Why We Teach Math as Logic, Not Formulas

A formula you memorized is a thing you can forget. A result you derived is a thing you can rebuild, even when the test puts it in a form you have never seen. That is the whole reason we teach our students to derive math rather than memorize it. It is not a style preference. It is the difference between knowledge that survives a hard problem and knowledge that does not.

This is the principle behind everything in our curriculum, so it is worth being precise about why we hold it, and what the evidence actually says.

Researchers split math knowledge into two parts. Procedural knowledge is knowing the steps: cross multiply, flip the second fraction, move the term across the equals sign. Conceptual knowledge is knowing why those steps are legal: why flipping a fraction divides, why a balanced equation stays balanced when you do the same thing to both sides.

The old assumption was that you drill the procedure first and understanding shows up later, if at all. The research does not support that picture. In a 2015 review in Educational Psychology Review, Rittle-Johnson, Schneider, and Star synthesize a body of studies showing that the relation between the two kinds of knowledge is bidirectional and develops iteratively: gains in understanding drive better procedures, and better procedures deepen understanding, in a loop. Neither is the junior partner. A student who only has steps is missing half of what makes the steps work.

The practical consequence is direct. When you teach the procedure alone, you are teaching one of two halves and hoping the other arrives on its own. When you teach the idea, the procedure tends to come with it, and it transfers to problems the student was never explicitly drilled on.

If memorizing procedures worked, the students who lean hardest on memorization would be the strongest. They are not.

In Scientific American (November 2016), Jo Boaler and Pablo Zoido reported on student strategy data from the international PISA math assessment, which tests hundreds of thousands of 15 year olds across dozens of countries. Their finding, stated plainly: "In every country, the memorizers turned out to be the lowest achievers." The United States had a high share of memorizers (in the top third) and also a high share of teens doing poorly on the assessment. By the authors' account, the students who leaned on memorization sat roughly half a year behind the students who used relational and self monitoring strategies, the ones who connected ideas and watched their own reasoning.

That is a large, real population of students, measured the same way across countries. The pattern is not subtle, and it points the same direction as the laboratory research: the path through understanding outperforms the path through rote.

We turn this into a rule we actually follow. A lesson does not open with a formula in a box. It opens with the situation the formula came from, and we walk the student to the result so that by the time the formula appears, it is something they just built rather than something they were handed.

The payoff shows up on a test. The SAT does not ask you to recite the slope formula. It hands you a graph, a table, or a word problem and expects you to recognize that slope is what the question is about, then produce it. A student who memorized the formula has to first recognize that this unfamiliar setup calls for it, which is exactly the step memorization does not teach. A student who understands slope as rise over run sees the same idea no matter how it is dressed up. That recognition under unfamiliar conditions is what researchers call transfer, and it is the skill a real exam measures.

This is also why memorization is not a shortcut, even when it looks faster up front. We say more about that in can math be memorized. And because understanding has to be built in order, not crammed, the derive it yourself approach pairs naturally with mastery learning: you do not move on until the current idea is solid, because the next idea is going to stand on it.

Deriving math is slower at the start. A student who is used to copying steps will find the first weeks of "but why does that work" harder than memorizing a procedure and moving on. That cost is real, and we are not going to pretend it away.

What the evidence says is that the cost is front loaded and the return compounds. The understanding you build does not evaporate the week after the quiz, it does not collapse when the problem is rephrased, and it is the thing that actually moves a score on a test like the SAT, where the bar at the most selective schools sits very high (see what SAT math score you need for the Ivy League). We teach math as logic because logic is what holds up when the formula slips your mind, and on a hard test, the formula always eventually does.