There’s a move in some parts of the country to get all eight graders to take algebra, rather than having students wait for high school. In “The Aftermath of Accelerating Algebra,” economists Clotfelter, Ladd, and Vigdor explain that a primary cause for the push is the observation that students who take algebra early have better performance later in school. The argument is that early algebra and good results are correlated, so let’s have more early algebra.
The authors point out that students who take algebra early are generally better students. Being a good student leads you to take algebra early and to do well in school otherwise as well. Early algebra doesn’t make you a better student.
To test this idea, the authors looked at an experiment in Charlotte-Mecklenberg where the proportion of students taking algebra in eighth grade skyrocketed and then fell back for a five year period starting in 2002/03. Because the change was due changes in district policy, it wasn’t just the best students taking algebra early. Even if the district didn’t intend to run a controlled “experiment,” that’s pretty much what they did.
What happened? Accelerated students performed worse in algebra and worse or no better in later math courses. Good evidence that acceleration is a bad idea.
This is why statisticians are prone to chant “correlation isn’t causation.”