Dear 2p2,

I was wondering if someone familiar with university curve grading could evaluate the strange method my professor used to set final grades for the class.

Keep in mind I only know the average, the standard deviation, my own specific score, and a rough grade distribution.


The average score for the class was a 193/400 or roughly a 48% and the standard deviation was 90.

For this class my professor failed half the class (D and F) and likely gave 25% Cs, 15% Bs and 10% As (I'm guessing with the Cs bs and as but it should be pretty close to the actual.

Now, I got 305.5/400 which is roughly a 76.35% and I received a B+. I believed that I deserved an A so I emailed my professor to ask why I got what I got.

She said that she graded on two curves. Anyone who got lower than 50% raw (200/400) failed. that was the first curve. The second curve was everyone above 50% and instead of using the 193 average as a benchmark, she used an average from one of her previous class from an earlier semester.

This doesn't make any sense to me and I was wondering what justification there is for using two curves as it definitely screwed me over.



Thanks

Doesn't your school have standard grade boundry distributions (that apply when scores haven't been rejiggled). I'm sure these would be useful to know aswell.
Also you're probably better off posting this in statistics/probability.

a B+ for a 76 seems like a pretty good deal unless this is some upper level aerospace engineering type class

even if she compared you to everyone, you are just a bit more than 1 SD above the mean. This puts you in like the 85th percentile. If she's only giving 10% As, then that puts you squarely at a B+

It seems ridiculous either way to fail half the class and then not grade the rest on the same curve. I feel like there is something we don't know about the class. OP, what major/year/size is the class?

I doubt this. Statisticians are no better than anyone else when determining fair grades. And besides, this question really doesn't have that much to do with statistics per se.

I'm sorry you got more than one letter grad more than you deserved. I hope that enough students complain about her so that you get the grade you actually earned.

Maybe she teaches that class pretty much the same each time, and you guys were really ****ty. The standard deviation also seems pretty high (maybe it isn't), so I assume quite a few people did well.

Class size was low 100s

Class was upper division biochemistry

4 year state university

Class mostly made up of 3rd and 4th year undergrad biochemistry majors

The only reason I'm posting is that whenever I've scored 20 or more % above the average? I've gotten an A. I realize it's possible that I didn't fall in the top 10% with the extremely high standard deviation but albeit unlikely

you just said that she gave out 10% As, so since you got a B its pretty natural for us to assume you were not in the top 10%

getting a 76 is an extremely thin argument for "deserving an A". If you got a 95/100 and she made some ******ed curve then fine.

sorry I was not clearer, I posted on my iPod earlier. I think when she failed half the class, she gave 10% of the people that passed As instead of giving 10% of the overall class an A. I am going to talk to her tomorrow. I just posted here to get some sort of confirmation that I wasn't just overestimating my ability to do well, but that's not going to happen, ha!

Wait, are you the OP? Two accounts?

lol at some random dude having gimmicks

yes, sorry, I am the OP

I didn't read a lot but you probably should be fighting for it, I REALLY think it makes no sense to use any data from previous semesters. To compare, sure, but to grade off of? No.

Unless it's part of the university written bylaws that professors have to give students a higher grade than they earned, then this is a losing argument. By the way, no university mandates that professors give a higher grade than they deserve. If this had resulted in a lower grade, then you could have a case.

I mean, go to a park and give a homesless guy $5. What do you do if he says, "wtf, dude yesterday gave me $10! I insist on you giving me an extra $5, right ****ing now!"

What do you mean "higher than they earned"? If you mean just the straight percentage grade then almost every engineering student would fail most of their classes. Most science/engineering classes have a curve because the averages are usually less than 60%.

I understand, but OP's argument is still not compelling.

I can see how you would want an A- but the SD is pretty high. But well done as your score is much higher than the average. Must be a difficult class/ or have a ton of idiots.

I got a B- this past semester in intermediate accounting and ended up getting 60's-70's on all 3 exams (no idea what i got on final). Sadly I did better than average on every exam. Only thing that tilted me was my teacher never discussed her curve so it was kinda nerveracking thinking I might get a C- or worse in a core class.

But the standard deviation here is over 20%. So your statement that I quoted doesn't seem to really be reasonable in this situation.

Well everyone for some reason assumes every single test will be normally distributed, for all we know given the test 1 SD puts him in the top 5%. Even then, it seems odd to fail half the people in an upper division class, by that point in the sciences people should have been through the weeder classes and there is no reason to do this.

Statistics, how do they work?

not sure if this is a shot at what i said or not, but all i'm saying is that it is sort of putting the cart before the horse to do statistics as if everything is normally distributed and then defend that the statistics should be used as though the scores follow a normal distribution.

A science/math/engineering class at an average of 50% is standard.

IME exam scores very rarely represent anything close to a normal (or discretized) normal distribution. They are much more commonly skewed way right.