Correcting Test for Extra Points

[HigherEd7]

What are your thoughts on letting students correct their exams to earn extra points? How many points should a question be worth? I guess my next question would be, will students even take the time to learn and correct the question or use the same answer from another student who missed the same question?

[the_geneticist]

What kind of questions are you using?  True/False, Multiple-Choice, short answer, diagram, calculate, etc?
For typical multiple choice, you should assume 1.5 minutes per question (if they don't need to calculate or draw to solve).

Regarding exam corrections:
I have tried offering exam corrections a few times* and won't do it again.  The logistics chaos of needing to re-grade & update scores wasn't worth it.  Now, these were mostly calculation & short answer questions for classes of ~60 students.

You also have to put a limit on how much they can earn back.

*Offered because I wrote an exam that was unreasonably long.  I let them choose 1 page to redo, usually the last one since that's the one they didn't get to.
Now, I'm just better at writing exams that 95% of students will finish 95% of the questions. 

I'd rather build in more formative, low-stakes assessments to give the chance to improve their learning.  Pedagogically, an exam is supposed to be a summative assessment - what can you achieve at a set point vs. what can you do with all the time & resources possible.

[Liquidambar]

I don't like doing it, but I've done it a couple times when there was a major disconnect between my expectations for the exam and the students' expectations (and therefore what they studied).  I've allowed students to earn up to half the points back.

The last time I did this was in a junior level class.  There were about 3 problems total, each of which had multiple parts.  I provided very specific instructions for the rewrite.  For example, part (a) could be worked in two different ways.  When you rewrite it, show both.  In part (b), explain what assumption was faulty in your original solution.  Etc.  Also, I didn't let students rewrite the whole exam.  For each student, I had flagged specific portions that would be pedagogically useful to rewrite.  It would be difficult to copy someone else's work and have it exactly match my criteria.  I'm pretty confident that the students learned something during the rewrite.

On the plus side, the students did learn, and I wasn't in the situation of giving embarrassingly low course grades to some of our top students.  On the minus side, the whole thing was a ton of work, about as much as writing and grading the exam in the first place.  Also my course evaluations were abysmal (not primarily because of this, but I'm sure it didn't help).  It's much better to write a good exam in the first place.

Next time I teach this course, I might have more exams so that each one is lower stakes.  At least I would space them differently so this exam didn't cover as much material.

ETA:  When I teach freshmen, I already have more exams built in for reasons that the_geneticist mentions.  I had higher expectations for juniors, though, and had opted to use that time for in-class problem solving and discussion instead of exams.

[Puget]

I don't do this, but I do give them the option of taking a cumulative final, which, if they do better on it, replaces their earlier exam grade(s). They questions are of course completely new, but it gives them a second chance to show what they learned if they struggled with one or both of the first two exams for whatever reason.

It also serves as the make-up for any student that misses an exam (which saves me from having to worry about who has a legitimate excuse or not -- everyone gets the same option).

Usually about 15-20% of the class takes this option. Of those, maybe 1/3 have a big improvement 1/3 a small improvement, and 1/3 little or no improvement (or even do worse).

[mythbuster]

I have done this in the past. I usually let them correct for half the value of the points missed. So if you originally earned a 3/10 (basically mercy point for writing anything) you can earn an additional 3.5 on that question to get to to a 6.5. This rewards those that got it the first time. I also require students to explain where their original answer went awry and where they found the correct information.
  It works well in terms of making you look caring and generous, and fixing things if there was some massive scoring snafu. But it does not work as well as I would like as a metacognitive exercise to change the approach to studying.

[Sun_Worshiper]

I did this once. It was a huge time suck for me and mostly helped the top students who had missed a few points but just had to have them back. Never again.

[fizzycist]

This is all the pedagogical rage in my circles, but I don't get it. If students now know the solutions (perhaps omitting a step or two, but still) and they then go back and write them out on their exams, how is this better than just giving them a fresh set of problems to work on?

I guess it is one of those mental/emotional things that works for many but not for me so I have trouble getting it. I'm the sort who gets great satisfaction from solving something on my own, and can tolerate not being able to get everything right, but hate the busy work of copying and pasting.

[statsgeek]

I've done this before, allowing students to earn back 1/2 of the points they lost for telling me the correct answer AND explaining why it is correct.  It is a lot of time to grade, but it does seem to help (at least some) update their understanding.  I've also experimented with doing this in one-on-one meetings, so their "explain" is not limited by their ability to explain in writing/I can ask follow up questions as needed and have a clearer picture of what they understand/we can discuss spots they're still "fuzzy".   

[HigherEd7]

Thank you everyone for the response. I agree with everything I have read here, and I like the idea of giving them the option to score higher on a cumulative final exam to replace their other exam grade.