Normal Distribution Explained: Bell Curve, Z-Scores & the 68-95-99.7 Rule

What Is the Normal Distribution?

The normal distribution — also called the Gaussian distribution or bell curve — is the most important probability distribution in statistics. It's symmetric around its mean, meaning half of all values fall above the mean and half below. The curve is tallest at the mean and falls off smoothly in both directions, producing its signature bell shape.

Two numbers completely define any normal distribution: the mean (μ), which sets where the center sits, and the standard deviation (σ), which controls how wide or narrow the bell is. A large σ produces a flat, spread-out curve; a small σ produces a tall, narrow one.

The probability density function is written as: f(x) = (1/σ√(2π)) × e^(−(x−μ)²/2σ²)

You don't need to memorize that formula. Our normal distribution calculator computes everything for you. But understanding the shape and the rules behind it unlocks a huge amount of statistical reasoning.

5 Properties of the Normal Distribution

1. Symmetric about the mean. The left half is a mirror image of the right. Mean, median, and mode are all equal.
2. Bell-shaped. Values near the mean are most probable; extreme values become increasingly rare as you move away from center.
3. Defined by just two parameters. μ and σ fully specify the distribution. Change either one and you get a different bell curve.
4. The total area under the curve equals 1. All probabilities must sum to 100%, and the normal curve satisfies this exactly.
5. Asymptotic tails. The curve never actually touches zero — it approaches the x-axis but extends infinitely in both directions. This is why extremely rare events are possible, just vanishingly unlikely.

The 68-95-99.7 Rule (Empirical Rule)

The most useful shortcut in all of statistics. In any normal distribution, the following always holds:

Applied to IQ scores (μ = 100, σ = 15): 68% of people score between 85 and 115. Only about 2.3% score above 130 (2σ above mean). A score above 145 (3σ) is rarer than 1 in 700 people.

Why This Rule Matters

The 68-95-99.7 rule lets you assess how unusual any observation is without doing any calculations. If a manufactured part is supposed to be 10 cm with a tolerance of ±0.3 mm, and the production process has σ = 0.1 mm, then 99.73% of parts fall within spec automatically. Quality engineers call values beyond 3σ “defects” — that's where the Six Sigma methodology gets its name (targeting 6σ from the mean, which theoretically allows only 3.4 defects per million).

Z-Scores: Converting to Standard Units

A z-score (also called a standard score) measures how many standard deviations a value sits from the mean. The formula:

z = (x − μ) / σ

Where:

• x = the observed value
• μ = population mean
• σ = standard deviation

Z-Score Examples

Z-scores are powerful because they let you compare observations across completely different scales. A student who scored 1270 on the SAT and 28 on the ACT — which is the better result? Convert both to z-scores and you can answer objectively.

Reading a Standard Normal (Z) Table

A z-table tells you what percentage of a normal distribution falls below a given z-score. For z = 1.00, the table returns 0.8413, meaning 84.13% of values fall below that point. To find the probability abovez = 1.00, subtract from 1: 1 − 0.8413 = 0.1587, or about 15.87%.

Our normal distribution calculator automates this lookup and shows the area visually on the bell curve.

4 Real-World Uses of the Normal Distribution

1. Standardized Testing

The College Board designs SAT scores to follow a normal distribution. According to College Board data (2023), the mean combined SAT score was approximately 1028 with a standard deviation near 210. This lets colleges compare students across graduating classes and years by converting scores to percentiles derived directly from normal probabilities.

2. Human Measurements

The CDC's National Health Statistics Reports (2018) found that the mean height of US adult men aged 20–39 is 69.1 inches with a standard deviation of roughly 2.9 inches. Women average 63.7 inches with σ ≈ 2.7 inches. Both distributions are approximately normal, which is why height is the canonical textbook example.

3. Manufacturing and Quality Control

Any repetitive production process introduces small random errors that add up to a normal distribution of part dimensions, weights, or chemical concentrations. Six Sigma quality methodology, used by companies like Motorola and General Electric since the 1980s, is built entirely on controlling the normal distribution of defects. The goal: push the process mean so far from specification limits that failures occur fewer than 3.4 times per million opportunities.

4. Financial Returns and Risk

Modern Portfolio Theory (Markowitz, 1952) models asset returns as approximately normally distributed — an assumption baked into tools like Value at Risk (VaR). A portfolio with annual expected return of 8% and σ of 15% implies a 95% chance of returns falling between −22% and +38% (μ ± 2σ). This is how fund managers and risk officers quantify the probability of large losses.

The Central Limit Theorem: Why the Bell Curve Is Everywhere

The reason the normal distribution appears so often isn't that the universe is especially fond of it. It's the Central Limit Theorem (CLT): when you average many independent, identically distributed random variables, the result approaches a normal distribution — regardless of what the original distribution looked like.

Individual dice rolls are uniformly distributed (each face equally likely). But roll 30 dice and take the average: that average is nearly normally distributed. This is why aggregated quantities — average test scores, total rainfall, mean blood pressure across patients — tend to be normal even when individual measurements are not.

The CLT is why statisticians can use normal-distribution-based tests (t-tests, ANOVA, regression) on data that isn't itself perfectly normal, as long as sample sizes are large enough (typically n > 30 is a common rule of thumb).

Standard Normal Distribution vs. General Normal Distribution

Every normal distribution can be converted to the standard normal by z-scoring. This is why statisticians only ever need one table (the z-table) to handle any normal distribution.

When Data Is Not Normal

Not everything follows a normal distribution. Income data is right-skewed (most people earn below the mean; a few earn enormous amounts). Lifetimes of components often follow an exponential or Weibull distribution. Stock prices themselves — as opposed to log-returns — are not normal.

You can test normality formally with the Shapiro-Wilk test or the Kolmogorov-Smirnov test, or visually with a Q-Q plot. If your data is severely non-normal and your sample is small, non-parametric tests (Mann-Whitney, Kruskal-Wallis) are more appropriate than t-tests.

But for most applied purposes, the normal distribution is a powerful and reliable approximation. Albert Einstein allegedly said statistics is “the art of saying something true about something that isn't quite right” — and the normal distribution embodies that tradeoff better than any other distribution.

Frequently Asked Questions

What is the normal distribution?

The normal distribution is a symmetric, bell-shaped probability distribution defined by its mean (μ) and standard deviation (σ). Most values cluster around the mean, tapering off equally on both sides. It describes countless real-world phenomena including human height, IQ scores, measurement errors, and exam results.

What is the 68-95-99.7 rule?

The 68-95-99.7 rule (also called the empirical rule) states that in any normal distribution: 68.27% of data falls within ±1 standard deviation of the mean, 95.45% within ±2 standard deviations, and 99.73% within ±3 standard deviations. This lets you quickly assess how unusual any observation is.

How do you calculate a z-score?

A z-score measures how many standard deviations a value is from the mean. The formula is z = (x − μ) / σ, where x is the observed value, μ is the population mean, and σ is the standard deviation. A z-score of +2 means the value is 2 standard deviations above average.

Why is the normal distribution so common in nature?

The Central Limit Theorem explains it: when you average many independent random variables, the result tends toward a normal distribution regardless of the original distributions. This is why measurements that reflect many small additive influences — height, test scores, biological traits — naturally follow the bell curve.

What is a standard normal distribution?

A standard normal distribution has a mean of 0 and a standard deviation of 1. Any normal distribution can be converted to standard normal form by computing z-scores. This standardization lets you compare values across different scales and use standard normal tables (z-tables) to find probabilities.

What percentage of SAT scores fall between 400 and 1600?

SAT scores are designed to follow a normal distribution with a mean near 1028 and standard deviation of about 210 (College Board, 2023). Scores between 400 and 1600 span roughly ±3 standard deviations from the mean, encompassing approximately 99.73% of all test takers per the empirical rule.