Check out this report on homework: Heavier Homework Load Linked to Lower Math, Science Performance, Study Says. A better / more truthful headline would suggest that these researchers recommend one hour of homework per night.
HOWEVER, this sort of research concerns me because it suggests a causal effect from # hours of homework and test scores. Better stated, my concern is that the public will read into the science that a causal effect is the TRUTH. Clearly this study could not imply a causal relationship, and in fact many other studies negate that assumption. Alfie Kohn presents a discussion of a few key studies about the effects of homework, and in particular Math & Science homework at Homework: New Research Suggests It May Be an Unnecessary Evil.
Although it is reasonable to argue that Alfie Kohn may have bias, it seems to me his “bias” is to question or deconstruct the invisible assumptions that schools and U.S. Education writ large seem to operate under.
Me personally, I cannot bring myself to fully accept what Kohn suggests and advocate for NO HOMEWORK at high school level. But my take is different than the norm for school homework routines, but I find it is not usually argued with by math teachers. I think kids should have 20-ish minutes of math homework 4-5 nights per week, plus once every 2 weeks or so a bit larger of a project to complete, such as a portfolio demonstrating learning or a more formal paper about a mathematical investigation undertaken (i.e. a “POW write-up”).
The nightly homework should balance practice of recently developed procedural strategies with some challenging applications, sharing/talking about learning with others (family), and brand new ideas to consider in preparation for the next class. Of course, not all each night. Certainly the HW is not to be scored right/wrong but rather “credit” be given for doing some mathematical thinking between class meetings.