In a blog post today, David Brooks, op-ed columnist at The New York Times, gives a brief overview of a forthcoming study in the Economics of Education Review that seems to suggest more homework has no value for students in any subject but math.
The study entitled “Are we wasting our children’s time by giving them more homework?” is certain to be taken out of context, with the result that babies will be washed out with the bathwater like farms below the Morganza Spillway.
Economists Ozkan Eren and Daniel J. Henderson took a representative sample of US eight graders and correlated the amount of homework they were assigned in math, science, English, and social studies with their test scores in those subjects. (Note that homework assigned does not mean homework done.)
The full study is worth reading. It is only 31 pages and a third is bibliography and appendices. If statistics make your eyes glaze over, you can skip the paragraphs with the Greek terms.
The researchers took great care to rule out all sorts of factors that might impact their results, including such things as:
- Student characteristics, such as race, gender, and family’s socioeconomic status.
- Teacher characteristics, such as race and gender as well as graduate degree and state certification status.
- Class characteristics, including class size, number of limited English proficiency students, number of hours the class met weekly, the amount of time the teacher spent administering tests and quizzes.
- Teacher evaluation of the overall class level, how much of the text the teacher covered, number of hours the teacher spent each week maintaining discipline.
The researchers apparently did not attempt to determine whether the homework activities and the test items covered the same ground. I suspect that the reason the math homework had a postive correlation with test scores was that the math homework questions and the math test questions were very similar.
I doubt very much that homework in science, English, and social studies would correlate well with questions on the test used in the study. American education has standardized tests that are taken nationwide, but no standard curriculum that guides study nationwide. The lack of standard curriculum is less obvious in math than in science, history, and English. Without agreement on, for example, social studies topics that all eight graders should study, the likelihood of social studies homework boosting test scores strikes me as pretty remote.
Moreover, science, history, and English focus (or should focus) at least as much on the thinking processes used in those disciplines as on specific facts. Those processes do not produce right answers in the same way that solving an algebraic equation produces right answers.
Homework in science, history, or English may be directed toward having students discover multiple options rather than toward one right answer. We tend to regard solving the math problem as mathematical thinking, but real mathematical thinking is as likely to result in several possible solutions as to find one “right” one.
Giving homework in which students develop a hypothesis to test empirically or having students write an essay about American history may be more important in the long run than giving homework on material that is more easily tested by blacken-the-bubble methods.
[11-27-2012 updated Eren-Henderson link]