Die Druckversion wird nicht mehr unterstützt und kann Darstellungsfehler aufweisen. Bitte aktualisiere deine Browser-Lesezeichen und verwende stattdessen die Standard-Druckfunktion des Browsers.

The Normal Distribution

Introduction

The normal distribution is one of the central models in statistics and probability. It describes a continuous, symmetric distribution whose values cluster around a mean and become progressively less common farther from the center. You will often see it called a Gaussian distribution or a bell curve.

This aiMOOC is designed for Grades 11–13. You will learn how to interpret the parameters of a normal model, calculate and explain z-scores, connect probabilities with areas under a density curve, use the empirical rule, judge whether a normal model is reasonable for data, and explain why normal-shaped patterns appear so often in statistics.

The graph above shows the standard normal density. The curve is highest at the mean, symmetric about the center, and approaches the horizontal axis without reaching it.

Learning Goals

By the end of the course, you should be able to:

Normal model: Describe the defining properties of a normal distribution and interpret its mean and standard deviation.
Density: Explain why probability is represented by area under a continuous density curve.
Standardization: Convert observations to z-scores and interpret the result in context.
Empirical rule: Use the 68–95–99.7 rule for quick estimates.
Cumulative probability: Interpret left-tail, right-tail, and interval probabilities.
Model checking: Use graphs and context to decide whether a normal model is plausible.
Sampling distributions: Distinguish a normally distributed variable from an approximately normal sampling distribution.

Understanding the Normal Model

A normal distribution is determined by two parameters: the mean $μ$ and the standard deviation $σ$ , where $σ > 0$ . We write

$X \sim N (μ, σ^{2})$ .

The mean $μ$ fixes the center of the curve. The variance $σ^{2}$ , or equivalently the standard deviation $σ$ , controls its spread. Increasing $σ$ makes the curve wider and lower because the total area must remain equal to one.

For an ideal normal distribution:

Symmetry: The curve is symmetric about $μ$ .
Mean: The mean, median, and mode coincide at the center.
Continuity: Every real number is a possible value in the mathematical model.
Tails: The tails extend indefinitely in both directions.
Probability density: The total area under the curve is exactly 1.

The curves above illustrate how changing the mean shifts a normal distribution and how changing the variance alters its spread.

The Probability Density Function

The density of a normal distribution is

$f (x) = \frac{1}{σ \sqrt{2 π}} e^{- \frac{(x - μ)^{2}}{2 σ^{2}}} .$

You do not usually calculate probabilities by evaluating this formula directly. Instead, the formula defines the shape of the density curve, while probabilities come from areas under the curve. For a continuous random variable, the probability of one exact value is zero; meaningful probabilities refer to intervals such as $P (a < X < b)$ .

Probability as Area

For a continuous distribution, an area under the density curve corresponds to probability. If $X$ is normally distributed, then

$P (a < X < b)$

is the area under the normal curve between $a$ and $b$ . The entire area under the curve is 1, so the probability of falling somewhere on the real number line is 100%.

A shaded normal curve makes tail and interval probabilities visible. Left-tail probabilities have the form $P (X < x)$ ; right-tail probabilities have the form $P (X > x)$ ; and middle areas describe probabilities such as $P (a < X < b)$ .

The Cumulative Distribution Function

The cumulative distribution function, abbreviated CDF, is

$F (x) = P (X \leq x) .$

For a normal variable, $F (x)$ gives the area under the density curve to the left of $x$ . Unlike the density, the CDF rises from values near zero to values near one.

Many calculators, statistics packages, and spreadsheets compute normal CDF values directly. The inverse normal function works in the opposite direction: given a cumulative probability such as 0.90, it returns the corresponding percentile value.

Standardization and Z-Scores

A z-score tells you how many standard deviations an observation lies above or below the mean. For a normally distributed variable,

$z = \frac{x - μ}{σ} .$

A positive z-score means the value is above the mean. A negative z-score means it is below the mean. A z-score near zero means the value is close to the center.

Example: Suppose examination scores are modeled by a normal distribution with mean 72 and standard deviation 8. A score of 86 has

$z = \frac{86 - 72}{8} = 1.75.$

So the score is 1.75 standard deviations above the mean. Standardization is useful because values measured in different units can be compared on the same scale.

The Standard Normal Distribution

When $μ = 0$ and $σ = 1$ , the model is the standard normal distribution, written $Z \sim N (0, 1)$ . Any normally distributed observation can be converted to this scale with a z-score. Probabilities for the original variable and its standardized z-score are identical when the corresponding regions are matched.

For example, if $X \sim N (170, 6^{2})$ , then $X = 182$ corresponds to $z = 2$ . Therefore

$P (X < 182) = P (Z < 2) \approx 0.9772.$

This means about 97.72% of values in the model lie below 182.

The Empirical Rule

The 68–95–99.7 rule gives quick probability estimates for any normal distribution:

Approximately 68.27% of values lie within one standard deviation of the mean.
Approximately 95.45% lie within two standard deviations of the mean.
Approximately 99.73% lie within three standard deviations of the mean.

The rule is especially useful for estimation and checking whether a more detailed calculator result is reasonable. It should not be applied automatically to strongly skewed, multimodal, or heavy-tailed data.

Using Symmetry

Because the normal curve is symmetric, equal distances on opposite sides of the mean have equal areas. For example, the probability of being more than two standard deviations above the mean is the same as the probability of being more than two standard deviations below the mean.

Symmetry also lets you convert left-tail probabilities into right-tail probabilities. If $P (Z < z) = p$ , then $P (Z > - z) = p$ and $P (Z > z) = 1 - p$ .

Checking Whether a Normal Model Is Reasonable

Real data are rarely perfectly normal. A normal model can still be useful when the distribution is reasonably symmetric, has one main peak, and has tails that are not much heavier than the normal curve.

Useful checks include:

Histogram: Look for a roughly symmetric, single-peaked shape.
Box plot: Look for strong skewness or unusually many extreme observations.
Q–Q plot: Points close to a straight line support a normal approximation.
Outlier: Investigate unusual observations because they can strongly affect the mean and standard deviation.
Context: Ask whether the data-generating process makes a normal model plausible.

The image above compares a normal density with a boxplot. A boxplot summarizes location and spread but does not show the smooth bell shape directly.

A common mistake is to conclude that any roughly bell-shaped sample must come from an exactly normal population. Sampling variation, measurement limitations, mixtures of groups, and hidden skewness can all complicate that conclusion.

Why Normal Shapes Appear So Often

One reason normal distributions are prominent is that many measurements combine numerous small influences. Another reason is the central limit theorem, which concerns sums and sample means.

The central limit theorem does not say that every large dataset is normally distributed. Instead, under suitable conditions, the sampling distribution of a sum or mean becomes approximately normal as the sample size grows, even when the individual observations are not themselves normal.

A Galton board offers a visual analogy: each ball experiences many small left-or-right deflections, and the accumulated outcomes often form an approximately bell-shaped pattern.

The convergence diagram illustrates how repeated aggregation can approach a normal form.

Applications and Limitations

Normal models are used in many areas, including measurement error, standardized testing, industrial quality control, biological measurements, finance, and statistical inference. Their value comes from a compact mathematical description and well-developed methods for calculating probabilities and standardized comparisons.

However, a normal model can be misleading when data are bounded, strongly skewed, multimodal, or dominated by rare extreme events. Income data, waiting times, and some financial returns are common examples where a normal model may fit poorly.

You should therefore treat normality as a modeling assumption to evaluate, not as a default truth about data.

Technology and Calculation

Depending on your course, you may use a graphing calculator, spreadsheet, or statistics software. Typical commands or functions let you:

Compute a cumulative probability below a given value.
Compute the probability between two values.
Find a percentile from a given cumulative probability.
Generate random values from a normal distribution.
Draw a normal curve and shade probability regions.

Even when technology gives the numerical answer, you should still sketch the region, identify whether it is a left tail, right tail, or interval, and interpret the probability in the original context.

Interactive Tasks

Quiz: Test Your Knowledge

Which parameter determines the center of a normal distribution? (The mean) (!The range) (!The sample size) (!The interquartile range)

What does the total area under a normal density curve equal? (One) (!Zero) (!The mean) (!The standard deviation)

What does a positive z-score indicate? (The value is above the mean) (!The value is below the mean) (!The value equals the median only) (!The distribution is skewed)

Approximately what percentage of a normal distribution lies within one standard deviation of the mean? (68 percent) (!50 percent) (!80 percent) (!99 percent)

What is the mean of the standard normal distribution? (Zero) (!One) (!Negative one) (!It depends on sample size)

What happens to a normal curve when its standard deviation increases while the mean stays fixed? (It becomes wider and lower) (!It shifts to the right) (!It becomes left skewed) (!Its total area becomes larger)

For a continuous normal variable, what is the probability of one exact value? (Zero) (!One half) (!One) (!It equals the density height)

Which graph is especially useful for checking normality by looking for an approximately straight pattern? (A Q Q plot) (!A pie chart) (!A network diagram) (!A stem map)

What does the central limit theorem primarily describe? (The distribution of sums or sample means) (!Every raw dataset becoming normal) (!Every population having equal variance) (!The median always equaling the mean)

Which statement best describes a z-score of minus two? (The value is two standard deviations below the mean) (!The value is two units below zero) (!The standard deviation is negative) (!The value is twice the mean)

Memory Game

Mean	Center parameter of the normal distribution
Standard deviation	Dispersion parameter controlling how wide the curve is
Z-score	Signed standardized distance from the center
Density	Curve height used to determine probability through area
Percentile	Cutoff below which a chosen proportion falls
Symmetry	Property of having matching left and right shapes

Drag and Drop

Match the correct terms.	Topic
Shifts the center	Mean
Controls the width	Standard deviation
Measures standardized distance	Z-score
Accumulates probability from the left	Cumulative distribution function
Checks linear agreement with normal quantiles	Q Q plot

...

Crossword Puzzle

Gaussian	What alternative name is often used for the normal distribution?
Symmetry	What property makes the left and right halves mirror each other?
Variance	What quantity equals the square of the standard deviation?
Percentile	What value marks a chosen cumulative proportion of observations?
Density	What function has area under its curve representing probability?
Standardize	What verb describes converting a value to a z-score scale?

LearningApps

Cloze Text

A normal distribution is symmetric about its

. Its spread is controlled by the

. Probability for a continuous normal variable is represented by

under the density curve. A standardized observation is called a

. The standard normal distribution has a standard deviation of

. The empirical rule places about 95 percent of values within

standard deviations of the mean. A cumulative distribution function gives the probability to the

of a chosen value. A Q Q plot can help you assess whether a normal model is

. The central limit theorem concerns the behavior of sums and sample

under suitable conditions.

Open-Ended Tasks

Easy

Bell curve annotation: Draw a normal curve, label the mean, mark one, two, and three standard deviations on each side, and explain what each region means.
Empirical rule infographic: Create a one-page infographic that communicates the 68–95–99.7 rule using your own wording and a clearly labeled diagram.
Z-score story: Invent a realistic school or sports context with a mean and standard deviation, calculate one z-score, and interpret it in a complete sentence.
Normal distribution in everyday life: Find one public example where a bell curve is used, summarize the claim, and explain whether the normal model seems reasonable.

Standard

Classroom data study: Collect at least thirty anonymous measurements or timings, create a histogram, calculate the mean and standard deviation, and discuss whether a normal model fits.
Galton board experiment: Build or simulate a simple Galton board process, record many outcomes, graph the results, and explain why the shape changes as the number of trials increases.
Spreadsheet normal model: Use a spreadsheet to generate normal random values, create a histogram, compare observed proportions with the empirical rule, and document your formulas.
Statistics interview: Interview a person who uses quantitative data in study or work, ask how standardized scores or probability models support decisions, and summarize the most important insight.

Advanced

Q Q plot investigation: Compare a roughly symmetric dataset with a clearly skewed dataset using histograms and Q Q plots, then justify which dataset is better modeled as normal.
Central limit theorem simulation: Simulate repeated sample means from a non-normal population for several sample sizes, graph the sampling distributions, and explain what changes.
Model comparison project: Choose a real dataset, compare a normal model with another plausible distribution, and defend which model better represents the data and why.
Tail probability explainer: Produce a short video or narrated slide sequence showing how z-scores, tail probabilities, and percentiles can be used to reason about rare events without exaggerating what the model proves.

Learning Assessment

Parameter reasoning: Explain how two normal distributions can have the same mean but different probabilities in their tails, using standard deviation in your reasoning.
Standardized comparison: Compare two students whose scores come from different exams by calculating z-scores and deciding who performed better relative to their own group.
Probability modeling: For a stated normal model, sketch and calculate a left-tail, right-tail, and interval probability, then interpret each result in context.
Inverse normal reasoning: Determine a cutoff value for a stated percentile and explain why an inverse normal calculation is appropriate.
Model critique: Evaluate a dataset description for skewness, outliers, and context, then argue whether using a normal model would be defensible.
Central limit transfer: Explain why the distribution of individual observations may be skewed while the distribution of sample means can still be approximately normal.

Evidence of Learning

Knowledge: You can define the normal distribution, identify its parameters, distinguish density from probability, explain the standard normal model, and state the role of the central limit theorem.

Skills: You can calculate and interpret z-scores, estimate probabilities with the empirical rule, use cumulative and inverse normal calculations, sketch probability regions, and check the plausibility of a normal model with graphical evidence.

Products: Strong evidence may include annotated bell curves, probability calculations with interpretations, a spreadsheet model, a histogram and Q Q plot analysis, a simulation report, an interview summary, or a short explanatory video.

Transfer achievements: You can decide when a normal model is useful in a new context, compare values measured on different scales, critique inappropriate normality assumptions, and connect sampling distributions with statistical inference.