Hacker News Reader

Comment by heartsucker | Hacker News Reader

At my university (UCSB '12), grades were close to meaningless. I got a BSc in Math and graduated with a 2.8 (or lower?) because I loaded up on upper division math courses. I had friends who got 3.9's because they got a BA in communications or biz. econ. So clearly GPA varies heavily with degree, not to mention institution.

There were some professors who straight up said they always failed the bottom X percent of the class. So if you knew the material and "passed," you still got an F. Or the professor whose final was literally "memorize this two page proof and regurgitate it verbatim." Or the <ethnic/gender> studies professor who fail your papers for disagreeing with them.

There should just be Pass / Fail. Enough with this GPA circus.

View on HN

I agree about moving to pass/fail, if the bar to pass is moved up a bit. I went to a school with relatively low grade inflation and I would say any I didn't really have a good understanding of a class unless I got a B or higher, barring some mathematical oddities (e.g. 50% of grade determined by one test).

I suppose it depends on the professor's philosophy of what a test should be as well. If the test is intended to make students stretch their knowledge then it might be OK to make a C. That would mean you understand what was covered in class but couldn't make the next intuitive leap on the test. However, most tests I took weren't like that. Anyone who understood the material should have been able to make at least a B.

UCSB '12 here as well. I got a BSEE and I found the contrast between grades in lower division and upper division classes to be quite stark. In the lower division classes you could get away with anything - for example people cheated on their online physics and math homework so everyone did well. The classes were too large so the TAs grade everything, and it's much easier to convince a TA to give you more credit for something than it is a professor. The class size makes professors resort to things like iClickers for participation points and the online homework I just mentioned. It degrades the quality of education while also creating a system with more loopholes that smart people can exploit, since everyone is optimizing for grades in the end.

In upper division, the classes got considerably smaller and considerably harder. Grades meant more, and the system was not easily played. When you have a class of 20 you can't screw around anymore. You can't hide behind your TA. You get what you put in.

That's not to say I didn't have funky grade experiences in upper div too - I had one class series that actually had a serious grade deflation problem. The average on our second quarter final was 18.5%. In the end, if you tried hard and did well in the lab, your grade was basically a participation score. Now, I learned an enormous amount in that class because it was so hard. It was demoralizing at times because everyone though they were going to fail (I vividly remember contemplating my future during that final, wondering if I would ever pass that class). So in that case, grades didn't really matter.

Personally I think the solution comes with class size. Grades become inflated because as more people are managed by a professor and their TAs, the less careful they have to be while handing them out. There's too many exceptions and they can't spend all their time tracking down who is being truthful when they say their dog ate their homework or their roommates keep them up all night because they're drunk. It's much the same problem as large corporations have with employee performance scores - as the body you are judging grows, your criteria become much more generalized and the judges much more disconnected from those being judged.

I think it's more than that. A college is worth $X because its graduates go on to Do Great Things. Now suppose there's only 10 colleges in the world. 9 inflate grades to a 3.8 average and 1 keeps it at 3.0. Who is going to get hired easier?

I think grade inflation is two fold. 1) students gaming the system and chasing grades rather than education, and 2) the system gaming against other participants.

The student graduating from the prestigious university in the difficult major will be interviewed first.

The student graduating from the state university in the difficult major will be interviewed second.

Somewhere down the line the consulting companies and body shops will look at gpa's.

I couldn't agree more with most of your points but I think that pass/fail isn't the right approach.

Taking an aggressive schedule in order to get your "I spent four years doing what I was told" paper and GTFO ASAP hurts your grades

Working a job hurts your grades

Extracurricular hurt your grades.

What I see is a simple duty cyle problem. Students have a finite productive capacity (not sleeping isn't a semester-term solution). If they take a schedule that requires more than a certain % duty cyle then they take a GPA hit. If you could account for the amount of productivity required by a class you can take duty cycle into account when calculating GPA. For example, 120 credits in 4yr is a lot easier than 120 credits in 3yr. Five 1xx or 2xx gen-eds in one semester is a ton easier than four 3xx or 4xx classes in a STEM discipline. Having some way to account for the extra effort required to run wide open for a semester or run close to wide open for 3yr would go a long way.

People playing the game (college) on easy mode (BA in business, no job, is a member of a frat with a really good test bank his major, etc) get the same GPA as people playing on hard mode (engineering degree with minor in something else, academic club involvement, works a job on the side, etc) and there's no way to tell them apart when it's all said and done.

If there was a way to score classes (and professors) based on typical grades then profs could grade how they want and students could take what they want and not be punished. You could do something like track the average grade for a class/prof combination and track average grades for students. You could eliminate the effect of one prof who teaches one of many sections of one required class as well as the effect of a bunch or really good/bad students enrolling in a class on one particular semester. You could also eliminate the effect of taking a ton of really hard classes or unknowingly enrolling in a section taught by a really tough prof.

Basically if you applied dynamic curves to classes at the university level then you can account for easy/hard classes, profs and schedules. There's got to be some catch but I can't think of one that couldn't be worked around with effort since it doesn't have to be perfect, just better than nothing. You don't need to make getting bad grades in four hard classes have an equivalent GPA to getting good grades in four easy classes, just making some way to partially account for the difference would help. Obviously this wouldn't be in the financial interest of colleges but that's a different issue.

Sure, there's probably an more effective way of differentiating students, but at the end of the day, you'll just have another system that can be gamed somehow. P/F is simplest approach, and sometimes worse is better.

Don't knock the pass/fail thing until you've tried it :)

I attended another UC that offered students the option to take classes pass/fail. I used it for gen-ed requirements so I could focus my attention on getting good grades in the classes for my major. It was really nice not to have to stress over a test in a non-core subject knowing that it wouldn't hurt my GPA. And, in retrospect, I think I actually remember more of the material from those classes than the classes I took for a grade. My guess is that stress is counter-productive to actually learning.

There is something to be said for pass-drop, with a high standard for passing.

Anything less than a pre-inflation "A" is stricken from your record. Any "B+" work (or worse) won't affect you. Retake until you succeed.

Only the advanced (mostly final year) classes are required. Take the beginner ones if you feel you need them, but they never count for anything and there isn't even a record that you took them.

> My guess is that stress is counter-productive to actually learning.

I'd qualify this with too much stress is counter-productive to learning.

Hard mode has better classmates. If your goal is truly is to learn, go where the smart people are.

Reply on news.ycombinator.com