The 68–95–99.7 rule
The empirical rule for normal distributions: about 68%, 95% and 99.7% of data lie within 1, 2 and 3 standard deviations of the mean, and using the symmetric pieces to find percentages.
Here is the payoff for learning standard deviation in Year 11. If data is normal, you don't need the full dataset to know how it's spread. The mean and standard deviation alone tell you what percentage of the data lies in any band of whole standard deviations.
The rule
For any normal distribution, approximately:
- $68\%$ of the data lies within $1$ standard deviation of the mean: between $\bar{x} - s$ and $\bar{x} + s$.
- $95\%$ lies within $2$ standard deviations: between $\bar{x} - 2s$ and $\bar{x} + 2s$.
- $99.7\%$ lies within $3$ standard deviations: between $\bar{x} - 3s$ and $\bar{x} + 3s$.
This is on the HSC reference sheet, but you need to use it fast, so learn it. It's also called the empirical rule.
The heights of Year 12 boys at a school are normally distributed with mean $176$ cm and standard deviation $7$ cm. Find the interval that contains the middle $95\%$ of heights.
Breaking the curve into pieces
Because the curve is symmetric, each band splits evenly in half. Work out every piece from the three big numbers:
| Region | Working | Percentage |
|---|---|---|
| Mean to $1s$ (one side) | $68 \div 2$ | $34\%$ |
| $1s$ to $2s$ (one side) | $(95 - 68) \div 2$ | $13.5\%$ |
| $2s$ to $3s$ (one side) | $(99.7 - 95) \div 2$ | $2.35\%$ |
| Beyond $3s$ (one side) | $(100 - 99.7) \div 2$ | $0.15\%$ |
| Above the mean | half the data | $50\%$ |
For the same heights (mean $176$, $s = 7$), what percentage of boys are (a) taller than $190$ cm, (b) between $169$ cm and $190$ cm?
Use the sliders to set the mean and standard deviation, then press 68%, 95% or 99.7% to shade that region. The interval appears underneath.
Change the standard deviation. Does the shaded percentage change? What does change? Then try Pick a band: tap two markers and see the percentage between them.
Your turn
The time Year 12 students take to complete a 5 km fun run is normally distributed with mean $32$ minutes and standard deviation $4$ minutes. What percentage of students take between $24$ and $36$ minutes? If $600$ students ran, how many is that?
Show the answer
$24 = 32 - 2 \times 4$ is $2s$ below, and $36 = 32 + 4$ is $1s$ above. So the percentage is $13.5 + 34 + 34 = 81.5\%$. Then $0.815 \times 600 = 489$ students.
For normal data, about what percentage lies within $2$ standard deviations of the mean?
Within $1$, $2$ and $3$ standard deviations: $68\%$, $95\%$ and $99.7\%$. So $2$ standard deviations gives $95\%$.
Marks are normal with mean $60$ and $s = 10$. About $68\%$ of marks lie between:
$68\%$ is within $1s$: $60 - 10 = 50$ to $60 + 10 = 70$.
What percentage of normal data lies more than $1$ standard deviation above the mean?
Outside the middle $68\%$ is $100 - 68 = 32\%$, split evenly between two tails: $32 \div 2 = 16\%$.
Marks are normal with mean $60$ and $s = 10$. What percentage of marks are below $40$?
$40 = 60 - 2 \times 10$, which is $2s$ below. Outside $95\%$ is $5\%$, and half of that is in the lower tail: $2.5\%$.
For normal data, what percentage lies between the mean and $2$ standard deviations above it?
$95\% \div 2 = 47.5\%$, or $34 + 13.5 = 47.5\%$.
Bags of chips have masses normally distributed with mean $170$ g and $s = 5$ g. What percentage of bags weigh between $160$ g and $175$ g?
$160 = 170 - 2 \times 5$ ($2s$ below) and $175 = 170 + 5$ ($1s$ above). Add the pieces: $13.5 + 34 + 34 = 81.5\%$.
Data is normal with mean $50$. About $99.7\%$ of values lie between $35$ and $65$. The standard deviation is:
$99.7\%$ is within $3s$, so $3s = 65 - 50 = 15$ and $s = 15 \div 3 = 5$.
What percentage of normal data lies more than $3$ standard deviations below the mean?
Outside the middle $99.7\%$ is $0.3\%$. Half of that is below: $0.3 \div 2 = 0.15\%$.
Exam style: The lifetimes of a brand of LED globe are normal with mean $25{,}000$ hours and $s = 2{,}000$ hours. What percentage of globes last more than $23{,}000$ hours?
$23{,}000 = 25{,}000 - 2{,}000$, which is $1s$ below. Above it is $34\%$ (up to the mean) plus $50\%$ (above the mean): $34 + 50 = 84\%$.
Exam style: At a Newcastle school, $1{,}080$ students sat a test. Marks were normal with mean $64$ and $s = 9$. How many students scored above $82$?
$82 = 64 + 2 \times 9$, which is $2s$ above. The percentage above is $5 \div 2 = 2.5\%$. Number of students: $0.025 \times 1{,}080 = 27$.
Coaching Centre