When I started teaching Calculus, I knew I wanted some regular form of homework. But I wanted to avoid something like "2-40, evens." Primarily, I think the way textbooks are designed don't allow for the kind of focused assignments I want. In 2015 I started rolling my own homework. The goal was something short and to the point that was on older material. I'd give it out on Monday, check it Friday, and provide a solution kids could check at their leisure.
Since implementing that system, my homework completion rates are huge. After 2 or 3, kids were used to the procedure of when it was due. The publicly posted key reduced the number of questions to almost zero, and the familiarity of the material kept kids from getting stuck and giving up.
Here's a sample assignment:
I skip weeks every so often depending on circumstances. So far I've given 12 assignments like this in AB Calc. This year, I finally decided to gather some data on homework habits. I had some ideas based on solution traffic (it's a bit.ly so I can track usage stats), but I wanted the real numbers.
I have the exact same system for BC, but their homework completion rate is 100% and it counts for a meaningless part of their grade, about 1% per assignment, 4 assignments per grading period. That crowd appreciates the opportunity for extra practice a little differently.
With 70 AB Calc respondents, here's what the kids had to say:
Strangely, the majority of kids doing homework on the day it's due are in my 1st period. Though perhaps if they stay up past midnight they counted that as Friday?
The even clustering of solution users is fascinating. I thought it'd be more consistent. I really don't think there's a good or bad takeaway from these percentages. I post the solutions intending for the kids to use them, and most of them use them most of the time.
The four kids who said the assignments are too long are the ones who consistently don't do them in the first place, which is kind of funny.
The two kids who admitted to 5+ (out of 12) copies are incidentally mediocre in overall classwork. My eyes popped a little when I saw how many had copied at least once.
In general, the system seems to be working as intended. A big majority are doing the assignments and checking their work. A small minority are abusing the system. At roughly 3% per assignment (~4 assignments per grading period), a student does not gain a whole lot from a blind copy.
You could argue I could make changes. Increase the weight or increase the frequency but it's likely to only amplify what is currently a small problem. Increase the frequency and it loses meaning (being seen as constant busy work). Increase the weight and it becomes punishment.
Probably the only clever thing would be to introduce a check digit of sorts into my solutions. I could add one intentional wrong/weird thing that'd signal a blind copy. Though I make natural errors with enough frequency that it could be hard to remember what was supposed to be wrong.