# Normal Distribution

> The bell curve that heights, test scores and errors tend to follow. The 68-95-99.7 rule, what a percentile tells you, and how to spot an outlier.

**Course:** [Probability and Statistics](https://learn.benibix.com/learn/general-math)  
**Time:** 12 min

**After this lesson you can:**

- Use the 68-95-99.7 rule
- Read a percentile
- Spot an outlier on a box plot

## What is a normal distribution?

A probability distribution lists each outcome and its chance, and all the chances add up to 1. The normal distribution is the most famous one. It is a bell-shaped curve. Most values sit near the middle, and few sit out at the edges. Heights, test scores and measuring errors often follow it closely.

A bell curve is described by two numbers: its mean, where the peak sits, and its standard deviation, how wide it is. About 68 percent of values fall within one standard deviation of the mean. About 95 percent fall within two, and 99.7 percent within three.

Many tests of reasoning are built to follow this curve, with a mean of 100 and a standard deviation of 15. So about 68 percent of people score between 85 and 115, and about 95 percent between 70 and 130. A score above 145 is three standard deviations up, rarer than 1 in 700.

A percentile says where a value ranks out of 100. A child at the 90th percentile for height is taller than 90 percent of children the same age. It is not a score out of 100. It is a rank, so it works for any kind of measure.

An outlier is a value far from all the others. A box plot is a quick way to spot one. The box runs from the 25th to the 75th percentile. A common rule marks as an outlier anything more than one and a half box widths beyond the box.

An outlier can be a mistake, like a height typed as 1,750 cm instead of 175. Or it can be the most interesting thing in the data. Check it before you delete it. Throwing out real outliers can hide exactly what you were looking for.

The common mistake is to think all data follows a bell curve. Incomes, city sizes and website visits do not. They have long tails, where a few huge values dwarf the rest. Bell-curve rules guess far too low on how extreme things can get.

![A bell curve with the mean at the centre. A dark band covers one standard deviation each side, marked 68%. A wider, lighter band covers two each side, marked 95%. The widest band covers three each side, marked 99.7%.](https://learn.benibix.com/diagrams/gen-normal-distribution.svg)

![A box plot over a number line. A box runs from the 25th percentile to the 75th percentile, with a line inside it at the 50th, the median. Whiskers reach out to the smallest and largest normal values. A dashed fence sits further out on the right. One dot lies far to the right, past the fence, marked outlier.](https://learn.benibix.com/diagrams/gen-box-plot.svg)

### Example: A baby's growth chart

At her check-up, baby Ava is at the 15th percentile for weight. Her father worries that she scored 15 out of 100.

The nurse explains that it means 15 percent of babies her age weigh less. She has stayed near the 15th line for months, which is what matters.

**Words in this lesson:** [Probability Distribution](https://learn.benibix.com/term/general/probability-distribution), [Normal Distribution](https://learn.benibix.com/term/general/normal-distribution), [Percentile](https://learn.benibix.com/term/general/percentile), [Outlier](https://learn.benibix.com/term/general/outlier)

**Question:** Test scores follow a bell curve with a mean of 100 and a standard deviation of 15. About what share score between 70 and 130?

**Answer:** 95 percent. 70 and 130 are two standard deviations either side of the mean. About 95 percent of values fall in that range.

---

Source: https://learn.benibix.com/learn/general-math/normal-distribution
Author: Behnia Zarei (a developer in Vienna, Austria, who builds and teaches data, AI and business intelligence at BENIBIX)
Publisher: BENIBIX Learn (https://learn.benibix.com/)
Last updated: 2026-11-19
Licence: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Quote freely with attribution.
