01 / 21

The Normal
Distribution

Master the bell curve that underlies nearly all of statistics: its shape, its parameters \(\mu\) and \(\sigma\), the Empirical Rule, and how to compute probabilities with z-scores.

AP Statistics

What Is a Normal Distribution?

A normal distribution is a continuous probability distribution that is symmetric about its mean. It produces the famous bell-shaped curve where data clusters around the center and tapers off equally on both sides.

\(\mu\) : Mean
The center of the distribution: where the peak occurs.
\(\sigma\) : Standard Deviation
Controls the spread: larger \(\sigma\) means a wider, flatter curve.

The Bell Curve Visualized

Why Does It Matter?

The normal distribution appears everywhere in nature and statistics. Understanding it lets you model real data, perform hypothesis tests, and construct confidence intervals.

Natural Data
Heights, test scores, blood pressure, and measurement errors often follow normal distributions.
Central Limit Theorem
Sample means of any distribution approach normality as \(n\) grows: the foundation of inference.
Statistical Tests
Z-tests, t-tests, and many confidence intervals rely on the normal model.

The Probability Density Function

1
\(\mu\) shifts the curve left or right along the x-axis.
2
\(\sigma\) controls width: small \(\sigma\) = tall and narrow, large \(\sigma\) = short and wide.
3
The coefficient \(\frac{1}{\sigma\sqrt{2\pi}}\) ensures the total area equals 1.

The Empirical Rule: 68-95-99.7

68%
Within \(\mu \pm 1\sigma\)
95%
Within \(\mu \pm 2\sigma\)
99.7%
Within \(\mu \pm 3\sigma\)

Standard vs General Normal

Standard Normal \(Z\) General Normal \(X\)
Mean \(\mu = 0\) Any \(\mu\)
Std Dev \(\sigma = 1\) Any \(\sigma > 0\)
Notation \(Z \sim N(0,1)\) \(X \sim N(\mu,\sigma)\)
Table lookup Direct: use z-table Convert to \(Z\) first

How to Apply z-Scores

1
Identify \(\mu\) and \(\sigma\) from the problem context.
2
Write the probability statement in terms of \(X\): e.g., \(P(X < 72)\).
3
Convert to z: \(z = \dfrac{x - \mu}{\sigma}\).
4
Look up the z-value in the standard normal table (or use a calculator).
5
Interpret the result in context: the area under the curve is the probability.

Example 1: Finding a Probability

SAT Math scores follow \(X \sim N(500, 100)\). What is the probability a student scores below 620?

1
Parameters: \(\mu = 500\), \(\sigma = 100\).
2
z-score: \(z = \dfrac{620 - 500}{100} = 1.20\).
3
Table lookup: \(P(Z < 1.20) = 0.8849\).

Example 2: Between Two Values

Heights of adult men: \(X \sim N(70, 3)\) inches. Find \(P(67 < X < 73)\).

1
Lower z: \(z_1 = \dfrac{67 - 70}{3} = -1.00\).
2
Upper z: \(z_2 = \dfrac{73 - 70}{3} = 1.00\).
3
Probability: \(P(-1 < Z < 1) = 0.8413 - 0.1587 = 0.6826\).

Example 3: Inverse Normal (Finding \(x\))

A factory produces bolts with diameters \(X \sim N(10, 0.2)\) mm. What diameter separates the top 5%?

1
Find the z-value: top 5% means area to the left is 0.95. From the table: \(z = 1.645\).
2
Back-transform: \(x = \mu + z \cdot \sigma = 10 + 1.645(0.2) = 10.329\).

Key Properties of the Curve

Symmetric
Mean = Median = Mode. The curve is a mirror image about \(\mu\).
Asymptotic
The tails approach but never touch the x-axis: extends from \(-\infty\) to \(+\infty\).
Area = 1
The total area under the curve is exactly 1, representing 100% probability.
Unimodal
Exactly one peak (mode) at \(x = \mu\). No other local maxima.

Interpreting z-Scores

How to Check for Normality

1
Histogram: look for a roughly symmetric, bell-shaped distribution.
2
Normal Probability Plot (QQ-plot): points should fall close to a straight line.
3
Empirical Rule check: about 68% of data within \(\pm 1\sigma\), 95% within \(\pm 2\sigma\).
4
Skewness and kurtosis: for a normal distribution, skewness \(\approx 0\) and excess kurtosis \(\approx 0\).

Common Mistakes

✕ Confusing \(\sigma\) and \(\sigma^2\)
The z-score formula uses \(\sigma\) (standard deviation), not \(\sigma^2\) (variance). Dividing by variance gives a z-score that is too small.
✕ Forgetting Complements
For "greater than" probabilities, compute \(P(Z > z) = 1 - P(Z < z)\). Many students read the table and stop.
✕ Assuming All Data Is Normal
Skewed data (income, wait times) and discrete counts are not normal. Always check the distribution shape first.

Summary Reference Table

Concept Formula / Rule Key Detail
z-score \(z = \dfrac{x - \mu}{\sigma}\) Standardizes any normal \(X\)
Inverse normal \(x = \mu + z\sigma\) Finds \(x\) from a percentile
68-95-99.7 \(\mu \pm 1\sigma,\; 2\sigma,\; 3\sigma\) Approximate areas quickly
\(P(a < X < b)\) \(P(Z < z_b) - P(Z < z_a)\) Subtract cumulative areas
\(P(X > a)\) \(1 - P(Z < z_a)\) Complement rule

Question 1 of 5

If \(X \sim N(100, 15)\), what is the z-score for \(x = 130\)?

Question 2 of 5

According to the Empirical Rule, approximately what percent of data falls within \(\mu \pm 2\sigma\)?

Question 3 of 5

For a standard normal distribution, \(P(Z < 0)\) equals:

Question 4 of 5

Test scores follow \(X \sim N(75, 8)\). What score is at the 84th percentile?
(Hint: \(P(Z < 1.00) \approx 0.84\))

Question 5 of 5

Which statement about the normal distribution is false?