View this PageEdit this PageUploads to this PageHistory of this PageHomeRecent ChangesSearchHelp Guide

Assessment Discussion

We spent the major portion of 10/28/98 talking about assessment.

Mike noted that there are entire journals devoted to the subject of
assessment. Assessment is hard. It can run the gamut from "multiple
guess" questions on an exam to observation of students performing a
task, and from artefact-based assessment (ie, the finished work
products) to observing how tasks are done (watching students work in
the labs and evaluating their process).

We need to decide what forms of assessment are most suited to each area
in the curriculum.

A good way to think about assessment is the GQM model: goals, questions,
metrics:


  • what are the goals of the lesson/practicum?
  • what types of questions are appropriate?
  • what metrics can we use to assess students based on their respones?


Our task for the coming week is to work out the details in our
individual practicums. We decided that assessment of the "course"
(nugget) parts of the curriculum is "the professor's job" and thus
eliminate some work for us; only practicums need be considered here.

Require extra, low-quality work?


Lex Spoon proposes that for each practicum requirement, students should have to do, say, three separate versions of whatever is being assessed, and only one of the three has to be of high quality. That way, students have less pressure to "just meet the checkboxes and move on". It also allows some freedom on certain projects. Someone learns something from doing most any freeform work that consumes time; furthermore, when someone is focussing on meeting specific objectives, they often completely stop thinking abstractly and start viewing the project as just a funny game to be beaten.



Who does the Assessment


Who can do practicum assessment? TAs? Project Managers? Group members? How much input is allowed from each?


Allowing input from team members and project managers can be very informative. However, there is a small "buddies" problem (friends grade each other higher, they guy nobody likes gets graded lower), and furthermore people will be "acting" to some extent around anyone who has much sayso in the grading. This isn't ideal for learning: students should be encouraged to experiment.