Introducing Confidence Intervals for Proportions — AP Statistics Study Guide

For: AP Statistics candidates sitting AP Statistics.

Covers: Point estimation for population proportions, confidence level definition, one-proportion z-interval formula, conditions for inference, interval calculation, and correct interpretation of confidence intervals for proportions, aligned with AP Statistics CED Unit 6.

You should already know: Sampling distribution behavior for sample proportions, properties of the standard normal distribution, how to calculate z-scores for given percentiles.

A note on the practice questions: All worked questions in the "Practice Questions" section below are original problems written by us in the AP Statistics style for educational use. They are not reproductions of past College Board / Cambridge / IB papers and may differ in wording, numerical values, or context. Use them to practise the technique; cross-check with official mark schemes for grading conventions.

1. What Is Introducing Confidence Intervals for Proportions?

A confidence interval for a population proportion is an inference technique that produces a range of plausible values for an unknown population proportion $p$, based on data collected from a random sample. This is the first core inference topic for proportions in AP Statistics CED Unit 6, which accounts for 12-15% of the total AP exam weight. This topic appears in both multiple-choice (MCQ) and free-response (FRQ) sections, typically as 1-2 standalone MCQ questions or 1-2 parts of a longer FRQ worth 2-4 total points.

Notation conventions are standardized for this topic: $p$ = the fixed, unknown true proportion of the population; $\hat{p}$ = the sample proportion, calculated from sample data; $n$ = sample size; $z^{\ast }$ = critical z-value for a given confidence level. Common synonyms include one-proportion z-interval and confidence interval for a population proportion. Unlike a point estimate (a single guess for $p$), a confidence interval explicitly quantifies sampling variability, giving a range that accounts for the fact that sample results naturally vary from sample to sample. This topic is the foundation for all subsequent inference on categorical proportions.

2. Confidence Intervals: Structure and Interpretation

All confidence intervals follow the same core structure: a point estimate plus or minus a margin of error. The point estimate is our best single guess for the unknown population parameter, and the margin of error accounts for random sampling variability. For a population proportion, the general form of the confidence interval is: $$\hat{p}\pm z^{\ast }\times SE(\hat{p})$$ where $SE(\hat{p})=\sqrt{\frac{\hat{p}(1-\hat{p})}{n}}$ is the standard error of the sample proportion, and $z^{\ast }$ is the critical value from the standard normal distribution corresponding to our chosen confidence level. Common $z^{\ast }$ values you should memorize for the AP exam are 1.645 for 90% confidence, 1.96 for 95% confidence, and 2.576 for 99% confidence.

The confidence level describes the long-run behavior of the interval method, not a single interval. A 95% confidence level means that if we repeated the sampling process many times, 95% of the resulting intervals would capture the true population proportion. To interpret a single calculated interval, the AP-expected phrasing is: We are C% confident that the true proportion of [context] is between lower bound and upper bound.

Worked Example

A political scientist asks 250 randomly selected registered voters whether they support a new local park bond, and finds 140 support it. (a) Identify the point estimate for the true proportion of all registered voters who support the bond. (b) Interpret what a 95% confidence level means in this context.

1. Confirm the goal: We need a point estimate and a context-specific interpretation of the 95% confidence level.
2. Calculate the point estimate: $\hat{p}=\frac{\text{Number of supporters}}{\text{Sample size}}=\frac{140}{250}=0.56$. This is our point estimate for the true population proportion $p$.
3. Recall that the confidence level describes the long-run success rate of the method, not a single interval.
4. State the interpretation in context: If we took many random samples of 250 registered voters from this population, about 95% of the resulting confidence intervals would capture the true proportion of all registered voters who support the park bond.

Exam tip: You will almost always lose points on the AP exam if you interpret a 95% confidence interval as "there is a 95% probability the true proportion is in this interval." The true $p$ is fixed, so it is either in the interval or not — always interpret confidence level as the long-run success rate of the method.

3. Conditions for Inference

Before you can reliably calculate a confidence interval for a proportion, you must verify three conditions to ensure that the sampling distribution of $\hat{p}$ is approximately normal and that our standard error calculation is valid. Skipping condition checks is one of the most common reasons for lost points on AP FRQs. The three required conditions are:

1. Random: The sample must be randomly selected from the population of interest. This ensures $\hat{p}$ is an unbiased estimator of $p$.
2. Independent: Individual observations must be independent. When sampling without replacement from a finite population, we use the 10% condition: the sample size $n$ must be no more than 10% of the total population size $N$. This ensures that sampling without replacement does not meaningfully change the probability of success for subsequent observations.
3. Large Counts: The sampling distribution of $\hat{p}$ is approximately normal only if we have at least 10 expected successes and 10 expected failures. We check this with our sample data: $n\hat{p}\ge 10$ and $n(1-\hat{p})\ge 10$ (the AP CED uses a threshold of 10, so stick with this for the exam).

If all three conditions are met, we can safely use the normal approximation to the sampling distribution of $\hat{p}$ to construct our z-interval.

Worked Example

A quality control inspector tests 40 randomly selected lightbulbs from a shipment of 5000, and finds 6 are defective. Check the conditions for constructing a confidence interval for the true proportion of defective bulbs in the shipment.

1. Check Random condition: The problem explicitly states the 40 bulbs were randomly selected, so this condition is satisfied.
2. Check 10% condition for independence: Total population size is 5000. 10% of 5000 is 500, and our sample size of 40 is less than 500. The 10% condition is satisfied, so we can assume independence.
3. Check Large Counts condition: First calculate $\hat{p}=\frac{6}{40}=0.15$. Then compute $n\hat{p}=40(0.15)=6$ and $n(1-\hat{p})=40(0.85)=34$.
4. Final conclusion: The Large Counts condition fails because $n\hat{p}=6<10$. We cannot safely construct a one-proportion z-interval for this data.

Exam tip: AP FRQs require you to explicitly name each condition, not just check the numbers. Always write "Random", "10% Condition", and "Large Counts" by name, and show your check for each, even if you have to state that a condition is assumed to be met.

4. Calculating a One-Proportion Z-Interval

Once you have confirmed all conditions are met, you can calculate the confidence interval using the standard one-proportion z-interval formula: $$\hat{p}\pm z^{\ast }\sqrt{\frac{\hat{p}(1-\hat{p})}{n}}$$ The term $z^{\ast }\sqrt{\frac{\hat{p}(1-\hat{p})}{n}}$ is the margin of error ($ME$), which measures how far we expect $\hat{p}$ to be from the true $p$ at our chosen confidence level. After calculating the lower and upper bounds, you must interpret the interval in context to earn full credit on FRQs.

Worked Example

A biologist wants to estimate the proportion of oak trees in a national forest that are infected with a certain fungus. They take a random sample of 180 oak trees, and find 63 are infected. Calculate and interpret a 90% confidence interval for the true proportion of infected oak trees.

1. Verify conditions (required before calculation): Random sample given, population of oak trees is far larger than $10\times 180=1800$ so 10% condition is met, $n\hat{p}=63\ge 10$ and $n(1-\hat{p})=117\ge 10$ so Large Counts is met. All conditions satisfied.
2. Calculate sample proportion: $\hat{p}=\frac{63}{180}=0.35$.
3. Find $z^{\ast }$ for 90% confidence: $z^{\ast }=1.645$.
4. Calculate margin of error: $ME=1.645\times \sqrt{\frac{(0.35)(0.65)}{180}}\approx 1.645\times 0.0356\approx 0.0585$.
5. Construct the interval: $0.35\pm 0.0585=(0.2915,0.4085)$, rounded to $(0.29,0.41)$.
6. Interpret in context: We are 90% confident that the true proportion of all oak trees in this national forest that are infected with the fungus is between 0.29 and 0.41.

Exam tip: Do not round early during calculation. Keep extra decimals for standard error and margin of error, only round your final interval bounds to 2-3 decimal places to avoid rounding error that can cost points.

5. Common Pitfalls (and how to avoid them)

• Wrong move: Interpreting a 95% confidence interval as "There is a 95% probability the true proportion $p$ is between 0.29 and 0.41." Why: Students confuse the long-run behavior of the method with probability for a single fixed interval; the true $p$ is fixed, not random, so it is either in the interval or not. Correct move: Phrase interpretations as "We are 95% confident that the true proportion of [context] is between [lower, upper]." Reserve the "95% of samples capture $p$" phrasing for interpreting the confidence level itself.
• Wrong move: Forgetting to check the 10% condition, only checking Random and Large Counts. Why: Students often dismiss the 10% condition as unimportant and skip it to save time. Correct move: Always list and check all three conditions by name on every FRQ; even if you have to assume the 10% condition is met, state that assumption explicitly to earn the point.
• Wrong move: Using $p=0.5$ to check the Large Counts condition instead of $\hat{p}$. Why: Students remember $p(1-p)$ is maximized at 0.5 from sampling distribution topics and incorrectly use it for condition checks. Correct move: Always use your calculated sample proportion $\hat{p}$ to compute $n\hat{p}$ and $n(1-\hat{p})$ for the Large Counts condition.
• Wrong move: Using approximate z-values (e.g. $z^{\ast }=2$ for 95% confidence) instead of precise critical values on FRQs. Why: Students rely on the empirical rule and use rounded values instead of the precise critical values AP expects. Correct move: Memorize and use the precise values: 1.645 for 90%, 1.96 for 95%, 2.576 for 99%.
• Wrong move: Claiming a 95% confidence interval means 95% of sample proportions lie in the interval. Why: Students confuse the range of the sampling distribution with a confidence interval for the population parameter. Correct move: A confidence interval estimates the fixed unknown population proportion, so always frame your interpretation around the population parameter, not sample results.

6. Practice Questions (AP Statistics Style)

Question 1 (Multiple Choice)

A researcher estimates the proportion of college students who work part-time. They take a random sample of 100 students and construct a 95% confidence interval of (0.62, 0.78). Which of the following statements is a correct interpretation of this interval? A) 95% of all college students work part-time between 62% and 78% of the time. B) There is a 95% probability that the true proportion of all college students who work part-time is between 0.62 and 0.78. C) We are 95% confident that the true proportion of all college students who work part-time is between 0.62 and 0.78. D) 95% of all samples of size 100 will have a sample proportion between 0.62 and 0.78.

Worked Solution: Eliminate incorrect options one by one. Option A misinterprets the parameter: the interval estimates the proportion of students who work part-time, not how much time individual students work, so A is wrong. Option B uses incorrect probability language for a fixed interval, so B is wrong. Option D describes the sampling distribution of sample proportions, not a confidence interval for the true population proportion, so D is wrong. Option C matches the AP-approved correct interpretation of a confidence interval. Correct answer: C.

Question 2 (Free Response)

A cereal company wants to estimate the proportion of boxes that have less cereal than the advertised 12 ounces, due to manufacturing variation. They randomly sample 75 boxes from a day's production run of 12000 boxes, and find 12 boxes have underweight cereal. (a) Check all conditions for constructing a 99% confidence interval for the true proportion of underweight boxes in the day's production. (b) Calculate the 99% confidence interval for the true proportion. (c) The company's quality control standard says that if the true proportion of underweight boxes is above 0.20, they need to adjust the machinery. Based on your interval, does the company need to adjust the machinery? Justify your answer.

Worked Solution: (a) 1. Random: The problem states the sample was randomly selected, so the Random condition is met. 2. 10% Condition: Population size is 12000, 10% of 12000 is 1200. Sample size 75 < 1200, so the 10% condition for independence is met. 3. Large Counts: $\hat{p}=12/75=0.16$, $n\hat{p}=12\ge 10$, $n(1-\hat{p})=63\ge 10$, so Large Counts is met. All conditions are satisfied. (b) Use the one-proportion z-interval formula: $\hat{p}\pm z^{\ast }\sqrt{\frac{\hat{p}(1-\hat{p})}{n}}$. For 99% confidence, $z^{\ast }=2.576$. Standard error = $\sqrt{\frac{(0.16)(0.84)}{75}}\approx 0.0423$. Margin of error = $2.576(0.0423)\approx 0.109$. Interval = $0.16\pm 0.109=(0.051,0.269)$. (c) The company does not need to adjust the machinery. Our interval includes plausible values for $p$ both below 0.20 (0.051 to 0.20) and above 0.20 (0.20 to 0.269). Because 0.20 is a plausible value for the true proportion, we do not have sufficient evidence to conclude the true proportion is above 0.20.

Question 3 (Application / Real-World Style)

Public health researchers want to estimate the proportion of adults in a city who have received at least one dose of a flu vaccine for the current season. They obtain a random sample of 320 adults with valid vaccine records from the city's population, and find that 184 have received at least one dose. Calculate a 95% confidence interval for the true proportion of vaccinated adults in the city, and interpret the result in context.

Worked Solution: First verify conditions: Random sample given, city population is far larger than 10320=3200 so 10% condition is met. $\hat{p}=184/320=0.575$, $n\hat{p}=184\ge 10$, $n(1-\hat{p})=136\ge 10$, so all conditions are satisfied. For 95% confidence, $z^=1.96$. Calculate the interval: $$0.575\pm 1.96\sqrt{\frac{(0.575)(0.425)}{320}}=0.575\pm 0.054$$ The 95% confidence interval is (0.521, 0.629). Interpretation: We are 95% confident that the true proportion of all adults in this city who have received at least one dose of the current season's flu vaccine is between 0.52 and 0.63.

7. Quick Reference Cheatsheet

8. What's Next

This topic is the foundation for all inference on proportions, and the core logic of interval estimation transfers to all other confidence intervals you will learn in AP Statistics. Next, you will learn how to calculate the required sample size to achieve a desired margin of error for a proportion, then move to significance testing for one proportion, followed by confidence intervals and significance tests for comparing two proportions. Without mastering the conditions, calculation, and interpretation of one-proportion confidence intervals, all subsequent inference topics for proportions will be extremely difficult, as they build directly on the core skills you learned here. These skills are tested repeatedly across the AP exam, and transfer to confidence intervals for means later in the course.

• Margin of Error and Sample Size for Proportions
• Significance Tests for One Proportion
• Confidence Intervals for Two Proportions
• Inference for Categorical Data: Chi-Square Tests