Import assessments with no assessment-level score detail
Currently the assessment data import wizard assumes that all assessment items have individual grades recorded. We should support the case where only the aggregate assessment grade for each student is recorded.
The schema for the assessment response data can remain the same, except that questionId will be null for all rows.
#2 Updated by Ryan McFall about 10 years ago
Ben, I want to verify that the EdML schema has a placeholder for the total score. If it does, then I would consider keeping this as part of the client-prototype, as I think it could be implemented in the upcoming month of work that I'm going to do. And, if it does exist, is it already displayed somewhere on the server? Are there assumptions in the data extraction code that the response data has been recorded at the item level?
I think that faculty should have the option to either upload individual scores, or to upload assessment level scores, but not both (in case they don't match up).
#3 Updated by ben leinfelder about 10 years ago
there's no place in the EdML (metadata) for the aggregate scores (data).
I think we though it could be accommodated in the same data file format that we currently use for item-level data, but we would leave the questionId column blank.
You can only have a single data file for any given assessment, so there is no real risk of totals and details not adding up.
I'm a little concerned that when we do the download over multiple assessments that the data will be misinterpreted - but we'll have to figure that out later down the road. Perhaps we will restrict users to only downloading assessments that have similar data (all totals only, or all-item-detail only, but never a mixture).
#4 Updated by Ryan McFall about 10 years ago
I'm going to mark this as after the client prototype, because it sounds like it
will require significant work on the server side, and it doesn't sound like we
have the resources for that. I think we need to have some design discussion as
to exactly how this should work. (Resubmit after fire)
#5 Updated by ben leinfelder about 10 years ago
I think you could [easily?] bypass the "assign question numbers to scores" step of the import wizard and end up with a good assessment-level data file. my previous comments about misinterpretation still apply, but i don't believe that's a "show stopper"
#6 Updated by Ryan McFall about 10 years ago
We could add another step in the import wizard that allows us to indicate whether we have individual item scores, or just aggregate scores. If I'm understanding you correctly, in the case of aggregate score only, we would then ask them to identify the column that contains the appropriate value, and that it will then be uploaded to the server like any other data file, just with no question ID.
I agree there are issues with potential confusion on download. It didn't sound like Mark/Diane thought this was that big of a deal during our teleconference on the 7th, so I think we should just not wory about it at this point.