I am teaching an introductory course on proofs in mathematics in a mid-size American public university, and trying to develop some kind of consistent grading meta-scheme for grading proofs. I am writing to see if anyone can refer me to math ed papers about grading proofs, or if anyone has such a meta-strategy that was helpful to use in the past.

Such a meta-strategy might look like: grade each proof out of 20 points, assigning 5 points to each of

  • Correctness - are definitions and previous results used correctly

  • Logic (micro) - does each step follow from the previous step

  • Logic (macro) - has the student understood what is to be proved, and chosen a path that is appropriate (i.e. correctly set-up a proof by contradiction, or direct proof, or etc)

  • Readability - has the writer used sufficient language to communicate the intended arguments

(Incidentally, this is what I came up with this morning, in an attempt to start grading the first exam, but I suspect there are educators else where who have thought longer and harder about this problem, and I am dying to hear what they have to say.)


1 Answer 1


This paper: "An assessment model for proof comprehension in undergraduate mathematics" can be used as a starter. As a starter, You have to work on it to convert the ideas discussed into grades.

  • 2
    $\begingroup$ Dear David. Proof learning and teaching has a vast literature. But, I am afraid to say that most of it is about what students could not learn and what lecturers could not achieve! There is hardly any concrete answer to such a concrete question as you asked. Even the paper that I have suggested though with a very nice and related title, just provides some very rough ideas never used in a real class. I would be very happy to be in touch if you make any concrete progress. $\endgroup$ Feb 1, 2015 at 23:13
  • 3
    $\begingroup$ Could you summarize the most relevant parts of the document. SE discourages link only answers. $\endgroup$ Jul 23, 2015 at 7:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.