Sarah's too cool for school, while JJ needs a lesson on how to wear
a backpack without attracting the attention of bullies.
I can't believe that summer is over and the kids are back to school. Yep, back to the cold, hard grind of first and third grades. All right, I'm exaggerating; aren't you allowed to do that in a blog? But what I dread most about returning to school is forcing my son to do his homework, and to do it well. (My 6-year-old daughter obviously reads and hews to the research showing that girls have a more positive attitude about homework.) You wouldn't think that 10 measly worksheets a week could provoke such angst, but how could they possibly lure my son away from the siren call of the Wii? Some of my friends send their kids to private schools that refuse to assign homework, and a couple of my more EBM-minded pals have even pointed out, "You know, there's no evidence that homework is effective." Could that be true?
The amount of research on homework is vast and generally crappy. One literature search on the topic found over 4,000 papers, which I was planning to summarize for you, until I was fortunate enough to run across a 2006 systematic review on the subject ("Does homework improve academic achievement?").
The authors found only four trials that randomized either students or classrooms to homework vs. no homework. All of the trials were small, and not one was published, which speaks to their quality. Most used as their outcome measure a test of the specific subject material. All four studies showed that students who were assigned homework performed better on the tests than those who did not. For example, in one randomized trial of high school students who were studying Macbeth, those who got the homework were able to correctly identify that MacDuff was a Scottish nobleman and not a Simpsons beer company mascot. ("Student and parental homework practices and the effect of English homework on student test scores.") The reviewers tried to inject a note of optimism into their summary of the meager evidence, stating, "While each set of studies is flawed, in general the studies tend not to share the same flaws.”
Despite the lack of experimental data, there are multiple large observational studies looking at the effect of homework on academic achievement. Many were based upon the National Education Longitudinal Study, which gathered achievement test scores, school grades and surveys filled out by tens of thousands of high school students, teachers and parents. It turns out that in this age group, there is indeed a strong positive correlation between homework and test scores and grades. And what was the optimal amount of homework for high school students? One study found it to be 7 to 12 hours a week, or about 1-2 hours per night. Spending over 20 hours a week (oh, the horror!) was no better than spending only 1-6 hours a week. Of course, the most obvious flaw in these studies is that it is impossible prove causality. Does homework lead to good test scores and grades, and by inference, improved learning? Or do better teachers in upper-income schools, where there is more parental involvement, tend to assign more homework?
And what about elementary school students? Here's where the evidence is at its shakiest. This paper's meta-analysis of cross-sectional studies on students in grades K-6 found that there was no correlation between the amount of time spent on homework and academic achievement. This latest review agrees with an analysis published 17 years prior that also showed no increasing benefit with increased amount of homework in this age group -- 15 minutes a week is as good as 15 hours.
So what are we left with? I would have to agree with the authors' conclusion that more experimental research is needed, particularly in elementary school, though who would fund such a study, or whether teachers and parents would find it acceptable, is unclear. Interestingly, in the past couple of decades, the average amount of homework assigned to the youngest age group, 6 to 8 years old, has doubled from 1 to 2 hours a week. I know an extra hour isn't much. But I suppose if I were a purist, I would demand that a randomized, controlled study be done comparing 2 hours a week of homework to 2 hours a week of Super Mario Brothers.
Nintendo, are you listening?