The Truth About Homework
Widespread misconceptions about learning keep our children busy with needless assignments.
There’s something perversely fascinating about educational policies that are clearly at odds with the available data. Huge schools are still being built, even though we know that students tend to fare better in smaller places that lend themselves to the creation of democratic caring communities. Many children who are failed by the academic status quo are forced to repeat a grade, even though research shows that this is just about the worst course of action for them. Homework continues to be assigned— in ever greater quantities—despite the absence of evidence that it’s necessary, or even helpful, in most cases.
The dimensions of that last disparity weren’t clear to me until I began sifting through the research for a new book. To begin with, I discovered that decades of investigation have failed to turn up any evidence that homework is beneficial for students in elementary school. Even if you regard standardized test results as a useful measure, homework (some versus none, or more versus less) isn’t even correlated with higher scores at these ages. The only effect that correlates with homework is a more negative attitude toward school on the part of students who get more assignments.
In high school, some studies do find a correlation between homework and test scores (or grades), but it’s usually fairly small, and it has a tendency to disappear when more sophisticated statistical controls are applied. Moreover, there’s no evidence that higher achievement is due to the homework, even when an association does appear. It isn’t hard to think of other explanations for why successful students might be in classrooms where more homework is assigned—or why they might spend more time on it than their peers do.
The results of national and international exams raise further doubts. One of many examples is an analysis of 1994 and 1999 Trends in Mathematics and Science Study (TIMSS) data from 50 countries. Researchers David Baker and Gerald LeTendre were scarcely able to conceal their surprise when they published their results in 2005: “Not only did we fail to find any positive relationships,” they wrote, but “the overall correlations between national average student achievement and national averages in [amount of homework assigned] are all negative.”
Finally, there isn’t a shred of evidence to support the widely accepted assumption that homework yields nonacademic benefits for students of any age. The idea that homework teaches good work habits or develops positive character traits (such as self-discipline and independence) could be described as an urban myth, except for the fact that it’s taken seriously in suburban and rural areas, too.
In short, regardless of one’s criteria, there is no reason to think that most students would be at any sort of disadvantage if homework were sharply reduced or even eliminated. Nevertheless, the overwhelming majority of American schools—elementary and secondary, public and private—continue to require their students to work a second shift by bringing academic assignments home. Not only is this requirement accepted uncritically, but the amount of homework is growing, particularly in the early grades. A large, long-term national survey found that the proportion of 6- to 8-year-old children who reported having homework on a given day had climbed from 34 percent in 1981 to 58 percent in 1997—and the weekly time spent studying at home more than doubled.
Sandra Hofferth of the University of Maryland, one of the authors of that study, has just released an update based on 2002 data. In it, the proportion of young children who had homework on a specific day has jumped to 64 percent, and the amount of time they spent on it has climbed by another third. The irony here is painful, because with younger children the evidence to justify homework isn’t merely dubious—it’s nonexistent.
Why Homework Persists
So why do we do something where the cons (stress, frustration, family conflict, loss of time for other activities, a possible diminution of interest in learning) so clearly outweigh the pros? Possible reasons include a lack of respect for research, a lack of respect for children (implicit in a determination to keep them busy after school), a reluctance to question existing practices, and the top-down pressures to teach more stuff faster in order to pump up test scores, so we can chant, “We’re number one!”
All of these explanations are plausible, but I think there’s also something else responsible for our continuing to feed children this latter-day cod-liver oil. Because so many of us believe that it’s just common sense that homework would provide academic benefits, we tend to shrug off the failure to find any such benefits. Our belief that homework ought to help is based on some fundamental misunderstandings about learning.
Consider the assumption that homework should be beneficial just because it gives students more time to master a topic or skill. (Plenty of pundits rely on this premise when they call for extending the school day or year. Indeed, homework can be seen as a way of prolonging the school day on the cheap.) Unfortunately, this reasoning turns out to be woefully simplistic. “When experimental psychologists mainly studied words and nonsense syllables, it was thought that learning inevitably depended upon time,” reading researcher Richard C. Anderson and his colleagues explain. “Subsequent research suggests that this belief is false.”
The statement “People need time to learn things” is true, of course, but it doesn’t tell us much of practical value. On the other hand, the assertion “More time usually leads to better learning” is considerably more interesting. It’s also demonstrably untrue, however, because there are enough cases where more time doesn’t lead to better learning.
In fact, more hours are least likely to produce better outcomes when understanding or creativity is involved. Anderson and his associates found that when children are taught to read by focusing on the meaning of the text (rather than primarily on phonetic skills), their learning does “not depend on amount of instructional time.” In math, too, as another group of researchers discovered, time on task is directly correlated to achievement only if both the activity and the outcome measure are focused on rote recall, as opposed to problem solving.
Carole Ames of Michigan State University points out that it isn’t “quantitative changes in behavior”—such as requiring students to spend more hours in front of books or worksheets—that help children learn better. Rather, it’s “qualitative changes in the ways students view themselves in relation to the task, engage in the process of learning, and then respond to the learning activities and situation.” In turn, these attitudes and responses emerge from the way teachers think about learning and, as a result, how they organize their classrooms. Assigning homework is unlikely to have a positive effect on any of these variables. We might say that education is less about how much the teacher covers than about what students can be helped to discover— and more time won’t help to bring about that shift.
Alongside an overemphasis on time is the widely held belief that homework “reinforces” the skills that students have learned—or, rather, have been taught—in class. But what exactly does this mean? It wouldn’t make sense to say, “Keep practicing until you understand,” because practicing doesn’t create understanding— just as giving kids a deadline doesn’t teach time-management skills. What might make sense is to say, “Keep practicing until what you’re doing becomes automatic.” But what kinds of proficiencies lend themselves to this sort of improvement?
The answer is behavioral responses. Expertise in tennis requires lots of practice; it’s hard to improve your swing without spending a lot of time on the court. But to cite an example like that to justify homework is an example of what philosophers call begging the question. It assumes precisely what has to be proved, which is that intellectual pursuits are like tennis.