In the "Hmwk Sets Editor" I marked a problematic question with "Mark Correct?" to take it out of the grading. I have now fixed the problem with the question.
Is there a way to undo the "Mark Correct?" feature and have the question be graded again?
I have looked everywhere and can't a way of doing this, but it is not unlikely that I have missed something. I am quite happy to edit configuration files directly or to query the mysql database if this is the only way to do this.
Andrew
The "Mark Correct" feature changes the status of the question for every student to 1 (i.e. gives them full credit). WeBWorK does not keep a record of scores that have been updated, so once you've used this, there is no way to get back the progress the students had before.
There are a couple of options I can think of to "reset" the question, sending everyone's score for that question back to zero. I'll start with a warning that none of these are tested, and there is significant danger of screwing up students' scores on other problems in the set if you are not careful.
My suggestion would be to simply add the problem again to the end of the homework set. This will mean that the total marks available for the assignment are one higher (e.g. it will be out of 11 instead of 10), which can be corrected when processing the grades if desired. If you're brave, you can then delete the original problem from the set, though I've had two many instances of weird behaviour when deleting a problem from a live set, so I personally am hesitant to do this.
With the right sql query you could reset the status for all students for that problem to 0. You should also reset their attempts if you go this route. I'm not going to try to craft a query here because there are too many ways to screw it up. The student status for individual problems is stored in the [coursename]_problem_user table. If you go this route, make sure you specify the set_id and problem_id.
Thanks for the ideas. I guess that this means that there is no way to get WebWork to remark existing quizzes? (Sorry, I realise this is probably a silly question, but I have only just started using WebWork.)
I was under the impression that if you go to "Scoring tools" and then re-score a quiz then all marks would be recomputed. It seems like a natural thing that people would want to do, but I assume this is not possible.
I was under the impression that if you go to "Scoring tools" and then re-score a quiz then all marks would be recomputed. It seems like a natural thing that people would want to do, but I assume this is not possible.
When you go to Scoring Tools and "Score selected sets" you are not asking it to reassess the correctness of the answers students have submitted. You are asking it to look up whatever score has already been written to the database for each question, and then tally that up for the set.
In theory it could actually re-load every problem and reassess student answers saved in the answer log. The first concern with that is it could take a long time. Sometimes students explore and try wrong answers after they have full credit for a question. So you could not do what you want just by testing their most recent answer to each question. You'd have to test the entire history of answers they had submitted, and that could be a lot of PG processing.
But more importantly, there is in general no guarantee that the current form of the question is the same as the one back when a student entered an answer. Imagine a question that is problematic for some seeds. For one student, it is not problematic, and they get 100% early after the set opens. But then you get reports of other students having issues, and you alter the .pg file to fix the bug. Maybe that first student's answer is now incorrect with the new version of the problem code, and what you would like to do with re-scoring would unjustly take away credit from that first student.
There are several other reasons why conditions may be different with a particular problem between when a student answers it for 100% credit and when you would re-score. So even though WW does store all the answer history for each student for each question, there would be problems if this kind of re-scoring were implemented.
I'll add to Danny's suggestion about adding an 11th problem. You could keep the bad problem, but change its weight to 0. I think that would keep your set at 10 total points, but leave the one question in place to avoid the concerns Danny raises about deleting a problem from an active set. Whatever you end up doing, keep in mind any students who believe they have completed the set and have 100%. If you add an 11th problem, or somehow take away the "marked correct" status of that one problem, then such a student may be oblivious to the fact that they should go back to keep working on the set. And therefore you may need to do some outreach to the class.
In theory it could actually re-load every problem and reassess student answers saved in the answer log. The first concern with that is it could take a long time. Sometimes students explore and try wrong answers after they have full credit for a question. So you could not do what you want just by testing their most recent answer to each question. You'd have to test the entire history of answers they had submitted, and that could be a lot of PG processing.
But more importantly, there is in general no guarantee that the current form of the question is the same as the one back when a student entered an answer. Imagine a question that is problematic for some seeds. For one student, it is not problematic, and they get 100% early after the set opens. But then you get reports of other students having issues, and you alter the .pg file to fix the bug. Maybe that first student's answer is now incorrect with the new version of the problem code, and what you would like to do with re-scoring would unjustly take away credit from that first student.
There are several other reasons why conditions may be different with a particular problem between when a student answers it for 100% credit and when you would re-score. So even though WW does store all the answer history for each student for each question, there would be problems if this kind of re-scoring were implemented.
I'll add to Danny's suggestion about adding an 11th problem. You could keep the bad problem, but change its weight to 0. I think that would keep your set at 10 total points, but leave the one question in place to avoid the concerns Danny raises about deleting a problem from an active set. Whatever you end up doing, keep in mind any students who believe they have completed the set and have 100%. If you add an 11th problem, or somehow take away the "marked correct" status of that one problem, then such a student may be oblivious to the fact that they should go back to keep working on the set. And therefore you may need to do some outreach to the class.