📖 Short lecture
Many measurements (heights, test scores, measurement errors) follow a bell-shaped normal distribution $N(\mu,\sigma^2)$, symmetric about the mean $\mu$, with standard deviation $\sigma$ measuring the spread. Total area under the curve is $1$, and areas are probabilities.
The empirical rule: about $68\%$ of values lie within $1\sigma$ of the mean, $95\%$ within $2\sigma$ and $99.7\%$ within $3\sigma$. The $z$-score $z=\frac{x-\mu}{\sigma}$ converts any normal variable to the standard normal $N(0,1)$.
⚠️ Common mistakes
- $z=\frac{x-\mu}{\sigma}$ divides by the standard deviation $\sigma$, not the variance $\sigma^2$.
- A table gives $P(Z\lt z)$; for “greater than” use $1-P(Z\lt z)$.
- The empirical rule (68–95–99.7) only applies to normal (bell-shaped) data.
🔑 Key formulas
✏️ Worked example
Exam scores are $N(70,10^2)$. What percentage of students score between $50$ and $90$? What is the $z$-score of $85$?
- $\displaystyle 50=\mu-2\sigma,\;90=\mu+2\sigma\Rightarrow\approx95\%$
- $\displaystyle z=\frac{85-70}{10}=1.5$
✔ About $95\%$; $z=1.5$.
🧮 More worked examples
Each step says what we do and why.
Example 2Heights are $N(165,8^2)$ cm. Find $P(X\gt181)$.
- $z$-score.$$z=\frac{181-165}{8}=2$$
- Empirical rule: $95\%$ within $2\sigma$.$$\frac{1-0.95}{2}=0.025$$
Example 3$\mu=50$, $\sigma=4$. Which value has $z=-1.5$?
- Reverse formula.$$x=50+(-1.5)(4)=44$$
Example 4Scores are $N(500,100^2)$. Find $P(400\lt X\lt700)$.
- Standardise.$$-1\lt Z\lt2$$
- Split at the mean.$$34\%+47.5\%$$
Example 5$X\sim N(50,8^2)$. Find $P(X\lt60)$ using $P(Z\lt1.25)=0.8944$.
- Standardise.$$z=\frac{60-50}{8}=1.25$$
- Read the table.$$P(Z\lt1.25)=0.8944$$
Example 6$X\sim N(50,5^2)$. Find $P(45\lt X\lt55)$.
- Both ends are one $\sigma$ away.$$-1\lt Z\lt1$$
- Empirical rule.$$\approx0.68$$
Example 7Find the interval holding the middle $90\%$ of $N(100,15^2)$.
- $z$ for $5\%$ in each tail.$$z=1.645$$
- Convert.$$100\pm1.645(15)=100\pm24.7$$
🖐️ Interactive demo & illustration
Move the sliders or press the buttons: the figure and the numbers update immediately. Try to predict the result before you look.
🎬 A Galton board makes a bell curve
Each ball bounces left or right at every row of pegs with probability $\tfrac12$. The bin it lands in is the number of right bounces, a binomial count. Drop many balls: the histogram takes the bell shape of the normal curve with $\mu=\frac n2$ and $\sigma=\frac{\sqrt n}{2}$, and about $68\%$ land within one $\sigma$ of the middle.
🎞️ Shade the probability under the normal curve
For $X\sim N(\mu,\sigma^2)$ the probability $P(a\lt X\lt b)$ is the area under the bell curve between $a$ and $b$. The panel converts $a$ and $b$ to $z$-scores with $z=\frac{x-\mu}{\sigma}$ and shows the area. Press play to grow the interval outwards from the mean and watch $68\%$, $95\%$ and $99.7\%$ appear.
🌍 Modelling problems
Real-life situations: translate the words into mathematics, solve, then interpret the answer.
🥤 Filling machine
Bottles hold $N(500,4^2)$ mL. What percentage hold less than $492$ mL?
Show the solution
- $z$-score.$$z=\frac{492-500}{4}=-2$$
- Below $-2\sigma$.$$2.5\%$$
✔ About $2.5\%$.
🎓 Scholarship cut-off
Test scores are $N(72,8^2)$. The top $16\%$ get a scholarship. Find the cut-off score.
Show the solution
- Top $16\%$ is above $\mu+\sigma$.$$z=1$$
- Convert.$$x=72+8=80$$
✔ $80$.
💡 Bulb lifetime
Bulbs last $N(1000,50^2)$ h. Out of $2000$ bulbs, about how many last between $950$ and $1100$ h?
Show the solution
- Standardise.$$-1\lt Z\lt2$$
- Probability and count.$$0.34+0.475=0.815,\quad0.815\times2000=1630$$
✔ About $1630$ bulbs.
🚪 Door height
Men's heights are $N(175,7^2)$ cm. How tall must a door be so that $99\%$ of men pass without bending? ($z_{0.99}=2.33$)
Show the solution
- Reverse formula.$$x=175+2.33(7)\approx191.3$$
✔ About $192$ cm.
🚌 Late for school
A commute takes $N(30,5^2)$ min. If you leave $40$ min before school starts, how likely are you to be late?
Show the solution
- Standardise.$$z=\frac{40-30}{5}=2$$
- Upper tail.$$P(Z\gt2)\approx0.025$$
✔ About $2.5\%$.
⚙️ Bolt diameters
Diameters are $N(10,0.02^2)$ mm and must lie in $10\pm0.05$. What fraction is defective? ($P(Z\lt2.5)=0.9938$)
Show the solution
- Standardise.$$z=\pm\frac{0.05}{0.02}=\pm2.5$$
- Both tails.$$2(1-0.9938)=0.0124$$
✔ About $1.2\%$.
🏆 Challenge problems
Harder problems for students aiming higher. Try first, then open the solution.
🏆 Find $\mu$ and $\sigma$
$X$ is normal with $P(X\lt40)=0.16$ and $P(X\gt70)=0.025$. Find $\mu$ and $\sigma$.
Show the solution
- Empirical rule.$$40=\mu-\sigma,\qquad70=\mu+2\sigma$$
- Subtract.$$3\sigma=30$$
✔ $\sigma=10$, $\mu=50$.
🏆 Average of a sample
Heights are $N(170,10^2)$. The mean of $25$ people has $\sigma_{\bar x}=\frac{10}{\sqrt{25}}$. Find $P(\bar x\gt174)$.
Show the solution
- Standard error.$$\sigma_{\bar x}=2$$
- Standardise.$$z=\frac{174-170}{2}=2\Rightarrow P\approx0.025$$
✔ About $2.5\%$.
✅ Multiple-choice quiz
📝 Short essay
Two students scored $80$ in Mathematics ($\mu=70,\sigma=5$) and $85$ in Physics ($\mu=75,\sigma=10$). In which subject did the student do relatively better? Explain.
Compare $z$-scores: Mathematics $z=\frac{80-70}{5}=2$, Physics $z=\frac{85-75}{10}=1$. The mathematics score is $2$ standard deviations above the mean, a much rarer result, so the student did relatively better in Mathematics.