There are some interesting criticisms of that research, as it is often difficult to determine whether "homework" means "homework assigned" or "homework completed", as well as the longitudinal length (is this an ongoing assignment stretched over many days like Grant Wiggins would suggest is most effective) and the nature of homework (is it reading? worksheets? exercises? high-order? low-order?)
Still, the research suggest that there is a positive correlation between homework and achievement at the secondary grades (0.25), but a very slight negative correlation at the elementary grades (-0.04). And, you can see that this troubles Marzano and other researchers, as they followed it up with additional studies to nuance what homework means.
Actually, Marzano and I are not that far apart on homework, even though I'm more adamant about major changes in homework policies for elementary grades and the types of homework given to secondary. What is very poignant are his list of conclusions he draws about homework.
- Homework should be structured to ensure high completion. If it's too long or difficult for students to finish, what's the point?
- While there's no magic amount, there is a law of diminishing returns with the amount of homework. Adding more will not result in more achievement. It should be limited.
- It needs to have a well-articulated legitimate purpose. As he states, "homework assigned for punishment or to demonstrate to the public that a school is a serious place of study is not very defensible".
- In the same vein, it should relate directly to learning goals.
- It should be designed so students can do it on their own. (Which reinforces the idea that homework is not for learning, it is for practice).
- This means parents shouldn't be required to do the homework for their kids. The school should have an articulated policy of parents' relationship to homework. Involved, yes. Responsible for, no.