Normal Distribution

Core Idea

The normal distribution with mean μ and standard deviation σ has PDF f(x) = (1/(σ√(2π))) × e^(-(x-μ)²/(2σ²)). It is symmetric, bell-shaped, and completely determined by its mean and variance. The normal distribution is ubiquitous in statistics because many naturally occurring phenomena approximate it, and because of the central limit theorem, which states that means of large samples are approximately normal regardless of the original distribution.

How It's Best Learned

Visualize how mean shifts and standard deviation stretches the bell curve. Use the empirical rule (68-95-99.7). Compare distributions with different μ and σ.

Common Misconceptions

Assuming all bell-shaped distributions are normal. Thinking the normal distribution can be negative (values are on ℝ, but probabilities decay in tails). Confusing standard deviation with variance.

Explainer

You have already encountered the idea of a continuous random variable — a quantity that can take any value in an interval, described by a probability density function (PDF). The normal distribution is the most important continuous distribution in all of statistics, and understanding why requires looking at both its shape and its origins.

The normal distribution is symmetric and bell-shaped, centered at its mean μ. The spread is controlled by the standard deviation σ: a small σ produces a tall, narrow bell, while a large σ produces a short, wide one. The mean and standard deviation are all you need to specify a normal distribution completely — there are no additional shape parameters. This makes it unusually tractable mathematically.

A practical rule that makes normals easy to reason about is the empirical rule: approximately 68% of values fall within one standard deviation of the mean (μ ± σ), about 95% within two, and about 99.7% within three. This rule is worth memorizing because it lets you quickly answer questions like "how unusual is a value 2 standard deviations above average?" without computing integrals. (Answer: only about 2.5% of values are that far above the mean.)

One of the most common misconceptions is treating any bell-shaped distribution as normal. The t-distribution, for instance, is also symmetric and peaked in the middle, but its tails are heavier — extreme values are more likely than the normal predicts. The normal is defined by a specific mathematical formula, not just by its visual shape. That said, the normal is ubiquitous because of the central limit theorem (a topic you will encounter soon), which says that the average of a large number of independent random variables is approximately normally distributed, regardless of the original distribution. This is why so many real-world measurements — heights, measurement errors, test scores — approximate the normal.

When you see a normal distribution, always read off μ and σ first. They tell you where the bulk of the data sits and how spread out it is. A score of 130 on a test with μ = 100, σ = 15 is two standard deviations above average — notable but not extraordinary (about 2.3% of people score this high). The same score on a test with μ = 100, σ = 5 would be six standard deviations above average — effectively impossible in a true normal. The same number means very different things depending on the distribution's parameters.