AP Environmental Science rewards a peculiar combination of skills: ecological literacy, quantitative reasoning, and the ability to design a fair test under time pressure. Of the three free-response questions on the exam, the design-experiment prompt is the one most students walk into under-prepared, because classroom labs tend to emphasise content recall rather than the language of experimental control. This article breaks down exactly how that question is scored, the rows on the rubric that separate a 3 from a 5, and the specific moves that turn a vague proposal into a defensible one.
What the design-experiment FRQ actually asks
The design-experiment prompt gives you a scenario, a hypothesis, and a context, then asks you to design a laboratory or field procedure to test it. You will normally receive a hypothesis such as 'Increased nutrient loading increases algal biomass in freshwater systems' and a constraint set: time, equipment, location. The prompt is not asking you to repeat the hypothesis back. It is asking you to demonstrate that you can isolate one variable, hold the others constant, and choose a measurable response. The 2025-26 course and exam description confirms that this question is part of the Investigative Skills cluster, weighted alongside data analysis and environmental solutions.
Three features separate a credible design from a weak one. First, a single independent variable is manipulated across a sensible range — for APES, 'across a range' usually means at least four treatment levels if you want to claim a defensible dose-response. Second, every other variable is named as a constant, with at least one example of how it will be held steady. Third, the dependent variable is operationalised with a unit and a measuring instrument. 'Measure biodiversity' is not enough. 'Count individuals per quadrat using a 0.25 m² frame, replicate five times per plot' is.
The four rubric rows and where students typically drop points
For most candidates reading this, the easiest gains live in the controlled-variables row and the replication row. The APES design rubric runs along four predictable lines.
| Rubric row | What a 5-level answer shows | What a 3-level answer usually misses |
|---|---|---|
| Identifies the independent variable | Names IV explicitly, gives a range with units | Names IV but no range, or conflates IV with the response |
| Identifies dependent variable | Names DV with units and a measurement method | Names DV without units or instrument |
| Controls other variables | Lists at least three controlled variables and how they are held constant | Mentions 'controlled variables' generically with no specifics |
| Replication and sample size | States n ≥ 3 per treatment and explains why this supports inference | Omits replication, or says 'repeat many times' without a number |
If you are losing two of these rows, you are sitting at a 3 even when your ecological content is strong. In my experience, the controlled-variables row is the single most common deduction across scoring samples I have reviewed.
Controlled variables: the row students treat as a footnote
Most candidates write 'control all other variables' in one phrase and move on. The rubric treats that as zero credit. A controlled variable is a specific factor that, if it varied, would offer an alternative explanation for your results. For the algae nutrient prompt, the obvious ones are light intensity, temperature, water volume, and starting algal density. Naming three of those is the floor for a top score. Naming them and stating how each will be held constant — same incubator, identical 250 mL beakers, shared stock culture split into equal aliquots — is what raises the row to full credit.
A useful working test: if a grader asked 'how exactly would you make sure X does not change between your treatment and control?', and you cannot answer in one sentence, the variable is not yet controlled on paper. Practise the language now. 'Light intensity held constant by placing all beakers 30 cm beneath the same 4000 K LED panel on a 12-hour timer' is one such sentence. The specificity is the score.
Replication, sample size, and the 90-second fix
Replication is the cheapest credit on the design rubric, and yet it is dropped constantly. The 90-second fix is mechanical: every time you describe a treatment level, append 'n = 5' or 'five replicates'. The fix is cheap because you are not doing new science, you are labelling what is already implicit. For field studies, the equivalent is a transect count or a quadrat count, and the threshold to claim 'enough' replication is rarely lower than three. Below three, your design cannot support a confidence interval, and the rubric will not credit the statistical thinking the course description demands.
Be honest with yourself about the trade-off. If your design uses a sensitive instrument that is slow to read, four replicates with one minute of measurement per sample is realistic; ten replicates with ten minutes each is not, and a reader will know. Match the number to the constraint set the prompt gives you, then justify it in one clause.
Quantitative output: graphs, units, and what the rubric expects
A 5-level design states how the data will be presented, not just how they will be collected. Specify the type of graph (bar, scatter, line), the axes with units, and a sentence on what pattern would support or refute the hypothesis. This is the row where students confuse 'design' with 'lab procedure'. Design includes the analytic plan. If you cannot state the expected shape of the result — a positive linear relationship, a logistic curve, a threshold response — the grader cannot tell whether you understand the hypothesis you were given.
For most APES prompts, the expected shape is monotone: more nutrient input yields more algal biomass, up to a point. Naming that expectation in one sentence — 'a positive linear relationship at low nutrient concentrations, plateauing as phosphorus becomes non-limiting' — is enough to earn the row. Saying 'we will graph the data' is not.
Common pitfalls and how to avoid them
- Conflating hypothesis and prediction. The hypothesis is given. Your job is the prediction — the specific pattern you expect, with direction and ideally a unit threshold.
- Designing a survey when the prompt asks for an experiment. 'Ask 50 students whether they recycle' is not an experiment. If the prompt gives a hypothesis about cause and effect, the design must manipulate an IV.
- Forgetting ethical and safety constraints. Field studies with vertebrates, working near waterways, or chemical use must include an explicit safety or permit note. One line is enough.
- Skipping the units row. Every measurement cited without a unit — 'temperature', 'distance', 'mass' — is a quiet deduction. Treat units as part of the design, not a finishing touch.
- Writing 'control' without saying what is controlled. A control group is not the same as a controlled variable. The design question usually rewards both, named separately.
Writing the design under timed conditions
You have roughly 22 minutes for the design question, and most of that should be spent in planning, not in prose. Spend three minutes sketching the variables on the page, two minutes listing three controlled variables with the mechanism of control, and one minute committing to a sample size. The remaining time goes into writing. If you cannot finish the full design, write the controlled-variables and replication sentences first — those are the cheapest credit per word on the rubric, and graders mark them line by line.
In practice, students who run out of time on this prompt usually spent too long on the introduction. The prompt has already given you the context. Two clauses acknowledging it is enough. Move to the design.
Conclusion and next steps
The design-experiment FRQ is the highest-leverage free-response point on AP Environmental Science because it rewards disciplined writing more than it rewards ecological breadth. Train the controlled-variable sentence, commit to a sample size, and always state the expected shape of the result. AP Courses' one-to-one AP Environmental Science programme marks past design FRQs against the four-row rubric, identifies which row you habitually drop, and turns the controlled-variable language into a timed drill you can run before the exam.