There's a bug in the coded correct third answer in Library/Michigan/Chap11Sec11/Q29.pg. The correct answer to the third part should represent two distinct values -d and +d where d is the sqrt of the discriminant. The form in which students have to enter this answer involves singleton intervals:
[-d,-d],[d,d]
However, the coded answer is accepting only the second of the two.
Murray Eisenberg
University of Massachusetts
Math & Stat
Amherst, MA 01003
I also reported this bug through the bugzilla link within WeBWorK. And I assume that eventually the bug will be repaired. But my issue now is how to deal with scoring this problem now that the set's due date has passed.
Obviously I cannot edit the question, since it's locked from me. If I made a copy and edited that so as to fix the answer to the third part, and substituted it for the original one from Library/Michigan in the assignment, would that allow me to correctly grade students' responses to the original (flawed) version?
If so, of course I'd have to code the answer to the third part right now so as to accept EITHER the (correct) answer of the form "[-d,d],[d,d]" OR the currently-accepted answer of the form "[d,d]". But I don't know how to do that.
Any other reasonable way than having to go through the answers for all 60 of my students one-by-one and add credit manually where deserved?
(No, I could not in good conscience just give everybody credit for the question, since some got all or part of the question wrong or didn't even do it.)
Obviously I cannot edit the question, since it's locked from me. If I made a copy and edited that so as to fix the answer to the third part, and substituted it for the original one from Library/Michigan in the assignment, would that allow me to correctly grade students' responses to the original (flawed) version?
If so, of course I'd have to code the answer to the third part right now so as to accept EITHER the (correct) answer of the form "[-d,d],[d,d]" OR the currently-accepted answer of the form "[d,d]". But I don't know how to do that.
Any other reasonable way than having to go through the answers for all 60 of my students one-by-one and add credit manually where deserved?
(No, I could not in good conscience just give everybody credit for the question, since some got all or part of the question wrong or didn't even do it.)
Hi Murray,
I hope to have a corrected version of this problem in the CVS later today.
There is a couple of difficulties impeding the effort to correctly mark students' scores in this case. Once we have the problem patched, if a student (or professor, acting as a student) submits her/his answer the system will indicate that the negative answer is correct. However, once the set is closed the student obviously can't submit the answer again to get credit, and in no case is the professor able to submit a student's problem for credit. Both of these behaviors are, I think, the most appropriate ones in general circumstances. Furthermore, my experience is that a student who has an answer marked wrong may then try another (very wrong) answer, with the result that the last submitted answer is not necessarily the correct one.
The result of this is that the effort of checking to see if a student should get credit may also require going back through her/his past answers. And I don't know of any way to automate the checking of all 60 students' current answers to see which of them had a correct answer that was not marked right.
If it is not appropriate to simply give all students credit for the problem (or to delete it) in this case, I would have any students who submitted a correct answer that was marked wrong contact me individually. This adds a small amount of work for them (they have to check their answer and e-mail me), and eliminates the substantial work required to go through each student to manually check whether they have a correct answer.
Gavin
I hope to have a corrected version of this problem in the CVS later today.
There is a couple of difficulties impeding the effort to correctly mark students' scores in this case. Once we have the problem patched, if a student (or professor, acting as a student) submits her/his answer the system will indicate that the negative answer is correct. However, once the set is closed the student obviously can't submit the answer again to get credit, and in no case is the professor able to submit a student's problem for credit. Both of these behaviors are, I think, the most appropriate ones in general circumstances. Furthermore, my experience is that a student who has an answer marked wrong may then try another (very wrong) answer, with the result that the last submitted answer is not necessarily the correct one.
The result of this is that the effort of checking to see if a student should get credit may also require going back through her/his past answers. And I don't know of any way to automate the checking of all 60 students' current answers to see which of them had a correct answer that was not marked right.
If it is not appropriate to simply give all students credit for the problem (or to delete it) in this case, I would have any students who submitted a correct answer that was marked wrong contact me individually. This adds a small amount of work for them (they have to check their answer and e-mail me), and eliminates the substantial work required to go through each student to manually check whether they have a correct answer.
Gavin