Arrow wrote:I do not think that subjectivity/arbitrariness is a big deal. Anytime a teacher or professor grades a paper this problem arises. It has been like this since your high school English teacher graded your final paper on Hamlet.
Here is how I think our school system gives out grades. First, the professors do not give out any letter grades. They assign point values anonymously to exams, and they give those points over to the registrar. The registrar then applies the forced curve.
I guess it is possible for professors to tweek numbers a little bit so more people get A's, but I doubt this occurs.
The registrar then averages the numbers and point values from the professor and calculates a standard deviation. Now, if you are 2 standard deviations above the mean, then you get an "A". For those of you who forget, 2 standard deviations from the mean puts you within the 5 to 95 percentile (al la 68-95-99.7 rule). Thus, only about 5% of the people will score over 2 standard deviations above the mean and get the "A". In addition, if you are 3 standard deviations above the mean, you will probably get an "A+".
Somewhere along the way, it also normalizes this into an 100 point scale if there are multiple choice questions or something.
Now, what this means is (excuse the pun), the curve + subjectivity may play a factor, both of which are out of your control. For example, using this curve, in one of my classes the 6 people got "A"s and in another, 2 people got "A"s.
Thus, if the exam is easier and the average is higher, then less people get A's. You can have the 3rd highest grade in the class out of 80 people, but it can be an "A-" instead of an A because you got screwed by the curve.
Now for another example. In one of my classes, the exam was tortuously hard. The average was very low. However, two people got "A+"s and another four got A's. In addition, this hard exam led to more people failing, thus allowing more A's to be given.
Then there is my legal writing class, which is another example of an "easy" grading/bad curve. No "A"s were given and only two "A-"s were given. The scores were just too squeezed together that no one made it 2 standard deviations above the mean. In addition, no one failed or did badly enough to push the "A-"s up to "A"s.
Finally, all the sections are compared together when the school does its ranking. Thus, if you have more teachers with easier exams that give out less "A/A+"s, then you can be in the top 2-4% even if you are number 1 in your section.
This is the luck that I am thinking of (all true for my section/year). It is simply just out of your control. Can you just imagine the frustration some people (*sigh* I admit, myself included) felt when they realized their section did not give any "A"s in one class while other sections did (and even gave out "A+"s)?
By the way, all this matters little in the grand scheme of things. Who cares how they rank? Just do the best you can. If you get screwed by the curve, it ultimately should not really mess with your rankings that much. At most, my GPA could have been number 1 (top 1%) or number 9 (top 3%).
Ungenerous T2 curves for the loss?
Lawyers love math, right?
Another extremely helpful post, Arrow.