New grading strategy, part I: a grading rubric.
I was going to tackle the list of skills first. In fact I did. Then I found myself waffling over how much info makes up a skill. If the pieces are too small, there will be a thousand of them to grade. If they’re too big, it will be impossible to summarize them in a single grade. I kept thinking about whether it was assessable — which meant knowing how I would assess.
So I decided to try the rubric first. Here’s a bit of blogosphere roundup:
|Max Score||To get full marks:|
|Continuous Everywhere but Differentiable Nowhere||4||Demonstrate the skill
Use algebra correctly
Use correct notation
|Teaching Statistics||4||Demonstrate skill
Solve a complex problem
Solve independently (lower scores for solving with assistance
|Brain Open Now||4||Demonstrate the skill
Use algebra correctly
Use correct notation
Draw valid conclusions
|MeTA Musings||4||Demonstrate full understanding|
|Sarcasymptote||5||Demonstrate comprehensive knowledge
Use it in novel situations
|Point of Inflection||4||Demonstrate mastery
Connect to other skills
A lot of rubrics differentiate the score based on algebra (3 if you get part of the idea but make conceptual errors; 4 if you get the idea but make algebra errors). This makes sense in a math class, but since I’ll be testing things like “can predict circuit behaviour”, I’m tempted to make algebra a separate standard. If you make an algebra mistake that’s serious enough to draw wrong conclusions from, it’s a conceptual error. But a lot of algebra mistakes have such small effects on the goal that they don’t change your predictions or problem-solving strategy. There’s a whole other course that tests algebra, so maybe it’s wasteful to make it the main criterion in my grading scheme. I think writing is also a separate standard. There will be chances to assess the students’ ability to integrate their own writing with their own measurements/troubleshooting as part of project work (which I think I’ll grade on a separate rubric).
It’s possible to score based on how much assistance someone needs, but I’ll be assessing only in situations where assistance is not available. “Demonstrate full understanding” and other similar language like “proficient” is not specific enough for me — I worry that students would have no definitive way of knowing what I consider proficient, short of asking me a million times a week (causes anxiety for them; causes insanity for me).
The last two, though, seem reasonable to me. What I’m really testing — if I had to pick one thing — is ability to apply the knowledge to a situation that’s not exactly like anything in the textbook (though it might be a mix-and-match combination of previous problems). This is also the hardest thing for students to understand. I get a lot of pushback from students who are angry that I’m testing them on things I “haven’t taught” them — this despite practise problems, in-class time Q&A or group work, and hands-on activities practising the skills from various angles. There’s a high correlation between that attitude and the Fs that came through on the last test. So I’d like clarify that, in fact, that is what I teach… and what I expect them to aim for.
I’m tempted to use a 5 point system. Someone wrote recently that they didn’t like 5 because it gave a “middle ground” score of 3 that was neither good nor bad. I disagree: there is always a score of 0. So it’s the 4-point scale that has a fence-sitting score. In the final weighting, 2/4 will probably translate into 50%, which is the cutoff grade for supplemental exams. That’s definitely something I want to avoid, both for my sanity and the students’. They will want to know which side of passing (60%) they’re on: a 3 means yes, a 2 means “not close”. So I think my synthesis so far is this:
- understands something about the concept
- understands lots of things about the concept but not everything
- gets the concept, has problems with strategy or application of skill
- applies the skill to a previously practised situation
- chooses appropriate concepts to combine with and/or applies to novel situations
The wording needs some work, but basically breaks down the way I want it to. Assuming I’ll average these for the final grade, a 4 is 80% (this is technically considered “honours” by the school, so I might have to make this worth 79% or something like that). To get a 5 you’d have to do something cool. If you did something cool and it mostly worked, I’d be ok giving a 4.5. Some students will probably be confused about what the “appropriate” other skills are for getting a 5, but I think I can draw up some guidelines for that. On the other end of the scale, if you don’t have a good grip on what the point is, you are below 60%. Seems sensible so far.
Next up: some really interesting ideas I’ve stumbled across for evaluating rubrics. I’ll put this one through the ringer and see what survives.