When do I use the Central Limit Theorem versus just saying the population is Normal?

If the problem states or implies the population itself is Normal, then x is Normal for any sample size and you do not need the CLT. Only invoke the CLT when the population is non-Normal or unknown and your sample size is large (typically n≥ 30). Say which case applies — readers award the shape point for the correct justification.

Do I need to check the 10% condition on every sampling-distribution FRQ?

Check it whenever sampling is done without replacement and the problem provides a population size. The condition n≤ 0.10N keeps the standard-deviation formula valid. If the population size is not given or the problem says sampling is with replacement, you can note the condition is reasonably met or not needed, but always mention it when the numbers are available.

Should I round the standard deviation before computing the z-score?

Keep extra decimal places in intermediate steps and only round at the end. Rounding _p or _x too early can shift your z-score enough to change the probability and cost you the final point. Carry at least three or four significant figures until your last calculation.

What is the difference between the standard deviation of the population and the standard deviation of the sample mean?

The population standard deviation describes variability among individual values. The standard deviation of the sample mean, _x=/sqrt(n), describes variability of the statistic x across many samples and is always smaller. Use for questions about a single observation and /sqrt(n) for questions about a sample mean.

AP-STATS-5-FRQ

U5 FRQ Practice

Master AP Statistics Unit 5 free-response questions on sampling distributions of p-hat and x-bar with a step-by-step attack plan and worked examples.

What you'll do in this lesson

A voice-first session with the Crimsora tutor on U5 FRQ Practice, then targeted practice and FRQs — with the tutor adapting to where you get stuck.

Start U5 FRQ Practice free

What this lesson covers

Unit 5 free-response questions ask you to reason about sampling distributions — the distributions of statistics like

\hat{p}

and

\bar{x}

across all possible samples. On the exam, these questions reward students who state conditions clearly, compute the correct mean and standard deviation, and interpret probabilities in context.

This guide is not about re-learning the formulas — you already met them in U5.1, U5.3, U5.5, and U5.7. Instead, it shows you how to package that knowledge into responses that earn every available point. You will learn a repeatable attack sequence, the exact language readers look for, and the mistakes that quietly cost points.

The Unit 5 FRQ Attack Plan

Almost every Unit 5 free-response part follows a hidden structure. Train yourself to hit these stages in order so nothing gets skipped under time pressure.

Step	What to do	What earns the point
1. Identify the statistic	Decide if the question is about $\hat{p}$ (proportion) or $\bar{x}$ (mean)	Correct parameter and symbol
2. State the shape	Justify Normal via CLT or the large-counts rule	Conditions checked with numbers
3. Center	$\mu_{\hat{p}}=p$ or $\mu_{\bar{x}}=\mu$	Correct mean value
4. Spread	Plug into the SD formula	Correct standard deviation
5. Compute	Standardize with a $z$ -score and find the probability	Correct probability
6. Interpret	Answer the actual question in context	Conclusion tied to the scenario

The most common lost point is Step 2. Writing "the distribution is Normal" is not enough — you must show why. For proportions, verify

np\ge 10

and

n(1-p)\ge 10

using actual numbers. For means, cite the Central Limit Theorem when

n\ge 30

, or state that the population is already Normal.

Always label your final probability statement in words: "There is about a 0.04 probability that a random sample of 50 would have a sample proportion this high or higher."

Sampling Distribution of p-hat on the FRQ

When a question describes counting successes — defective parts, voters, people who prefer a brand — you are working with

\hat{p}

. The reader wants three things spelled out.

First, the center:

\mu_{\hat{p}}=p

, the true population proportion. Second, the spread:

\sigma_{\hat{p}}=\sqrt{\dfrac{p(1-p)}{n}}

. Third, the shape justification via the large-counts condition: both

np\ge 10

and

n(1-p)\ge 10

.

A subtle exam trap involves the 10% condition. Because sampling is usually done without replacement, the standard-deviation formula is only valid when the sample is no more than 10% of the population. Many FRQs award a point for stating

n\le 0.10N

. If the problem gives you a population size, mention it.

When you standardize, use

z=\dfrac{\hat{p}-p}{\sigma_{\hat{p}}}

. Keep at least three decimal places in intermediate work; rounding

\sigma_{\hat{p}}

too early can push a

z

-score across a boundary and change your probability. Finish with a contextual sentence, not just a decimal.

A misconception to avoid: the sampling distribution of

\hat{p}

is not the distribution of the data. The data are 0s and 1s (success/failure); the sampling distribution describes the statistic

\hat{p}

over many samples. Confusing these two leads to using the wrong standard deviation.

Sampling Distribution of x-bar on the FRQ

When the scenario involves a measured quantity — weights, times, temperatures — the statistic is

\bar{x}

. The center is

\mu_{\bar{x}}=\mu

and the spread is

\sigma_{\bar{x}}=\dfrac{\sigma}{\sqrt{n}}

.

Shape justification splits into two cases you must argue explicitly.

Situation	Justification to write
Population stated Normal	"Since the population is Normal, $\bar{x}$ is Normal for any $n$ ."
Population not Normal or unknown, $n\ge 30$	"By the Central Limit Theorem, with $n\ge 30$ the sampling distribution of $\bar{x}$ is approximately Normal."
Population not Normal, small $n$	You cannot assume Normal; say so.

Standardize with

z=\dfrac{\bar{x}-\mu}{\sigma/\sqrt{n}}

. Notice the denominator uses

\sigma/\sqrt{n}

, not

\sigma

. Forgetting the

\sqrt{n}

is the single most frequent arithmetic error on mean FRQs; it makes the spread far too large and the probability far too moderate.

Expect a two-part structure: one part about a single observation (uses

\sigma

) and one part about a sample mean (uses

\sigma/\sqrt{n}

). The contrast is deliberate — readers want to see you know that averaging reduces variability. State that idea in words if asked to compare: larger samples produce sample means clustered more tightly around

\mu

Writing Answers That Earn Points

AP readers score communication, not just arithmetic. A correct number with a vague sentence can still lose the interpretation point.

Use full subscript notation so the reader knows exactly which quantity you mean: write

\mu_{\bar{x}}

and

\sigma_{\bar{x}}

, not just "mean" and "SD." Show the formula, then the substitution, then the result — three lines. If you only write the final number and it is wrong, you get zero; if you show substitution, a single arithmetic slip may still earn partial credit.

When a question asks whether an outcome is "unusual" or "surprising," you must connect a computed probability to a judgment. For example: "Because the probability of a sample mean this extreme is only about 0.02, which is small, this result would be surprising if the claimed mean were true."

Never answer a probability question with just a calculator command like normalcdf output stripped of context. Define your variable, state the distribution and its parameters, then report the probability. Finally, reread the prompt's last sentence — that is the exact question you must answer. Many students compute a beautiful

z

-score and forget to state the conclusion the problem requested.

Key terms

Sampling distribution.: The distribution of a statistic (such as $\hat{p}$ or $\bar{x}$ ) over all possible samples of a fixed size $n$ from a population.
Standard deviation of the statistic.: A measure of how much the statistic varies from sample to sample; $\sigma_{\hat{p}}=\sqrt{p(1-p)/n}$ for proportions and $\sigma_{\bar{x}}=\sigma/\sqrt{n}$ for means.
Large-counts condition.: The requirement $np\ge 10$ and $n(1-p)\ge 10$ that justifies approximating the sampling distribution of $\hat{p}$ as Normal.
Central Limit Theorem.: The result that the sampling distribution of $\bar{x}$ becomes approximately Normal as $n$ grows, typically for $n\ge 30$ , regardless of population shape.
10% condition.: When sampling without replacement, the sample size should be no more than 10% of the population so the standard-deviation formula stays valid.
z-score (standardizing).: The number of standard deviations a statistic lies from its mean, computed as $z=\frac{\text{statistic}-\text{mean}}{\text{SD of statistic}}$ .
Unbiased estimator.: A statistic whose sampling distribution is centered at the true parameter, so $\mu_{\hat{p}}=p$ and $\mu_{\bar{x}}=\mu$ .

Worked example

A bottling machine fills bottles with a mean of 500 mL and a standard deviation of 8 mL. The fill amounts are approximately Normal. A quality inspector selects a random sample of 16 bottles. (a) Describe the sampling distribution of the sample mean fill amount. (b) Find the probability that the sample mean is less than 496 mL. (c) Interpret this probability.

Part (a): The statistic is

\bar{x}

. Center:

\mu_{\bar{x}}=\mu=500

mL. Spread:

\sigma_{\bar{x}}=\dfrac{\sigma}{\sqrt{n}}=\dfrac{8}{\sqrt{16}}=\dfrac{8}{4}=2

mL. Shape: because the population of fill amounts is approximately Normal, the sampling distribution of

\bar{x}

is approximately Normal for any sample size, so we do not need the Central Limit Theorem here.

So

\bar{x}

is approximately Normal with mean 500 mL and standard deviation 2 mL.

Part (b): Standardize the value 496.

z=\dfrac{496-500}{2}=\dfrac{-4}{2}=-2

. We want

P(\bar{x}<496)=P(z<-2)

. From the standard Normal distribution,

P(z<-2)\approx 0.0228

.

Part (c): There is about a 0.0228 probability that a random sample of 16 bottles has a mean fill amount below 496 mL. Because this probability is small, observing such a sample mean would be somewhat surprising if the machine truly averages 500 mL — it could signal the machine is underfilling.

Practice questions

A polling firm knows that 60% of a large city supports a ballot measure. In a random sample of 100 residents, which value is closest to the standard deviation of the sampling distribution of

\hat{p}

0.0024
0.049
0.24
0.60

Answer: 0.049

Use

\sigma_{\hat{p}}=\sqrt{\frac{p(1-p)}{n}}=\sqrt{\frac{0.60\times 0.40}{100}}=\sqrt{0.0024}\approx 0.049

. The value 0.0024 is the variance (before taking the square root), and 0.24 is

p(1-p)

without dividing by

n

or rooting. The sample size is well under 10% of a large city and

np=60\ge10

n(1-p)=40\ge10

, so the Normal model applies.

A machine produces bolts whose lengths have mean 4.0 cm and standard deviation 0.10 cm, but the distribution of lengths is right-skewed. An engineer takes a random sample of 49 bolts. Describe the sampling distribution of

\bar{x}

and find the probability that the sample mean exceeds 4.03 cm. Show your reasoning.

Answer: Approximately Normal with mean 4.0 cm and SD $0.10/\sqrt{49}\approx 0.0143$ cm; $P(\bar{x}>4.03)\approx 0.018$ .

Even though the population is right-skewed,

n=49\ge 30

, so by the Central Limit Theorem the sampling distribution of

\bar{x}

is approximately Normal. Center:

\mu_{\bar{x}}=4.0

. Spread:

\sigma_{\bar{x}}=0.10/\sqrt{49}=0.10/7\approx 0.0143

. Standardize:

z=\frac{4.03-4.0}{0.0143}\approx 2.10

, giving

P(z>2.10)\approx 0.018

. Full credit requires naming the CLT (because the population is skewed), showing the SD uses

\sqrt{n}

, and stating the probability in context.

Explain why a sample mean based on

n=100

is more likely to fall within 1 mL of the population mean than a sample mean based on

n=25

, assuming the same population.

Answer: Because $\sigma_{\bar{x}}=\sigma/\sqrt{n}$ decreases as $n$ increases, the sampling distribution for $n=100$ is narrower, so more of its area lies within 1 mL of the mean.

The center of the sampling distribution is

\mu

for both sample sizes, so neither is biased. What changes is the spread: with

n=100

the standard deviation is

\sigma/10

, while with

n=25

it is

\sigma/5

— twice as large. A tighter distribution places more probability near

\mu

, so the larger sample is more likely to land within any fixed distance of the true mean. This is the core intuition behind why larger samples give more precise estimates.

FAQ

When do I use the Central Limit Theorem versus just saying the population is Normal?: If the problem states or implies the population itself is Normal, then $\bar{x}$ is Normal for any sample size and you do not need the CLT. Only invoke the CLT when the population is non-Normal or unknown and your sample size is large (typically $n\ge 30$ ). Say which case applies — readers award the shape point for the correct justification.
Do I need to check the 10% condition on every sampling-distribution FRQ?: Check it whenever sampling is done without replacement and the problem provides a population size. The condition $n\le 0.10N$ keeps the standard-deviation formula valid. If the population size is not given or the problem says sampling is with replacement, you can note the condition is reasonably met or not needed, but always mention it when the numbers are available.
Should I round the standard deviation before computing the z-score?: Keep extra decimal places in intermediate steps and only round at the end. Rounding $\sigma_{\hat{p}}$ or $\sigma_{\bar{x}}$ too early can shift your $z$ -score enough to change the probability and cost you the final point. Carry at least three or four significant figures until your last calculation.
What is the difference between the standard deviation of the population and the standard deviation of the sample mean?: The population standard deviation $\sigma$ describes variability among individual values. The standard deviation of the sample mean, $\sigma_{\bar{x}}=\sigma/\sqrt{n}$ , describes variability of the statistic $\bar{x}$ across many samples and is always smaller. Use $\sigma$ for questions about a single observation and $\sigma/\sqrt{n}$ for questions about a sample mean.

Learn this with a teacher, not a page

The Crimsora tutor teaches U5 FRQ Practice live — explaining on a whiteboard, asking you questions, and adapting to where you get stuck.

Try a live session free