plot of say high school GPA vs say college graduation rate or college GPA would be rather informative.

by ylem 6y ago

throwawaygh 6y ago
> I mean a plot of say high school GPA vs say college graduation rate or college GPA would be rather informative. You'd have to be very careful in how to measure both the inputs and the outputs. First, "weighting" makes high school GPAs pretty difficult to interpret without a lot of context. This is compounded by the fact that "weighting" is a bad solution to the underlying problem of trying to compare fundamentally incomparable coursework. There's really no numerical way to compare a grade in a remedial Biology course to a grade in an AP Biology course. Second, grading varies wildly from school to school and teacher to teacher, and what an "A" means will also vary wildly. Some teachers emphasize understanding, some emphasize busy work, and some hand of "A"s like candy. To give one example that might be useful to the HN audience: I've volunteered in several AP CS courses: - In one course, getting an "A" required writing thousands upon thousands of lines of if-else conditions following inane string formatting guidelines. Barely any looping. AFAICT the teacher was the least competent math teacher that was removed from teaching any math courses that were tested for in state standardized tests (never taught algebra or geometry -- just all the random remedial courses and electives). - In another course, getting an "A" required writing real programs, using some data structures appropriately, knowing the difference between constant/linear/"not linear" algorithm, etc. Students even wrote some simple proofs. Teacher was a CS major. - In a third course, basically everyone got in "A". Teacher was, I shit you not, a gym teacher. Literally a warm body. Course was basically self-study on days I wasn't there. So, even in relatively "standardized" AP courses, there's often not a nice bell curve, teachers aren't using the same rubrics or even assignments, and you really have no clue whether that grading reflects ability to put up with bullshit busywork or actual understanding. Third, you also have to be careful about how you're measuring the results. College GPA and graduation rates are not field-agnostic. A 4.0 in CS or Physics or Math is not easily comparable to a 4.0 in Communications. Also, universities aren't comparable. A degree/GPA at Small Noncompetitive Liberal Arts College is simply not comparable to a degree/GPA at MIT or Stanford. Also, dropping your lucrative major for a useless major can have an even bigger impact on your life than dropping out. The number of people who start in STEM and end up with a Communications or Business degree is just as relevant, or arguably even more relevant, than the number of Communications or Business majors who end up without a degree.
- ylem 6y ago
  These are all excellent points! But, I think this is the discussion we should be having. I doubt we can get the data, but universities have data on both the performance of students they have admitted (along with their schools) as well as their performance at the university. It seems like they could do their own analysis. You mentioned the question of what the A means. I wonder if just the act of working to get the A tells us something and that something may be more predictive of college success than the SAT/ACT. Or, another thing to imagine looking at would be say AP exam scores. I am biased, but I would imagine these would be more predictive than SAT/ACT and probably more informative than GPA. The question of change in major is interesting--but I think even the coarse question of say graduates/doesn't graduate would be interesting to see how it correlates with high school GPA (as compared to SAT/ACT) before delving into the more difficult cases that you rase.
TrackerFF 6y ago
They're not useless, but it's rather that when everyone's scoring high, it's not the best measure to differentiate candidates. What happens in other countries, where they only use grades/GPA, is that you start using more decimals - but then again, that's only useful when the size/population of "perfect" scores is low enough. If you have 10000 students with a perfect 4.0, and 2000 spots, then it's obviously impossible to just use grades - so it kinda loses its function for selection.
- ylem 6y ago
  Ok, so you think that say some state schools it might work ok, but not at say Stanford or Harvard. What about say folding in AP exam scores?
  - TrackerFF 6y ago
    If they're not gonna use tests as a measure, then it's probably gonna fall more on interviews / motivational letters / extracurricular activities. Again, when you have so and so many perfect students, you need to use other features - otherwise you might as well draw random names out of a hat. These schools (Ivy Leagues, Stanford, etc.) first and foremost want the students that are most likely to succeed at their schools. They know fully well that there are far more good students (on paper) than spots, but they also know that two 4.0 students are not alike. Some will fail, no mater how good they are in HS.
  - whotheffknows 6y ago
    Not every school has AP classes. In fact there's a direct correlation with AP classes existing in schools in neighborhoods where income is higher. This is why MIT got rid of considering Subject II tests because even on top of already relatively wealthy select schools that offer AP classes only a select few of those offer classes that prep you for subject II test. It's essentially exponentiating the amount of opportunities rich kids have to slice themselves within percents different from other rich kids at their school but overall entirely leaves behind students who weren't lucky enough to born in a neighborhood with a tax bracket that determines the schools funding.
Spivak 6y ago
High Schools are on the student’s side when it comes to college admissions. So when you’re on the honors track of your HS you almost literally can’t fail because your HS is trying to get you into better schools. Your HS is playing the game with you. The secret of being an honors kid isn’t aptitude but being told that you’re smart and your parents/peers expecting it of you.
- malandrew 6y ago
  > The secret of being an honors kid isn’t aptitude but being told that you’re smart and your parents/peers expecting it of you. This condescension isn't even based on fact. I was in some honors classes but not in others and for the classes I was in an honors class, I had more aptitude than my peers. For those in the classes I was not in, those peers had more aptitude than me (probably more motivation too since it was for subjects I didn't care for like English Literature)
  - Spivak 6y ago
    I mean I don’t think it’s condescending. I was the “smart kid” at my HS. People on the honors track weren’t any smarter than the rest, we just cared a lot more about our grades and looking good for colleges. The “low performers” during HS, now that they’re unburdened by a stifling institution, are showing that they’re just as smart and talented but in areas that school didn’t bring out. One is a pastry chef, quite a few are artists and graphic designers, one is a copyrighter, one does industrial design, two are in theater now. They could have done just fine in maths and physics but they, rightfully, didn’t really care.
- thaumasiotes 6y ago
  > The secret of being an honors kid isn’t aptitude but being told that you’re smart and your parents/peers expecting it of you. How much teaching experience do you have? I'm including any time you tried to convey a skill to any other person.
  - Spivak 6y ago
    Sure! I tutored the majority of my classmates in Calc and Physics and I did not find one person I believed couldn’t do it. I did find a lot of people who believed they couldn’t do it, weren’t good at math, and wanted to give up though. Some of the main problems: * A lot more people than you would expect have attention issues and just needed someone to give them one-on-one attention. They would get distracted in class for a minute and then be lost for the rest of the lecture. I got permission from our teacher to record the lectures and then had people work through them with the advantage of a pause and rewind button. * Some people were really independent and our textbooks were utter trash. I brought them “more advanced” books like Spivak ayyy that didn’t treat the reader as an idiot too stupid for a rigorous explanation. Having someone not “teach down” to them opened up so much. Easiest tutoring sessions ever when they just needed some quiet space to read. It’s amazing how much lack of understanding comes from years of teachers treating them as stupid and trying to have them memorize shortcuts that convey no understanding and don’t generalize. * A few people were very tactile and so I brought in my laptop with Mathematica and had a bunch of examples coded up and let them play with the sliders. So many people groked limits and integrals with this. * Some people needed more or less abstraction. Symbol overload is real. I loved saying something like “and this is true for any 17.” Connecting problems to “real life” wasn’t as effective as I had hoped but putting numbers in or having a variable cheat sheet they could quickly reference the meaning / value worked great! * Some people learn by machine learning. They need a bunch of problems to work algorithmicly to feel the pattens. I called them numeric tactile. Once they’ve done it 50 times or so they’re ready to hear the abstract explanation because they’ve oriented themselves. * A lot of people struggled because HS has so little focus on rote memorization. People didn’t have the toolbox of identities and little tricks at the forefront of their minds but some flashcards and cheat sheets fixed that right quick. I had like 5 people suddenly get trig identities once I made them memorize a few of the core ones. Constantly asking them to repeat something back to me whenever it came up was way more effective than I thought. As you can clearly see I didn’t have much of a social life in HS.
0d9eooo 6y ago
No. Empirically, they're as predictive as or more predictive of future educational performance (usually first-year college GPA) than test scores, despite grade inflation (which isn't as extreme as some would have you think). The advantage of GPA is it's an average over a long time period, unlike a test, which is a slice at a very brief moment. The disadvantage of GPA is variability in course difficulty and grading standards. What's interesting to think about is how variable grades and courses really are across institutions. For example, colleges do try to coordinate content a bit because of transfer issues etc.
- thaumasiotes 6y ago
  > [GPAs a]re as predictive as or more predictive of future educational performance (usually first-year college GPA) than test scores, despite grade inflation (which isn't as extreme as some would have you think). But in the context of this finding, the returns to grade inflation are limited by the fact that, when you report your GPA, you also report a standardized test score. This theoretically nullifies any potential benefit to grade inflation. (In practice, there's still a benefit.) When there is no outside check on the reliability of GPAs, might they adjust by inflating a lot more than they already have?
- kevin_thibedeau 6y ago
  They can be gamed too easily. My high school gave a 1 point bonus for AP courses. It it fair to compare someone with a 4.2 against students without that opportunity?
  - mennis16 6y ago
    Yeah, also private schools make a business out of their college admissions profile, I am sure as this becomes the norm they will inflate grades even more. I do think along with the application the school is supposed to submit a profile of the classes they offer and the average grades at the school and whatnot, but I do wonder how much admissions committees actually look closely at this. It's also hard to penalize a student for having all As, because maybe they did deserve them! I think we are going to end up having even more emphasis on legit extracurriculars, at least for the average middle class student.
pks016 6y ago
My viewpoint: GPAs are not useless. They are a good prediction of studying, learning and seriousness in those. But definitely not entirely predictive. May be less weightage to overall GPAs and more to final year GPAs(considering difficulty of courses).

Dumb question, but are GPAs useless? I mean a plot of say high school GPA vs say college graduation rate or college GPA would be rather informative.

by ylem 6y ago

Dumb question, but are GPAs useless? I mean a plot of say high school GPA vs say college graduation rate or college GPA would be rather informative.

throwawaygh 6y ago

> I mean a plot of say high school GPA vs say college graduation rate or college GPA would be rather informative. You'd have to be very careful in how to measure both the inputs and the outputs. First, "weighting" makes high school GPAs pretty difficult to interpret without a lot of context. This is compounded by the fact that "weighting" is a bad solution to the underlying problem of trying to compare fundamentally incomparable coursework. There's really no numerical way to compare a grade in a remedial Biology course to a grade in an AP Biology course. Second, grading varies wildly from school to school and teacher to teacher, and what an "A" means will also vary wildly. Some teachers emphasize understanding, some emphasize busy work, and some hand of "A"s like candy. To give one example that might be useful to the HN audience: I've volunteered in several AP CS courses: - In one course, getting an "A" required writing thousands upon thousands of lines of if-else conditions following inane string formatting guidelines. Barely any looping. AFAICT the teacher was the least competent math teacher that was removed from teaching any math courses that were tested for in state standardized tests (never taught algebra or geometry -- just all the random remedial courses and electives). - In another course, getting an "A" required writing real programs, using some data structures appropriately, knowing the difference between constant/linear/"not linear" algorithm, etc. Students even wrote some simple proofs. Teacher was a CS major. - In a third course, basically everyone got in "A". Teacher was, I shit you not, a gym teacher. Literally a warm body. Course was basically self-study on days I wasn't there. So, even in relatively "standardized" AP courses, there's often not a nice bell curve, teachers aren't using the same rubrics or even assignments, and you really have no clue whether that grading reflects ability to put up with bullshit busywork or actual understanding. Third, you also have to be careful about how you're measuring the results. College GPA and graduation rates are not field-agnostic. A 4.0 in CS or Physics or Math is not easily comparable to a 4.0 in Communications. Also, universities aren't comparable. A degree/GPA at Small Noncompetitive Liberal Arts College is simply not comparable to a degree/GPA at MIT or Stanford. Also, dropping your lucrative major for a useless major can have an even bigger impact on your life than dropping out. The number of people who start in STEM and end up with a Communications or Business degree is just as relevant, or arguably even more relevant, than the number of Communications or Business majors who end up without a degree.

ylem 6y ago
These are all excellent points! But, I think this is the discussion we should be having. I doubt we can get the data, but universities have data on both the performance of students they have admitted (along with their schools) as well as their performance at the university. It seems like they could do their own analysis. You mentioned the question of what the A means. I wonder if just the act of working to get the A tells us something and that something may be more predictive of college success than the SAT/ACT. Or, another thing to imagine looking at would be say AP exam scores. I am biased, but I would imagine these would be more predictive than SAT/ACT and probably more informative than GPA. The question of change in major is interesting--but I think even the coarse question of say graduates/doesn't graduate would be interesting to see how it correlates with high school GPA (as compared to SAT/ACT) before delving into the more difficult cases that you rase.

TrackerFF 6y ago

They're not useless, but it's rather that when everyone's scoring high, it's not the best measure to differentiate candidates. What happens in other countries, where they only use grades/GPA, is that you start using more decimals - but then again, that's only useful when the size/population of "perfect" scores is low enough. If you have 10000 students with a perfect 4.0, and 2000 spots, then it's obviously impossible to just use grades - so it kinda loses its function for selection.

ylem 6y ago
Ok, so you think that say some state schools it might work ok, but not at say Stanford or Harvard. What about say folding in AP exam scores?
- TrackerFF 6y ago
  If they're not gonna use tests as a measure, then it's probably gonna fall more on interviews / motivational letters / extracurricular activities. Again, when you have so and so many perfect students, you need to use other features - otherwise you might as well draw random names out of a hat. These schools (Ivy Leagues, Stanford, etc.) first and foremost want the students that are most likely to succeed at their schools. They know fully well that there are far more good students (on paper) than spots, but they also know that two 4.0 students are not alike. Some will fail, no mater how good they are in HS.
- whotheffknows 6y ago
  Not every school has AP classes. In fact there's a direct correlation with AP classes existing in schools in neighborhoods where income is higher. This is why MIT got rid of considering Subject II tests because even on top of already relatively wealthy select schools that offer AP classes only a select few of those offer classes that prep you for subject II test. It's essentially exponentiating the amount of opportunities rich kids have to slice themselves within percents different from other rich kids at their school but overall entirely leaves behind students who weren't lucky enough to born in a neighborhood with a tax bracket that determines the schools funding.

Spivak 6y ago

High Schools are on the student’s side when it comes to college admissions. So when you’re on the honors track of your HS you almost literally can’t fail because your HS is trying to get you into better schools. Your HS is playing the game with you. The secret of being an honors kid isn’t aptitude but being told that you’re smart and your parents/peers expecting it of you.

malandrew 6y ago
> The secret of being an honors kid isn’t aptitude but being told that you’re smart and your parents/peers expecting it of you. This condescension isn't even based on fact. I was in some honors classes but not in others and for the classes I was in an honors class, I had more aptitude than my peers. For those in the classes I was not in, those peers had more aptitude than me (probably more motivation too since it was for subjects I didn't care for like English Literature)
- Spivak 6y ago
  I mean I don’t think it’s condescending. I was the “smart kid” at my HS. People on the honors track weren’t any smarter than the rest, we just cared a lot more about our grades and looking good for colleges. The “low performers” during HS, now that they’re unburdened by a stifling institution, are showing that they’re just as smart and talented but in areas that school didn’t bring out. One is a pastry chef, quite a few are artists and graphic designers, one is a copyrighter, one does industrial design, two are in theater now. They could have done just fine in maths and physics but they, rightfully, didn’t really care.
thaumasiotes 6y ago
> The secret of being an honors kid isn’t aptitude but being told that you’re smart and your parents/peers expecting it of you. How much teaching experience do you have? I'm including any time you tried to convey a skill to any other person.
- Spivak 6y ago
  Sure! I tutored the majority of my classmates in Calc and Physics and I did not find one person I believed couldn’t do it. I did find a lot of people who believed they couldn’t do it, weren’t good at math, and wanted to give up though. Some of the main problems: * A lot more people than you would expect have attention issues and just needed someone to give them one-on-one attention. They would get distracted in class for a minute and then be lost for the rest of the lecture. I got permission from our teacher to record the lectures and then had people work through them with the advantage of a pause and rewind button. * Some people were really independent and our textbooks were utter trash. I brought them “more advanced” books like Spivak ayyy that didn’t treat the reader as an idiot too stupid for a rigorous explanation. Having someone not “teach down” to them opened up so much. Easiest tutoring sessions ever when they just needed some quiet space to read. It’s amazing how much lack of understanding comes from years of teachers treating them as stupid and trying to have them memorize shortcuts that convey no understanding and don’t generalize. * A few people were very tactile and so I brought in my laptop with Mathematica and had a bunch of examples coded up and let them play with the sliders. So many people groked limits and integrals with this. * Some people needed more or less abstraction. Symbol overload is real. I loved saying something like “and this is true for any 17.” Connecting problems to “real life” wasn’t as effective as I had hoped but putting numbers in or having a variable cheat sheet they could quickly reference the meaning / value worked great! * Some people learn by machine learning. They need a bunch of problems to work algorithmicly to feel the pattens. I called them numeric tactile. Once they’ve done it 50 times or so they’re ready to hear the abstract explanation because they’ve oriented themselves. * A lot of people struggled because HS has so little focus on rote memorization. People didn’t have the toolbox of identities and little tricks at the forefront of their minds but some flashcards and cheat sheets fixed that right quick. I had like 5 people suddenly get trig identities once I made them memorize a few of the core ones. Constantly asking them to repeat something back to me whenever it came up was way more effective than I thought. As you can clearly see I didn’t have much of a social life in HS.

0d9eooo 6y ago

No. Empirically, they're as predictive as or more predictive of future educational performance (usually first-year college GPA) than test scores, despite grade inflation (which isn't as extreme as some would have you think). The advantage of GPA is it's an average over a long time period, unlike a test, which is a slice at a very brief moment. The disadvantage of GPA is variability in course difficulty and grading standards. What's interesting to think about is how variable grades and courses really are across institutions. For example, colleges do try to coordinate content a bit because of transfer issues etc.

thaumasiotes 6y ago
> [GPAs a]re as predictive as or more predictive of future educational performance (usually first-year college GPA) than test scores, despite grade inflation (which isn't as extreme as some would have you think). But in the context of this finding, the returns to grade inflation are limited by the fact that, when you report your GPA, you also report a standardized test score. This theoretically nullifies any potential benefit to grade inflation. (In practice, there's still a benefit.) When there is no outside check on the reliability of GPAs, might they adjust by inflating a lot more than they already have?
kevin_thibedeau 6y ago
They can be gamed too easily. My high school gave a 1 point bonus for AP courses. It it fair to compare someone with a 4.2 against students without that opportunity?
- mennis16 6y ago
  Yeah, also private schools make a business out of their college admissions profile, I am sure as this becomes the norm they will inflate grades even more. I do think along with the application the school is supposed to submit a profile of the classes they offer and the average grades at the school and whatnot, but I do wonder how much admissions committees actually look closely at this. It's also hard to penalize a student for having all As, because maybe they did deserve them! I think we are going to end up having even more emphasis on legit extracurriculars, at least for the average middle class student.

pks016 6y ago

My viewpoint: GPAs are not useless. They are a good prediction of studying, learning and seriousness in those. But definitely not entirely predictive. May be less weightage to overall GPAs and more to final year GPAs(considering difficulty of courses).