Dear JE,

Thanks for your response. Yes, you've got it right. If we're married to
the idea that homework scores should be computed as total achieved/total
possible and not average(achieved/possible) for each problem set, then I
think this is the fairest way to interpret "dropping the lowest", at least
from the POV of the student.

In practice, there are about 35 problem sets which are mostly the same size
although there are some larger and some smaller. Most of the time the
scores dropped are zeroes. So the situation you describe -- I'm impressed
by your example-choosing skills, btw -- would be rare for me.

Oh, and this component amounts to 10-15% of a student's course grade, and we
grade on a curve, so I'm not sure how much a difference in course grades
this policy actually makes. But the students appreciate the policy because
it's (somewhat) forgiving.

--Matt


On 6/3/05 12:34 PM, in article
, "JE McGimpsey"
wrote:

I'm not sure I follow. Do you really mean that you drop the scores that
achieve the most improvement? That seems counterintuitive to me.

To simplify, take an example where there are four scores, and you drop
one.

A B C D
1) 9 7 37 9
2) 10 10 50 10

So before dropping a score, the average is SUM(A1:D1)/SUM(A2:D2) = 77.5%

Dropping the lowest score, in column B (7/10 = 70%) results in (55)/(70)
= 78.57%

Dropping the score in column C (37/50 = 74%) results in (25)/(30) =
83.33%

So you really want to drop a higher % score (74% vs 70%) on a more
important assignment (50 vs 10)???

That seems bizarre!


In article ,
Matthew Leingang wrote:

First, let me say that I have a solution to this problem but I am looking
for a better one. Second, I apologize if this gets a little long.