OttoLearn

L-INF-1: Sampling Distributions and the Central Limit Theorem

Module 4 · Statistical Inference

Section 1: Introduction

▾

A tag such as (C2 + C3) after a title names the Core Concepts parts the problem draws on, so you know where to look back.

A school board ranks its schools by their students’ average score on the same provincial exam. When the results come out, a pattern jumps off the page: of the ten highest-scoring schools, nine are small, with fewer than 150 students tested. A board member proposes splitting the large schools into smaller ones.

Before you read on, make a prediction. If you looked at the ten lowest-scoring schools, what kind of schools would you expect to find there: small ones, large ones, or a mix?

Pause and write. Write your prediction and one sentence of reasoning on paper or in your notes. You will test it in C4, where Figure 6 shows the whole district.

The puzzle has the same root as a very ordinary question. Imagine you work in quality control at a factory that fills cereal boxes. Each box is supposed to contain 500 g of cereal. You can’t weigh every single box — there are thousands per hour. So you pull a random sample of 40 boxes, weigh them, and compute the average.

Your sample mean comes out to 497 g. Should you sound the alarm? Is the machine underfilling? Or is 497 g just normal random variation you’d expect to see even when the machine is working perfectly?

To answer that question, you need to understand something that isn’t obvious at first: the average of a random sample is itself random. If you took a different sample of 40 boxes, you’d get a slightly different average. Take 1,000 different samples, and you’d get 1,000 different averages — and those averages would follow their own distribution.

That distribution — the distribution of all possible sample means — is called the sampling distribution. The Central Limit Theorem tells us when that distribution can be treated as approximately normal, which lets us judge whether an observed average is ordinary or surprising.

After this lesson, you will be able to:

  • Explain what a sampling distribution is and how it differs from a population distribution
  • Describe the centre and standard error of the sampling distribution of and explain how sample size changes its spread
  • Decide whether a normal model for is justified by the population shape, sampling process, and sample size
  • Compute probabilities for sample means by standardizing:

This is Lesson L-INF-1 — the first lesson in the Statistical Inference module. It combines the distribution and z-score work from L-DS-5 and L-PR-6. Everything that follows—confidence intervals, hypothesis tests, and regression inference—builds on this shift from individual observations to repeated sample statistics.

Section 2: Prerequisites

▾

A tag such as (C2 + C3) after a title names the Core Concepts parts the problem draws on, so you know where to look back.

Statistical inference is built on distribution, variability, and probability ideas from earlier modules.

From L-DS-4 — spread: Standard deviation describes the typical distance of individual values from their mean. Today it becomes the starting point for describing the spread of sample means.

From L-DS-5 — z-scores: A z-score measures distance from a mean in standard-deviation units. Today the same structure uses a different yardstick: standard error.

From L-PR-6 — normal probabilities: Left-tail areas and complements translate z-scores into probabilities. Today those probabilities describe sample means rather than individuals.

From L-PR-1 — random sampling: A probability-based sampling process supports independence and representative inference. Today that condition determines whether the sampling-distribution model is trustworthy.

Retrieval Checkpoint

A population has mean and standard deviation . One observation has value . What is its z-score?

Bridge to today: Earlier z-scores use because they locate one observation. Today the object is a sample mean, which needs its own yardstick; in C2 you will predict how large that yardstick is before you see the formula. Throughout L-INF-1, is treated as known; L-INF-3 introduces the t-distribution when it must be estimated with .

Section 3: Core Concepts

▾

A tag such as (C2 + C3) after a title names the Core Concepts parts the problem draws on, so you know where to look back.

How the five concepts fit together:

  1. C1: identify the new random object—the sample mean across repeated samples;
  2. C2: locate its centre and measure its spread with standard error;
  3. C3: decide whether a normal shape is justified;
  4. C4: predict what changes when sample size changes; and
  5. C5: standardize an observed sample mean and compute its probability.

C1 — The Sampling Distribution of

Here’s the key idea, stated plainly: a sample mean is a random variable.

When you take a random sample of size from a population and compute , the value you get depends on which observations happened to land in your sample. A different random sample would give a different . Because varies from sample to sample, it has a distribution.

See it before you name it. The figure below is a sampling machine: it draws a random sample from a population, computes the sample’s mean, and drops that mean into panel ③.

Colour key for every figure in this lesson: blue shows the population (individual values), green shows the data in one sample, and orange shows sample means .

Population:
Sampling machine: population, latest sample, and sample means Three stacked panels share one axis, with a dashed line at the population mean. Panel 1 shows the right-skewed population, mean 22 and standard deviation 18. Panel 2 will show the latest sample. Panel 3, sample means so far, is empty until you draw samples.
Available after 30 samples (0 so far).

Figure 1: Press Draw 1 sample a few times, then Draw 100 samples. Compare how wide panel ③ spreads with panel ①. Panel ③ is built from sample means — one dot per sample — not from individual data values.

Draw one sample at a time first: panel ② changes completely with every draw, while panel ① never changes. Then draw 100 samples and compare the spread of the orange dots in panel ③ with the spread of the population in panel ①. Panel ③ is the new object this lesson is about, and it has a name.

Sampling Distribution of the Sample Mean

The sampling distribution of is the probability distribution of all possible values of that could arise from repeatedly drawing random samples of size under the same specified sampling process.

Think of it as: if we could repeat the sampling process infinitely many times, each time computing the sample mean, what would the histogram of those means look like?

This is different from the population distribution (which describes individual observations) and from the sample distribution (the data in one specific sample). The sampling distribution lives at a higher level — it describes the behaviour of the statistic across many repetitions.

The sampling distribution is not the distribution of the data in your sample. It is the distribution of values you would get if you repeated your study many times. You never actually observe the sampling distribution directly — it is a theoretical construct that tells you how reliable your estimate is.

C2 — Mean and Standard Error

The sampling distribution has two key parameters. We write to mean the long-run mean of the sample means, and to mean their standard deviation, also called the standard error.

Before any formula, commit to a guess. One box from a filling line varies with a standard deviation of 12 g. What happens to the averages of 36 boxes? Shape the curve you expect, then reveal the answer.

Graphical Prediction
Predict the sampling distribution before revealing it
A filling process is exactly normal with mean 500 g and population standard deviation 12 g. From a very large production run, teams repeatedly take random, independent samples of 36 boxes and record each sample mean. Shape the curve you expect for those sample means.
Coordinate graph comparing the learner prediction with the observed relationship.

Figure 2: Predict the centre and spread before revealing the sampling distribution. The dashed curve records your expectation; the solid curve shows the exact normal model for sample means. Compare both position and width, then explain which one averaging changes.

Whatever you predicted, the reveal shows the two facts this concept is about: averaging leaves the centre where it was and shrinks the spread. Here is why, in symbols.

Mean of the Sampling Distribution

The mean of the sampling distribution equals the population mean . This says that is an unbiased estimator of : on average, sample means hit the target.

This makes intuitive sense — if you average many sample averages, you should get close to the true population average. The remarkable thing is that this is exactly true, not approximately.

The second parameter is the spread — how much do sample means vary around ?

Standard Error of the Mean

The standard error (SE) of the mean measures the typical distance between a sample mean and the population mean . It is not the same as the population standard deviation — it shrinks as sample size increases.

The formula is exact for independent draws, including sampling with replacement. For a random sample drawn without replacement from a finite population, the exact SE also includes the finite-population correction from Challenge 2. When the sample is no more than about 10% of the population, is usually a close approximation.

Notice the in the denominator. A sample of has a standard error one-tenth that of a sample of . Larger samples give sample means that cluster more tightly around .

Why and not ? Roll dice and average them. A high roll and a low roll in the same sample partly cancel, so averages pile up near the middle. The cancelling is only partial (sometimes every die rolls high), which is why the spread falls by rather than by . Figure 3 shows the exact distributions; no simulation is involved.

One die — individual rolls Average of several dice — sample means
Averages of 1, 2, 4 and 10 dice on one height scale Exact distributions of the average roll, drawn on one shared height scale so bar areas are probabilities. One die is flat from 1 to 6 with SD about 1.71, called σ. The average of two dice is a triangle peaking at 3.5 with SD σ/√2, about 1.21. The averages of four and ten dice are ever narrower humps centred on 3.5, with SD σ/2, about 0.85, and σ/√10, about 0.54. High and low rolls partly cancel, so averages pile up near 3.5, but the spread shrinks by the square root of the number of dice, not by the number itself.

All four panels share one height scale; each bar's area is the probability of that average.

Exact mean and standard deviation of the average of k dice
Dice averaged (k) Mean Exact SD Rounded SD
One die 3.5 √(35/12) ≈ 1.708
Average of 2 dice 3.5 √(35/24) ≈ 1.208
Average of 4 dice 3.5 √(35/48) ≈ 0.854
Average of 10 dice 3.5 √(7/24) ≈ 0.540
Figure 3: One die is flat. Averaging pulls results toward 3.5 because high and low rolls partly cancel. Four dice halve the spread; ten dice divide it by √10 ≈ 3.2, not by 10 — the square root in σ/√n.

The most common error in this lesson: Using (population SD) instead of (standard error) when computing probabilities for sample means. The formula applies to a single observation. For a sample mean, you must use . Using where belongs makes your z-score times too small — a significant error.

Individual observation — spread = σ = 15 Sample mean x̄ — spread = SE = σ/√n = 5.00
Individual observations compared with sample means Two overlapping bell curves centered at μ = 50. The blue curve shows the spread of individual observations (σ = 15, fixed). The orange curve shows the spread of sample means (SE = σ/√n), which shrinks as n increases. Bracket annotations below the axis show ±σ and ±SE.
At n = 9, sample means are 3.0× more concentrated than individual values.
SE = 15/√9 = 5.00
Figure 4: The blue curve (σ = 15) never changes — it describes individual observations. The orange curve narrows as n increases, showing how sample means cluster more tightly. Drag the slider to n = 100 to see the full contrast, then press Compare with 4 × n: the dashed curve shows that four times the data only halves the spread.

Drag the slider to n = 9, 25, 100 and watch the orange curve (sample means) compress while the blue curve (individuals) stays fixed. At n = 100, a sample mean that deviates from μ by 5 units would be extremely unusual — but for an individual, a 5-unit deviation is ordinary.

C3 — The Central Limit Theorem

We’ve established that has mean and standard error . But what shape does its distribution have? This is where the Central Limit Theorem comes in — and it’s one of the most powerful results in all of statistics.

The Central Limit Theorem (CLT)

Let be a random sample from a population with mean and finite standard deviation . Then, for sufficiently large , the sampling distribution of the sample mean is approximately normal:

Three conditions must hold:

  1. Independence: Observations must be independent of one another. Sampling with replacement gives independence; for sampling without replacement, checking that is no more than about 10% of the population supports the usual approximation.
  2. Sample size and population shape: If the population is exactly normal, any produces an exactly normal sampling distribution. Otherwise, the approximation improves as grows; is a common starting rule, not a universal guarantee for every shape.
  3. Finite variance: The population must have a finite standard deviation .

What makes the CLT remarkable is what it does not require: the population does not need to be normal. It may be skewed, bimodal, or uniform. As grows, the distribution of approaches normality, but an extremely skewed or heavy-tailed population may require much more than 30 observations for a good approximation.

This is why the normal distribution appears everywhere in statistics: even when individual observations are non-normal, averages tend to be normal.

The CLT says the sampling distribution of is approximately normal — it says nothing about the distribution of individual observations becoming more normal as grows. The data in your one sample still has whatever shape the population has. It is only the distribution of the statistic (across many hypothetical samples) that becomes normal.

The rule guards against non-normality of the population — it does not rescue a violated independence condition. If your data are a time series, a cluster sample, or repeated measures on the same individuals, the CLT may not apply regardless of how large is. Independence must be verified separately, and no amount of additional data can substitute for it.

So how large is large enough? It depends on the population. Before you look, predict: which row of Figure 5 will need the largest before its sample means look normal?

  • Population (simulated draws)
  • Sample means x̄ (simulated)
  • Normal curve with the same mean and SE

Scroll sideways to see all five columns.

How large must n be before x̄ looks normal? Rows are five population shapes: normal and uniform (skew 0), moderate skew (1), strong skew (2) and extreme skew (about 6.18). Columns show the population, then simulated sample means for n = 2, 5, 30 and 100, each zoomed to its own spread with a matching normal curve. Normal means look normal at every n; uniform means are close at n = 2 (a triangle) and normal from n = 5; moderate skew is close at 5 and normal by 30; strong skew is still skewed at 5, close at 30 and normal by 100; extreme skew is still skewed at n = 100. Population x̄, n = 2 x̄, n = 5 x̄, n = 30 x̄, n = 100 Normal skew 0 ≈ normal ≈ normal ≈ normal ≈ normal Uniform skew 0 close triangle shape ≈ normal ≈ normal ≈ normal Moderate skew skew 1 still skewed skew 0.71 close skew 0.45 ≈ normal skew 0.18 ≈ normal skew 0.10 Strong skew skew 2 still skewed skew 1.41 still skewed skew 0.89 close skew 0.37 ≈ normal skew 0.20 Extreme skew skew 6.18 still skewed skew 4.37 still skewed skew 2.77 still skewed skew 1.13 still skewed skew 0.62

Badge: skew of x̄ = population skew ÷ √n. “≈ normal” up to 0.25, “close” up to 0.5, otherwise “still skewed”.

Every histogram is simulated: 4,000 draws or sample means per cell, from a fixed seed, so the picture is the same on every visit.

Shape of the sample mean x̄ by population and sample size
Population Population skew n = 2n = 5n = 30n = 100
Normal 0 ≈ normal≈ normal≈ normal≈ normal
Uniform 0 close (triangle shape)≈ normal≈ normal≈ normal
Moderate skew 1 still skewed (skew 0.71)close (skew 0.45)≈ normal (skew 0.18)≈ normal (skew 0.10)
Strong skew 2 still skewed (skew 1.41)still skewed (skew 0.89)close (skew 0.37)≈ normal (skew 0.20)
Extreme skew 6.18 still skewed (skew 4.37)still skewed (skew 2.77)still skewed (skew 1.13)still skewed (skew 0.62)
Figure 5: Each cell is zoomed to its own spread, so only shape is compared; badges show how skewed x̄ still is. Moderate skew is fine by n = 30, strong skew needs about 100, and extreme skew is still skewed at 100. n ≥ 30 is a rule of thumb, not magic.

Read the grid row by row. A normal population gives normal sample means at every . Uniform and moderately skewed populations get there quickly. A strongly skewed population needs about 100 observations, and an extremely skewed one (incomes, insurance claims) is still visibly skewed at . That is why is a starting rule, not a guarantee.

C4 — Effect of Sample Size on the Sampling Distribution

Two things happen as you increase sample size :

  1. The standard error shrinks: decreases. Sample means cluster more tightly around .
  2. The normal approximation improves: Even for very non-normal populations, larger samples produce a more bell-shaped distribution of .

A concrete example: suppose .

Sample size Standard error Interpretation
No averaging benefit
SE is one-third of
SE is one-sixth of
SE is one-twelfth of

Notice the diminishing returns: doubling only reduces SE by a factor of . To halve the SE, you need to quadruple (press Compare with 4 × n in Figure 4 to see it). This is why there’s a cost-benefit tradeoff in sampling design.

Back to the school ranking. A school’s average is a sample mean, so its standard error is , where is the number of students tested. Small schools have large standard errors: their averages wander further from the overall mean, in both directions. Check your prediction from the introduction with Figure 6.

One simulated school Top 10 averages Bottom 10 averages Funnel 500 ± 2 SE
School average score against number of students tested Scatter plot of 200 simulated schools. Every student's score comes from one population with mean 500 and standard deviation 100. Small schools scatter widely above and below 500; large schools sit close to 500.

What the funnel shows: the dashed lines sit at 500 plus or minus 2 standard errors, where the standard error is 100 divided by the square root of the number of students tested. At 25 students the lines run from 460 to 540; at 100 students, from 480 to 520; at 400 students, from 490 to 510. Averages of few students wander further from 500, so the highest and the lowest averages in the district both come mostly from small schools, not because those schools are better or worse.

200 simulated schools. Use the buttons to highlight the extremes.

Figure 6: Every school in this simulated district draws its students from the same population, yet small schools fill both the top and the bottom of the ranking. The funnel, 500 ± 2·SE with SE = 100/√n, shows why: averages from small samples wander further from 500.

Every school in this simulated district draws its students from the same population, so none is truly better, yet small schools fill both ends of the ranking. This pattern was not invented for this lesson. In the early 2000s a large US education foundation invested heavily in creating smaller schools, partly because small schools topped the rankings. Statistician Howard Wainer later called “the most dangerous equation” (American Scientist, 2007) because ignoring it leads to decisions like this one.

Doubling the sample size does not halve the standard error. Because sits under a square root, halving SE requires four times as many observations.

C5 — Computing Probabilities for

Once the sampling process and population shape justify a normal model, we can use the z-table to compute probabilities. We write for the standardized location of a sample mean. The only adjustment from L-PR-6 is the denominator in the z-score formula.

Z-Score for a Sample Mean

This converts a sample mean into a standard normal z-score. The denominator is the standard error , not .

Once you have , use the standard normal table exactly as in L-PR-6: is the left-tail area.

The workflow for any probability question about :

  1. Compute the standard error:
  2. Compute the z-score:
  3. Look up the probability in the z-table.
  4. If the question asks for a right-tail probability, use the complement: .

A small tail probability does not prove that the population mean changed. It says the observed sample mean would be unusual under the stated model, which is evidence to investigate alongside the sampling design and context.

Step 1
SE = σ / √n
—
Step 2
z = (x̄ − μ) / SE
—
Step 3 — Φ(z)
Standard normal probability area Mini standard normal N(0,1) curve. The orange shaded area to the left of the z marker represents Φ(z) = P(Z ≤ z).
Step 4 — Answers
P(x̄ < x̄₀) —
P(x̄ > x̄₀) —
Two-tail —

Figure 7: A live four-step pipeline. Edit any input field to see all four steps update instantly. The starting values deliberately use n = 50, so computing σ/√n remains visible rather than collapsing to a convenient integer. The shaded area in Step 3 represents Φ(z) — the left-tail probability.

Retrieval Bridge — Individual Z-Scores

You now know the full workflow for sample means. Before moving to worked examples, confirm you can still apply z-scores to individual observations — the contrast between the two cases is central to Section 4. A standard normal z-table shows . A student scores 82 on an exam where and . What is the probability that a randomly chosen student scores lower than this student?

Section 4: Worked Examples

▾

A tag such as (C2 + C3) after a title names the Core Concepts parts the problem draws on, so you know where to look back.

Example 1 — Parameters of the Sampling Distribution (Fully Worked)

A population has mean and standard deviation . We draw random samples of size .

(a) What are the mean and standard error of the sampling distribution of ?
(b) Is it reasonable to use the normal distribution for probability calculations? Why?

I notice that the question asks about repeated sample means, so I choose the sampling-distribution parameters rather than the distribution of individual values.

Part (a):

So the sampling distribution of has mean 80 and standard error 3.

Part (b): We need to check the CLT conditions. We’re not told the population shape, so we apply the rule of thumb: is ? Here .

This means the CLT rule of thumb is not met on its own. However, if the population is exactly normal, the sampling distribution is exactly normal for any ; if the population is only approximately normal, the model may still be a reasonable approximation. If the population is strongly skewed, we should be cautious with .

Key takeaway: Always check CLT conditions before computing probabilities. State your assumption about the population shape if .

Example 2 — Computing a Probability for (Partially Scaffolded)

Cereal boxes are filled by a machine with g and g. The fill weights are normally distributed. A quality inspector randomly selects boxes.

Find — the probability the sample mean is below 498 g.

Before revealing the model, predict which denominator belongs in the z-score— or —and explain why. Committing even an uncertain answer makes comparison with the solution more useful.

Pause and write. Record your answer and reasoning on paper or in your notes before continuing. Even if you are unsure, committing your thinking helps you see what you understood and what you need to revisit.

Show Solution

Step 1 — Standard error:

Step 2 — z-score:

Step 3 — probability:

I chose standard error because the random object is an average of 64 boxes. If the process is centred at 500 g, only about 2.3% of random samples have a mean below 498 g, so this result is unusual enough to investigate; it does not by itself prove underfilling.

Example 3 — Effect of Sample Size on Probability (Minimally Scaffolded)

Using the same cereal machine ( g, g), compute for samples of size and . Compare the three results (include from Example 2).

Decide what must be recomputed when changes, carry out both probability calculations, and explain the pattern in your own words.

Pause and write. Record your answer and reasoning on paper or in your notes before continuing. Even if you are unsure, committing your thinking helps you see what you understood and what you need to revisit.

Show Solution

For :

For (from Example 2):

For :

Summary:

SE
162.00−1.000.1587
641.00−2.000.0228
1000.80−2.500.0062

As increases, the SE shrinks, the z-score magnitude grows, and the probability of being far from drops sharply. Larger samples make it much easier to detect a real deviation from target.

Example 4 — Individual Observation vs. Sample Mean (Application Twist)

This problem looks like the previous ones, but there’s a twist. Read carefully.

A student exam score in a large course is normally distributed with and .

(a) What is the probability that a single randomly selected student scores above 78?
(b) What is the probability that the average score of a randomly selected class of 25 students is above 78?

Choose the denominator for each part before calculating, then explain why the two probabilities differ so much.

Pause and write. Record your answer and reasoning on paper or in your notes before continuing. Even if you are unsure, committing your thinking helps you see what you understood and what you need to revisit.

Show Solution

Part (a) — single observation: Use (not SE).

Part (b) — sample mean of 25: Use SE.

Contrast: A single student scoring above 78 is moderately likely (27%). But a class average above 78 is extremely unlikely (0.13%) — because the average of 25 scores is far less variable than any individual score. This is the power (and the purpose) of the SE formula.

Individual score X
Individual score distribution Normal curve for a single student score: N(72, 10²). Right tail above 78 is shaded. This panel never changes.
P(X > 78) ≈ —
Sample mean x̄ of n scores
Sample mean distribution Normal curve for the sample mean x̄: N(72, SE²) where SE = 10/√n. As n increases, the curve narrows and the tail probability shrinks dramatically.
P(x̄ > 78) ≈ —
Figure 8: Same cutoff (78), same μ = 72, σ = 10. The blue curve (individual) never changes — P(X > 78) ≈ 27%. The orange curve (sample mean) narrows as n grows, making P(x̄ > 78) collapse toward zero. Drag to n = 25 to see the contrast from Example 4.

Example 5 — Where Do Most Sample Means Land? (Working Backwards)

Every example so far started from a cutoff and asked for a probability. This one runs the other way. The cereal machine fills boxes with g and g, and inspectors average samples of boxes. Between which two weights do the middle 95% of sample means fall?

Decide which spread to use, and which z-value leaves 2.5% in each tail, before you open the solution.

Pause and write. Record your answer and reasoning on paper or in your notes before continuing. Even if you are unsure, committing your thinking helps you see what you understood and what you need to revisit.

Show Solution

Step 1 — Standard error:

Step 2 — z-value: the middle 95% leaves 2.5% in each tail, so (the table gives ).

Step 3 — Margin:

Step 4 — Interval:

In 95% of samples of 36 boxes, lands within 3.92 g of 500 g. Keep this sentence in mind: in L-INF-2 it is turned around. If is usually within 3.92 g of , then is usually within 3.92 g of , and that is a confidence interval.

Section 5: Guided Practice

▾

A tag such as (C2 + C3) after a title names the Core Concepts parts the problem draws on, so you know where to look back.

Problem 1 — Sample-Mean Probability (C2 + C5)

Problem 2 — Central Limit Theorem Conditions (C3)

For each scenario below, decide whether the CLT justifies using a normal distribution for the sampling distribution of .

Scenario A: A population is moderately right-skewed with no severe outliers. You draw a random, independent sample of .

Scenario B: Test scores in a large population are normally distributed. You draw a random sample of .

Scenario C: A population of daily wait times has an extreme bimodal shape (clustered near 2 min or 20 min, nothing in between). Sample size is .

Scenario D: A café wants the average wait time of all its customers. During one lunch rush it times 60 customers in a row; each wait depends heavily on how long the queue was for the customer before. Wait times are moderately right-skewed.

Problem 3 — Comparing Sample-Size Effects (C4 + C5)

Section 6: Independent Practice

▾

A tag such as (C2 + C3) after a title names the Core Concepts parts the problem draws on, so you know where to look back.

Problem 1 — Which Distribution Is It? (C1)

Sort into groups
Population, one sample, or sampling distribution?

A café records the wait time of every customer it serves; wait times are right-skewed. A manager takes a random sample of 25 customers and computes their mean wait, x̄. Sort each statement by the distribution it describes.

Drag an item into its group, or select it and then choose the group. On a keyboard: select an item and press the group’s number (1–3); 0 sends it back to Unsorted. Each item also has its own “move” buttons.
Unsorted

Problem 2 — When Does the CLT Apply? (C3)

Sort into groups
Can we use a normal model for x̄?

Each card describes how data were collected. Sort it by what the Central Limit Theorem lets you conclude about the sample mean.

Drag an item into its group, or select it and then choose the group. On a keyboard: select an item and press the group’s number (1–3); 0 sends it back to Unsorted. Each item also has its own “move” buttons.
Unsorted

Problem 3 — Sampling Distribution Probability (C2 + C5)

Problem 4 — Three Inspection Plans (C2 + C4 + C5)

Complete the table
Same process, three sample sizes
Fill in every empty cell, then check the whole table once. The cells depend on one another.

Problem 5 — Probability for a Non-Normal Population (C5)

Problem 6 — One Box or the Average of 36? (C1 + C2)

Structural Difference
One observation or a sample mean?
Identify what key word or condition separates Case A from Case B, then select the method for each.
Scenario A
Fill weights are normal with μ = 500 g and σ = 12 g. One box is chosen at random. Find P(weight > 506).
Scenario B
Fill weights are normal with μ = 500 g and σ = 12 g. A random sample of 36 boxes is averaged. Find P(x̄ > 506).

Problem 7 — Where Do Most Sample Means Land? (C2 + C5)

Problem 8 — Diagnose a Sampling-Distribution Error (C2 + C3 + C5)

Mixed Review — Retrieval from Earlier Lessons

These problems draw on concepts from earlier in the course. Attempting them without re-reading prior lessons is the point — retrieval practice strengthens long-term memory more than re-reading.

Review Problem 1 — Variability and Spread (L-DS-4)

Review Problem 2 — Normal Distribution Probability (L-PR-6)

Section 7: Mastery Check

▾

A tag such as (C2 + C3) after a title names the Core Concepts parts the problem draws on, so you know where to look back.

Begin each item without hints. After the explanation, each question checks one idea cold: what the sampling distribution is, when a normal model is justified, how sample size moves the standard error, and a complete probability decision.

Question 1 — Feynman Test

A classmate who missed this lesson asks you: “What is the Central Limit Theorem and why does it matter? What conditions does it need?”

Explain the lesson in concise, plain language on paper or in your notes. Address all three prompts:

  1. What is the sampling distribution of ?
  2. What does the CLT say, and what conditions make its normal approximation trustworthy?
  3. Why does standard error matter when judging a sample mean?

Pause and write. Record your answer and reasoning on paper or in your notes before continuing. Even if you are unsure, committing your thinking helps you see what you understood and what you need to revisit.

Compare with a model answer

The sampling distribution of describes how the mean changes across all random samples of the same size. It stays centred at , while its standard error is , so averages vary less than individual observations.

For independent observations from a population with finite variance, the CLT says the shape of this sampling distribution approaches normal as grows. If the population itself is normal, the sampling distribution is exactly normal for any . For a non-normal population, is only a starting rule; severe skew or heavy tails may require more data.

When a normal model is justified, we write and use the standard error to judge how unusual an observed average is. This supports evidence-based decisions without claiming that one unusual sample proves the population changed.

Question 2 — Which Distribution?

Read the situation carefully: is it about one observation or about the average of a sample?

Question 3 — Is a Normal Model Justified?

Question 4 — Scaling the Sample

Question 5 — A Quality-Control Decision

Work out the probability yourself first. If your answer is off, open the steps to find where it went wrong.

Self-Assessment

Still confusedReady for the Boss Fight

If you are below the midpoint, revisit the concept behind any question you missed (Question 2 → C1, Question 3 → C3, Question 4 → C4, Question 5 → C2 and C5) and try a new version. If you are near the upper end, continue and test that confidence on one complete Boss path.

Section 8: Boss Fight

▾

A tag such as (C2 + C3) after a title names the Core Concepts parts the problem draws on, so you know where to look back.

Choose one of two equally demanding missions. Neither tells you which method to use: choosing the tool is part of the fight. Each starts with your own complete attempt; open the steps only if you need to find where it went wrong. Lose, and you get a fresh rematch with new numbers.

🔬 The Quality Analyst

Decide whether a production result is ordinary variation or evidence worth investigating.

🏫 The School Board Analyst

Back to the opening puzzle: is the top-ranked small school really the strongest?

Section 9: Challenge Problems

▾

A tag such as (C2 + C3) after a title names the Core Concepts parts the problem draws on, so you know where to look back.

Optional stretch. These go beyond the lesson objectives. Attempt them when you feel confident with the core material.

Challenge 1 — Precision by Design

Find the smallest sample size that meets a randomly generated standard-error goal. The algebra is only part of the challenge: you must also round in the direction that preserves the design requirement.

Challenge 2 — The Finite Population Correction

The standard-error formula assumes an effectively infinite population. When a sample covers a meaningful share of a small population, the standard error is smaller:

For , , and , the sampling fraction is . Without the correction, ; with it, .

Now compute one yourself.

Challenge 3 — Why Averaging Reduces Variability

The formula follows from basic properties of variance. Here’s the argument — answer the two questions below.

Let be independent draws from a population with variance . Then:

. Which property lets you bring the outside the variance?

After extracting , we have . What allows us to write ?

See the complete derivation

Because the are independent:

Taking the square root: .

This is the exact formula for the standard error — derived from first principles, not memorized. The key ingredients are (1) the variance scaling rule and (2) independence of the observations.

Section 10: Solutions Reference

▾

A tag such as (C2 + C3) after a title names the Core Concepts parts the problem draws on, so you know where to look back.

Complete, step-by-step solutions for all problems in Sections 5–9 are available on the solutions page. Solutions include worked arithmetic, common mistakes to watch for, and interpretation guidance.

View Full Solutions →

If you’re stuck: Re-read the relevant Core Concept in Section 3, then find the Worked Example that maps to that concept (e.g., Example 1 maps to Concept 1). The solutions page shows the reasoning behind every step, not just the final answer.

Quick-Reference Formulas

Mean of the Sampling Distribution:

Standard Error of the Mean:

Normal Model When Conditions Support It:

Z-Score for a Sample Mean:

Population ShapeWhen is the sampling distribution normal?
NormalAlways (any sample size )
Unknown, skewed, or otherwise non-normalThe approximation improves as grows; is a starting rule, while severe skew or heavy tails may require more