AP Statistics FRQ Tips: State, Plan, Do, Conclude

Q: What does State, Plan, Do, Conclude stand for?

It is a four-step template for inference free-response answers. State gives the parameter and the hypotheses, Plan names the procedure and checks conditions, Do shows the statistic, test statistic, and p-value or interval, and Conclude compares a test result to α or interprets the interval, then answers in context. Labeling the four steps helps a scorer find each component.

Q: How many free-response questions are on the AP Statistics exam?

Section II has 4 free-response questions worth 10 points each, answered in 90 minutes, and it counts for 50% of the exam score. They appear in a fixed order, and Question 3 is the dedicated inference question, either a hypothesis test or a confidence interval.

Q: Do I lose points for using a calculator on the FRQ?

No, a graphing calculator with statistical capabilities is expected for both sections. You do lose the work-shown point if you write only the calculator command, such as a function name, without reporting the statistic, test statistic, and p-value. Show the values the procedure produces alongside any calculator use.

Q: Should I write hypotheses about the sample or the population?

Always write hypotheses about the population parameter, such as p (the population proportion) or μ (the population mean). Hypotheses about a sample statistic like p-hat (the sample proportion) lose credit, because inference makes claims about the population, not the sample you already measured.

By Jude Wallis · Published July 28, 2026

Write inference free-response answers in four steps: State names the parameter and hypotheses, Plan names the procedure and checks conditions, Do shows the statistic, test statistic, and p-value or interval, and Conclude compares a test to alpha or interprets an interval, then answers in context.

This maps to Section II of the Fall 2026 AP Statistics exam, where Question 3 is a 10-point inference free-response and the CED task verbs shape what each part asks for.

How the Section II free-response section is structured

The AP Statistics exam is 3 hours long. Section II is the free-response section: 4 questions worth 10 points each, answered in 90 minutes, and together worth 50% of your exam score. Each question is worth 12.5% of the exam score.

The four questions always come in the same order:

Question 1: Multi-Focus on Practices 1 and 2 (formulating questions and collecting data).
Question 2: Multi-Focus on Practices 3 and 4 (analyzing data and interpreting results).
Question 3: Inference, either a hypothesis test or a confidence interval.
Question 4: Multi-Focus on Practices 2, 3, and 4.

A graphing calculator with statistical capabilities is expected, and a formula sheet and tables are provided for both sections. This guide focuses on Question 3, the inference question, because that is where a clear template earns the most points.

State, Plan, Do, Conclude: the four-step inference template

Most inference free-response answers follow a four-step structure often called State, Plan, Do, Conclude. State sets up the parameter and the hypotheses. Plan names the procedure and checks its conditions. Do runs the numbers. Conclude interprets the result in the context of the problem.

The scorers award points component by component on an analytic scale, so they look for specific things in each step. Missing any one component can cost a point even when your arithmetic is correct. The sections below map each step to what earns credit.

State: hypotheses and parameters

Define the parameter in words before you write symbols. Use $p$ for a population proportion or $\mu$ for a population mean, and say what it refers to, for example: let $p$ be the true proportion of all customers who prefer the new layout.

Then write both hypotheses. The null hypothesis $H_0$ states a specific value, and the alternative hypothesis $H_a$ states the direction or difference you are testing. For a one-sided test of a proportion this looks like $H_0: p = 0.5$ and $H_a: p > 0.5$ .

Two mistakes lose credit here. Writing hypotheses about a sample statistic such as $\hat{p}$ (the sample proportion) instead of the parameter $p$ is wrong, because hypotheses are always claims about the population. Leaving the parameter undefined also costs a point. For more on setting these up, see null vs alternative hypothesis.

Plan: name the procedure and check conditions

Name the exact procedure you will use, such as a one-sample $z$ -test for a proportion or a two-sample $t$ -interval for a difference of means. If you are unsure which one fits, which statistical test to use walks through the choice.

Then verify the conditions and show that you checked them with the numbers, not just their names. For a one-proportion $z$ -test you check that the sample was random, that the sample is no more than 10% of the population, and that the expected counts $np_0$ and $n(1-p_0)$ are each at least 10, where $n$ is the sample size and $p_0$ is the null value. Write the actual products, for example $np_0 = 100(0.5) = 50 \ge 10$ .

Stating a condition without showing it holds does not earn the condition point.

Do: statistic, test statistic, and p-value or interval

Show the sample statistic first, then the test statistic, then the p-value or interval. The formula sheet defines the standardized test statistic as

\text{test statistic} = \frac{\text{statistic} - \text{parameter}}{\text{standard error of the statistic}}

For a hypothesis test about a proportion, use the null value $p_0$ in the standard deviation of the sampling distribution, $\sqrt{\frac{p_0(1-p_0)}{n}}$ .

You may use your calculator, but a bare calculator command is not a substitute for showing the value. Writing 1-PropZTest with no numbers can lose the work-shown point. Report the test statistic and the p-value that the procedure produces. See reading calculator output for how to translate the screen into a written answer.

Conclude: link the p-value to alpha, in context

A complete conclusion has two parts: a decision and an interpretation in context. Compare the p-value to the significance level $\alpha$ (often 0.05). If the p-value is less than or equal to $\alpha$ , reject $H_0$ ; if it is greater, fail to reject $H_0$ .

Then state what that decision means for the original question, using the context and units. For example: because the p-value of 0.0082 is less than $\alpha = 0.05$ , we reject $H_0$ , so there is convincing statistical evidence that more than 50% of all customers prefer the new layout.

Never write accept $H_0$ , and never conclude in bare statistical language with no reference to the setting. A conclusion that could be copied onto any problem earns no context credit. For a confidence interval, interpret the interval in context and, if a claim is involved, say whether the interval supports it. See what does a p-value mean for the reasoning behind the comparison.

Common point-losers and AP Statistics FRQ tips

These AP Statistics FRQ tips address the mistakes that cost the most points on inference questions:

No context in the conclusion. A decision without a sentence tied to the setting loses the interpretation point.
Bare calculator commands. Listing a calculator function name with no statistic, test statistic, or p-value does not count as showing work.
Wrong parameter symbols. Using $\hat{p}$ or $\bar{x}$ (the sample mean) in hypotheses, or mixing up $\mu$ and $p$ , signals you are testing the wrong quantity.
Naming conditions without checking them. Write the numbers that show each condition holds.
Skipping the parameter definition. Say in words what $p$ or $\mu$ represents before you use it.

Answering in complete sentences and labeling each of the four steps makes it easy for a scorer to find every component.

What the CED task verbs are asking for

The wording of each part tells you what to produce. The Course and Exam Description defines the task verbs, and matching your response to the verb keeps you from over- or under-answering:

Calculate: perform mathematical steps to arrive at a final answer, such as a test statistic or an interval.
Justify: provide evidence to support, qualify, or defend a claim, and provide statistical reasoning to explain how that evidence supports the claim.
Interpret: describe the connection between a mathematical result and its meaning within the realistic context of the problem.
Compare: describe similarities and differences; numerically, this requires a quantitative comparison of two or more values.
Complete: identify parts, justify conditions, and calculate the results for the appropriate statistical inference method.

A Justify part needs reasoning, not just a number, while an Interpret part needs a sentence in context. For the full exam framework, see AP Statistics or the College Board course page.

A full-credit one-proportion z-test, step by step

A store manager claims that more than half of all customers prefer a new store layout. In a random sample of 100 customers, 62 prefer the new layout. Test the manager's claim at the $\alpha = 0.05$ significance level, and write the response using State, Plan, Do, Conclude.

State: Let $p$ be the true proportion of all customers who prefer the new layout, a population proportion. $H_0: p = 0.5$ versus $H_a: p > 0.5$ , using $\alpha = 0.05$ .
Plan: Use a one-sample $z$ -test for a population proportion. Random: the 100 customers are a random sample. 10%: 100 customers is less than 10% of all customers. Large counts: $np_0 = 100(0.5) = 50 \ge 10$ and $n(1-p_0) = 100(0.5) = 50 \ge 10$ . All conditions are met.
Do, statistic: the sample proportion is $\hat{p} = \frac{62}{100} = 0.62$ .
Do, standard deviation under the null: $\sqrt{\frac{p_0(1-p_0)}{n}} = \sqrt{\frac{0.5(0.5)}{100}} = \sqrt{\frac{0.25}{100}} = \sqrt{0.0025} = 0.05$ .
Do, test statistic: $z = \frac{\hat{p} - p_0}{\sqrt{p_0(1-p_0)/n}} = \frac{0.62 - 0.5}{0.05} = \frac{0.12}{0.05} = 2.40$ .
Do, p-value: for a right-tailed test, p-value $= P(Z > 2.40) = 1 - 0.9918 = 0.0082$ .
Conclude: because the p-value $0.0082$ is less than $\alpha = 0.05$ , reject $H_0$ . There is convincing statistical evidence that more than 50% of all customers prefer the new layout.

$z = 2.40$ and the p-value is $0.0082$ . Since $0.0082 < 0.05$ , reject $H_0$ : there is convincing statistical evidence that more than half of all customers prefer the new layout.

Frequently asked questions

What does State, Plan, Do, Conclude stand for?

It is a four-step template for inference free-response answers. State gives the parameter and the hypotheses, Plan names the procedure and checks conditions, Do shows the statistic, test statistic, and p-value or interval, and Conclude compares a test result to $\alpha$ or interprets the interval, then answers in context. Labeling the four steps helps a scorer find each component.

How many free-response questions are on the AP Statistics exam?

Section II has 4 free-response questions worth 10 points each, answered in 90 minutes, and it counts for 50% of the exam score. They appear in a fixed order, and Question 3 is the dedicated inference question, either a hypothesis test or a confidence interval.

Do I lose points for using a calculator on the FRQ?

No, a graphing calculator with statistical capabilities is expected for both sections. You do lose the work-shown point if you write only the calculator command, such as a function name, without reporting the statistic, test statistic, and p-value. Show the values the procedure produces alongside any calculator use.

Should I write hypotheses about the sample or the population?

Always write hypotheses about the population parameter, such as $p$ (the population proportion) or $\mu$ (the population mean). Hypotheses about a sample statistic like $\hat{p}$ (the sample proportion) lose credit, because inference makes claims about the population, not the sample you already measured.