How to Differentiate One Science Test Three Ways (Without Writing Three Tests)
Every teacher knows the moment. You finish a solid unit test — aligned, rigorous, fair — and then you look at your roster. Three students have IEPs that require modified assessments. Four are emergent bilingual students still building academic English. Two finish everything early and ask what's next. The usual answer is three tests, three nights of work, and a grading pile that never quite matches up.
There is a better structure, and it starts with a rule: one standard, one test — three on-ramps. You do not write three tests. You write one strong test, then build two deliberate layers around it.
Layer 1: Core — the test as written
Start with the real thing. Twenty questions, full rigor, distractors built from the misconceptions students actually bring ("cold flows," "just average the two leg speeds"). Tag every item to its standard as you write it — that tag is what makes the other layers fast, and it powers your reteach data later. This Core version is the anchor: if a question isn't worth asking here, it doesn't belong anywhere.
Layer 2: Scaffolded — same standards, supported access
The Scaffolded version is not an easier test. It is the same standards with the barriers removed:
- Fewer choices: cut the weakest distractor (four options to three) instead of cutting questions.
- Simpler language: shorten stems, keep the academic vocabulary — students still meet words like dissolve and equilibrium, because the state test will use them.
- Sentence frames: "Thermal energy moves from ___ to ___ because ___" carries the structure so EB students can show the science.
- A glossary with Spanish cognates: la temperatura, la fuerza, el equilibrio — most core science terms have a cognate, and using it is a strength, not a shortcut.
- Graphics first: a diagram beside the question and an icon beside each glossary term, so meaning arrives before decoding does.
Build the Scaffolded layer while the Core test is fresh — it takes a fraction of the time because every decision is a transformation, not a rewrite. That is exactly how the Matter & Energy and Force & Motion assessment packs in my store are built: Core test, Scaffolded test, and the reteach and reassessment already keyed to the same standard tags.
Layer 3: Honors Extension — deeper, not more
GT and honors students do not need forty more questions. They need three harder ones. An Honors Extension takes the same standard and raises the thinking: evaluate a claim with a trap hidden in it ("oxygen is half the atoms in glucose — true or false?"), design a fair experiment and defend the variables, or critique a graph whose story contradicts its own data table. Grade it with a short worked key, and never let it punish — a missed extension should feed curiosity, not drag down a grade earned on the standard.
Why "starts together" matters
Because all three layers assess the same standards, the whole class starts together, takes the test in the same room, and reviews the same ideas afterward. Nobody is handed a visibly different test. Your tracking stays simple: one row per standard shows who needs reteach — regardless of which layer they took — and the reassessment closes the loop.
Make it a system, not a weekend
The first time takes an afternoon. After that, it is a template: Core first, transform down for Scaffolded, push up for Honors. If you use AI tools, this is one of the best jobs to hand them — my AI Prompts for Teachers pack includes a Three-Layer Differentiator prompt that turns one activity into all three versions in a single pass (you still check the keys — you are the expert, and the AI works for you).
One test. Three on-ramps. Same standards — and every student starts together.