Why can averages look bell-shaped when individual values do not?
You will be able to: Explain the CLT’s role and limits using repeated sample means.
Why can averages look bell-shaped when individual values do not?
Most visits to a help desk are short, but a few take much longer. Individual times are right-skewed; averages of larger independent samples are more balanced.
A useful starting point: What varies when you keep taking new samples? →
Words and symbols before equations
- Central limit theorem CLT
- Under suitable conditions, a standardized mean of many independent observations approaches a normal distribution.
- Sample size n
- Number of observations in each average.
- Standard error
- Standard deviation of a statistic’s sampling distribution.
- Finite variance
- A population spread with a finite squared-deviation average.
What this picture assumes
Independent draws with replacement from a fixed finite-variance population. Each plotted observation in the sampling histogram is one sample mean. Changing the number of repeated samples changes Monte Carlo detail; changing n changes the theoretical spread.
Read the picture in three steps
- Read the axes and labels first. Identify what each symbol and line represents. Read the units and fixed conditions before comparing quantities.
- 200 independent samples, each of size 5. Population mean 1; theoretical SD of sample means 0.447. Histogram bins include the lower endpoint and exclude the upper, except the last includes both. Larger n reduces theoretical spread; more repetitions refine the simulation.
- Check what the picture assumes below. Use the Explore task to predict one change before moving a control.
Connect the picture to the mathematics
For independent, identically distributed observations with finite mean μ and finite positive variance σ², the sample mean centers at μ and its standard deviation is σ/√n. The standardized mean becomes approximately normal as n grows.
The CLT changes the distribution of sample means, not the population’s individual values. Strong skew or rare extremes may require much larger n; 30 is not a universal guarantee.
Our right-skewed toy population has values 0,1,4 with probabilities 0.6,0.3,0.1. Its mean is 0.7 and variance 1.41. Means from larger independent samples cluster more tightly around 0.7. More simulation repetitions smooth the display but do not substitute for larger n.
A worked example, step by step
An independent population model has μ=20 minutes and σ=12 minutes. Compare the spread of sample means for n=4 and n=36.
- Both sampling distributions have mean 20 minutes.
- For n=4, standard error is 12/√4=6 minutes.
- For n=36, standard error is 12/√36=2 minutes.
- The larger sample produces a narrower mean distribution. Approximate normality still depends on population shape and the conditions; the population itself is unchanged.
The CLT does not make raw data normal, and larger n does not repair biased sampling or dependence.
Does n=30 guarantee a normal approximation for every population?
Compare with an explanation
No. How quickly means become approximately normal depends on population shape and tail behavior.
Predict. Change one thing. Explain.
Use the skewed population and compare n=1,5,20,40. Observe center, spread and shape. Change only the number of replications afterward and explain the different effect.
On narrow screens, swipe or scroll diagrams sideways to read all labels.
200 independent samples, each of size 5. Population mean 1; theoretical SD of sample means 0.447. Histogram bins include the lower endpoint and exclude the upper, except the last includes both. Larger n reduces theoretical spread; more repetitions refine the simulation.
| Population value | Probability |
|---|---|
| 0 | 0.5 |
| 2 | 0.5 |
| Quantity | Value |
|---|---|
| Population mean μ | 1 |
| Population SD σ | 1 |
| Theoretical mean of sample means | 1 |
| Theoretical SD σ/√n | 0.447 |
| Simulated average of sample means | 1.08 |
Independent draws with replacement from a fixed finite-variance population. Each plotted observation in the sampling histogram is one sample mean. Changing the number of repeated samples changes Monte Carlo detail; changing n changes the theoretical spread.
Explain what you noticed: Answer the investigation prompt above. State one observation and explain it using the probability values, reference groups, graph scales or model assumptions. Identify what the representation cannot tell you.
Apply the idea to a fresh problem Practice →Show what you understand.
Two original questions are a starting check, not proof of mastery. Explain your choice before revealing the answer.
Original written challenge
4 points · self-check · not an official AP questionA population has μ=8 and σ=6. Under independent sampling, compare means for n=9 and n=36 and state one limit of the CLT.
This response is not submitted or saved. Copy it before leaving.
Compare with the answer and four-point rubric
- 1 point: Both sample-mean distributions center at 8.
- 1 point: Standard errors are 6/3=2 and 6/6=1.
- 1 point: The n=36 mean distribution is less variable.
- 1 point: Approximate normality is not guaranteed solely by n; severe skew, dependence or nonfinite variance can invalidate a simple application.
Accept equivalent correct methods and explanations. This is a Refresh Kid teaching rubric, not an official AP scoring guideline.
Retrieve it before you reveal it.
RECALL 1Which distribution becomes approximately normal?
The appropriately standardized sample-mean distribution under the CLT conditions.
RECALL 2Does more n remove selection bias?
No.
RECALL 3How does standard error change with n?
For independent observations it scales as 1/√n.
Revisit these tomorrow and a week later. Try a fresh problem and explain why the method applies.
Why can averages look bell-shaped when individual values do not?
- For independent identical observations: mean(x̄)=μ, SD(x̄)=σ/√n.
- CLT describes means, not individual data.
- Normal approximation needs suitable conditions and sufficiently large n.
Remember: The CLT does not make raw data normal, and larger n does not repair biased sampling or dependence.
Conditions: Independent draws with replacement from a fixed finite-variance population. Each plotted observation in the sampling histogram is one sample mean. Changing the number of repeated samples changes Monte Carlo detail; changing n changes the theoretical spread.
Refresh Kid · AP Statistics Unit 2 · Objectives 2.12.A · Review edition
Framework, scope and review status
Mapped to College Board, AP Statistics CED, Topic 2.12, objectives 2.12.A. Framework effective Fall 2026, checked September 17, 2026. Unit 2 includes probability, random variables, probability models and introductory sampling distributions; it is part of the revised five-unit course.
Examples and datasets are synthetic, independently authored teaching material. Assumptions about independence, replacement and equal likelihood are stated before calculations. Simulation estimates fluctuate. Discrete probability is summed; continuous probability is area. Sampling distributions and randomization distributions use different repetition mechanisms. Formal inference comes in later units.
The Organic Chemistry Tutor companion title and destination were checked; the full video was not reviewed. Khan Academy’s destination was checked, but its lesson content was not fully readable by the research tool. OpenStax provides optional reference reading. No provider scripts, questions or graphics were copied. Refresh Kid is not affiliated with these providers.
GitHub’s 3D website collection informed optional spatial inspection. Our original three-toss outcome cube uses self-hosted Three.js with its MIT license. Eight corners represent eight equally likely sequences of three fair independent tosses. Conditioning removes ineligible sequences; camera rotation never changes probabilities. Quantitative graphs remain 2D to avoid perspective distortion. Complete labeled diagrams, outcome lists and explanations remain available without 3D.
Independent teacher review and observation of students remain pending. Technical checks do not certify statistical accuracy, accessibility or learning effectiveness. This is a review edition.
Released AP Statistics questions and scoring guides are optional. Older exams use the earlier framework, so check alignment before selecting parts. All practice on this page is original, not official AP material.
Learn → Explore → Practice → Review is informed by the IES learning guide. This implementation has not yet been evaluated with learners.
Want to work through this with a tutor?
Bring your question about Why can averages look bell-shaped when individual values do not? Your explanation and answers remain free to access.