What science says about the benefits of homework

[Igel]

4. There have been a lot since Cooper's book was published.

And his was a meta-analysis, so it included a lot of divergent data collected under different conditions.

That's part of the problem.

I give homework to high schoolers and a lot of it is pointless. Kids don't do it. Or they don't "get it." Worse yet, they do it wrong and it reinforces mistakes, misunderstanding, and frustrates them.

I give homework to high schoolers and a lot of it is "pointful." Kids do it. Or they finally "get it." Best yet, they did it right in class (or not) but finally clarify their misunderstandings, learn the proper procedures, and come back with a feeling of satisfaction.

Now, average those two things together, weighting it for your particular sample.

I tried flipping the classroom, but I found that those kids that had the hardest time were the least likely to watch the videos at home and come in ready to review, reinforce, correct misunderstandings, and work on problems. They were also the least likely to have a stable home environment to work in, the least likely to value education over a part-time (or even full-time job).

Heck, I find that giving assignments in class with the warning, "You have to finish it at home" is a recipe for disaster. The kids who are least likely to do their homework are the most likely to say, "I'll socialize and play games now; I'll do this later". Then they don't do it, missing both in-class help and home reinforcement.

My default hypothesis is that these kinds of differences matter, which is what a lot of the research does show. If you're looking at AP students, you get one answer to the question, "Is homework any use?" If you're looking at homes where the kid who has homework and didn't "get it" in class has parents to help explain it, great. If you're looking at homes where the kids don't get pushed to do their homework, and any confusion or misunderstanding can't be corrected, it's a hindrance. No homework accomplishes the goal of decreasing the achievement gap. Mostly by lowering the upper end of the achievement curve, not by doing anything to raise the low end. But if the primary thing you think of as "helpful" is reducing the standard deviation of that average, hey, it's a no-brainer.