### I'm Pretty Much Done for the Semester

OK, so I ignored the Title IV nonsense when I was entering the grades (need to go back and do that). And then there's the whole end-of-semester office-cleaning (including storing the finals and discarding the old finals). But I'm mostly done.

The calculus students bombed the final. Mostly because while they have acquired skill at integrating and differentiating, they are still unclear how to choose which is the one called for in a word problem. (I sorta knew that was going to happen, based on their dismal showing on the "integrate or differentiate?" problems that we did about a month ago.)

The math department would like to propose the following new spin on grading.

For our lower division courses that are graded on a straight percentage basis, you know, the type of course with a well-defined function from numerical averages to letter grades, we'd like to do away with letter grades. We're sick of hearing from students near letter grade borders explaining why they should be moved up to the next higher letter grade. (No one ever wants to be rounded down -- at least not under our current system.)

Now, you may be thinking that this just trades in one set of problems for another. Do we really grade precisely enough so that there is a meaningful different between a 78% and a 79%. No, we don't. Which is why we would record each student's average as well as the confidence interval. Imagine two C students. The one making solid Cs on everything might have a small confidence interval. The one who did really well on some things and then bombed some others? Much wider confidence interval.

(We already have someone who calculates quiz averages using "Olympic grading": He drops the lowest AND the highest score.)

Yes, this is what we think about instead of grading.

The calculus students bombed the final. Mostly because while they have acquired skill at integrating and differentiating, they are still unclear how to choose which is the one called for in a word problem. (I sorta knew that was going to happen, based on their dismal showing on the "integrate or differentiate?" problems that we did about a month ago.)

The math department would like to propose the following new spin on grading.

For our lower division courses that are graded on a straight percentage basis, you know, the type of course with a well-defined function from numerical averages to letter grades, we'd like to do away with letter grades. We're sick of hearing from students near letter grade borders explaining why they should be moved up to the next higher letter grade. (No one ever wants to be rounded down -- at least not under our current system.)

Now, you may be thinking that this just trades in one set of problems for another. Do we really grade precisely enough so that there is a meaningful different between a 78% and a 79%. No, we don't. Which is why we would record each student's average as well as the confidence interval. Imagine two C students. The one making solid Cs on everything might have a small confidence interval. The one who did really well on some things and then bombed some others? Much wider confidence interval.

(We already have someone who calculates quiz averages using "Olympic grading": He drops the lowest AND the highest score.)

Yes, this is what we think about instead of grading.