[LON-CAPA-users] questions about our first exam - partial marks

Mills, Douglas G dmills at illinois.edu
Mon Feb 3 15:05:15 EST 2014

I don’t know for sure if Gerd is ALWAYS right but upon closer examination
he IS right about this approach. We have not used it much but on looking
at it more closely now I see that using a variable as the value of the
weight for a part in a problem does not carry over into the grading
context and that particular part ends up being worth zero points whether
you have it “full credit” correct or “partial credit” correct. We missed
this in the couple places we’ve used it because at 0 points it doesn’t
show up in the possible total either, so students can get perfect scores ―
this part of the problem just doesn’t count.  Thank you Gerd for raising
the issue.

The approach we use most of the time, which DOES work is with
customresponse problems. We are not using the number of tries as a factor
in either case, so I don’t know (and now hesitate to venture!) whether or
not that’s a possibility with customresponse, but my guess would be that
it would be possible...


Douglas Mills
Director of Instructional Technologies
Department of Chemistry
University of Illinois
dmills at illinois.edu
(217) 244-5739

On 2/3/14, 11:33 AM, "Mills, Douglas G" <dmills at illinois.edu> wrote:

>So far we have not seen a problem with this. We’ve got a few problems set
>up like this across two very high enrollment courses. In these courses,
>both lab courses, the deciding factor on whether they get full credit for
>the question or half credit is the accuracy of their lab data (full credit
>if they do their calculations correctly and their data is accurate enough
>to lead them to the right conclusion, half credit if their data is sloppy
>but they do their calculations correctly at least, no credit if their
>answer is not supported by their data). These courses have from one to 2.4
>thousand students a semester and in general anything that can go wrong
>does for some percentage of them, so it *seems* like it’s working
>reliably. Some of these we’ve done as customresponse but I know some of
>them are other types as well.
>Douglas Mills
>Director of Instructional Technologies
>Department of Chemistry
>University of Illinois
>dmills at illinois.edu
>(217) 244-5739
>On 2/3/14, 10:27 AM, "Gerd Kortemeyer" <korte at lite.msu.edu> wrote:
>>On Feb 3, 2014, at 11:16 AM, Mills, Douglas G <dmills at illinois.edu>
>>> We have done something similar to your first challenge below by setting
>>> the weight of the problem equal to a variable and setting the variable
>>> the Perl scrip at the head of the document based on how many tries have
>>> been taken on the problem using the relevant EXT functions. I can
>>> code examples if you¹d like.
>>I am sorry, but I don’t think that is a save solution. This cannot
>>reliably feed back into the grade book, I would expect “surprises.”
>>- Gerd.

More information about the LON-CAPA-users mailing list