The standard rhythm of school is test and move on. A unit is taught, a test is given, a grade is recorded, and the class advances whether or not anything stuck. The grade certifies that the student knew the material on the day of the test — which, it turns out, is a very different claim from knowing it at all. We have built an entire system that measures the peak of a curve we know is about to fall.
And earth science falls faster than almost any other subject, for a reason worth understanding.
Why earth science decays so fast
Some knowledge is sticky because it connects to something you already feel — a story, an image, a lived experience. Earth Science, taught badly, has none of that to hold onto. It is names and relationships: this mineral's hardness, that cloud type, the order of the rock cycle, which boundary makes which landform. A fact you have memorized but not understood is held in place by nothing. The moment you stop rehearsing it, it slides off.
Two topics in particular are notorious for this, and both are load-bearing for everything that comes after:
- Plate tectonics. It is the organizing framework of earth science — nearly every landform, earthquake, and volcano runs through it. Memorize the boundary names without grasping why the plates move and what each boundary builds, and the whole apparatus evaporates the week after the test, taking the rest of the year's geology with it.
- The rock cycle. It asks the student to hold a genuinely dynamic idea — that any rock can become any other kind of rock through melting, weathering, heat, and pressure, on a loop with no beginning — and reason about the paths through it. Crammed, it becomes a fog of arrows and half-remembered rock names. Mastered, it becomes intuition.
When these decay, they don't fail quietly. They pull down weathering and erosion, the history of Earth's landscapes, oceans and atmosphere, and astronomy and Earth in space with them, because those units assume the earlier ones are still standing.
Learn, Master, Retain
The course replaces the test-and-move-on cycle with a three-stage one: Learn → Master → Retain. Learn is the first encounter with the idea, on Concept Day and at the bench. Master is the harder threshold — the student can reproduce the reasoning, explain it, and apply it to a problem they haven't seen before. And Retain is the part the ordinary model skips entirely: deliberately returning to the idea after time has passed, so it is rebuilt rather than allowed to fade.
The engine for that last stage is two well-established practices that the course bakes into its schedule:
- Spaced practice. Instead of one concentrated burst before a test, a concept is revisited at widening intervals. Each return is slightly effortful — you have to reconstruct a little — and that effort is precisely what cements it.
- Retrieval. The student is asked to produce the answer from memory before checking it, not to re-read until it feels familiar. Pulling weathering, erosion & soil out of your own head, repeatedly, is what makes it stay there. Recognition feels like learning and isn't; retrieval feels harder and is.
Mastery is not seat-time. A student does not understand plate tectonics because the calendar spent two weeks on it. They understand it when they can rebuild the reasoning on demand — and that is what we measure.
Why mastery beats seat-time
The old model confuses coverage with learning. It assumes that if a topic was taught, and time was spent, and a test was passed, then learning occurred. But the forgetting curve does not care how many days the syllabus allotted. It only responds to whether the knowledge was built deeply and revisited deliberately.
So in this course a student advances through a concept when they have actually mastered it — demonstrated, in their own words and their own work, that they can reproduce and apply it — and not merely because the unit is "over." "Not yet" is an honest and expected default, not a failure. The rubrics are what make that judgment fair and repeatable. The goal was never to get the student through the test in October. It was to make sure they can still do the earth science in March — and in the year after that.