# The Normal Distribution

> AP Statistics · AP Stats 2024-2026
> Source: https://www.owlsprep.com/study/ap-statistics-u11-the-normal-distribution/

This module covers normal distribution properties, the 68-95-99.7 rule, z-score standardization, probability calculations, and inverse normal percentile finding for AP exam success.

**Prerequisites:** [Basic descriptive statistics (mean, standard deviation)](https://www.owlsprep.com/study/ap-statistics-u1-descriptive-measures/); [Continuous random variable fundamentals](https://www.owlsprep.com/study/ap-statistics-u10-continuous-random-vars/)

## Learning objectives

- Identify core symmetric, bell-shaped properties of normal distributions and the 68-95-99.7 empirical rule
- Calculate z-scores to standardize raw normal values to the standard normal distribution
- Use normal cumulative distribution function (CDF) to find interval and tail probabilities for normal random variables
- Apply inverse normal calculations to find percentile cutoff values for specified areas under the normal curve

## Core Properties of the Normal Distribution

Normal distributions are the most widely used probability model in statistics, as they naturally describe many real-world datasets like heights, test scores, and measurement errors. All normal distributions are symmetric, unimodal, and asymptotic, meaning the curve never touches the horizontal x-axis.

**Normal Distribution** — A continuous probability distribution where the total area under the curve equals 1, and the mean, median, and mode are all equal at the center of the bell shape.

*Notation:* $X \sim N(\mu, \sigma^2)$

> **Empirical Rule Mnemonic**
>
> Remember 68-95-99.7: 68% of data within 1σ of μ, 95% within 2σ, 99.7% within 3σ. No extra memorization needed for quick MCQ estimates.

**Worked example:** Scores on a standardized exam follow a normal distribution with μ=500 and σ=100. Use the empirical rule to estimate the range that contains the middle 95% of all scores.

1. Step 1: Recall the 95% interval for normal data falls 2 standard deviations above and below the mean.
2. Step 2: Calculate lower bound: $\mu - 2\sigma = 500 - 2*100 = 300$

   $$300$$
3. Step 3: Calculate upper bound: $\mu + 2\sigma = 500 + 2*100 = 700$

   $$700$$
4. Final answer: The middle 95% of scores lie between 300 and 700.

**Check your understanding**

Test your understanding of the empirical rule:

1. What percentage of normal data lies above μ + 2σ?

   - 2.5%
   - 5%
   - 95%
   - 16%

   *Why:* Since 95% of data is within ±2σ, the remaining 5% is split equally between the two upper and lower tails.

*Calculator:* allowed

## Z-Scores and Standard Normal Standardization

Z-scores let you convert any normal distribution to the standard normal N(0,1) distribution, so you can use shared probability tables or calculator functions for all normal problems instead of memorizing unique formulas for every possible mean and standard deviation.

**Z-Score** — The number of standard deviations a raw observation $x$ is from the population mean.

$$z = \frac{x - \mu}{\sigma}$$

**Worked example:** A student scores 720 on the standardized exam from the previous example (μ=500, σ=100). Calculate their z-score.

1. Step 1: Identify given values: $x=720$, $\mu=500$, $\sigma=100$
2. Step 2: Substitute into z-score formula:

   $$z = \frac{720 - 500}{100} = 2.2$$
3. Final answer: The student's score is 2.2 standard deviations above the mean.

> **tip**
>
> AP graders deduct points for z-scores rounded to 1 decimal place. Always keep 2 or 3 decimal places for z-scores to avoid calculation errors.

*Calculator:* allowed

## Normal CDF for Probability Calculations

The normal cumulative distribution function (CDF) returns the total area under the normal curve to the left of a specified x value, which equals the probability that a random observation is less than that x value. You can extend this to find interval or upper tail probabilities with simple arithmetic.

**Exam command terms**

AP exam questions use specific cues for normal probability problems:

- **Find the proportion of scores** — You are being asked to calculate a normal CDF probability

- **Estimate the percentage of observations** — You may use the empirical rule for a quick approximate answer

**Worked example:** Using the same exam distribution N(500, 100²), find the probability a random student scores between 600 and 700.

1. Step 1: Recognize P(600 < X < 700) = P(X < 700) - P(X < 600)
2. Step 2: Use normal CDF to find P(X < 700): that is the area left of 700, equal to 0.9772
3. Step 3: Use normal CDF to find P(X < 600): that is the area left of 600, equal to 0.8413
4. Step 4: Subtract to get the interval probability:

   $$0.9772 - 0.8413 = 0.1359$$
5. Final answer: ~13.6% of students score between 600 and 700.

*Calculator:* allowed

## Inverse Normal for Percentile Cutoffs

Inverse normal is the reverse operation of the normal CDF: you input a cumulative left area, mean, and standard deviation, and it returns the raw x value that corresponds to that percentile. This is used to find cutoff scores, margin of error bounds, and critical values for hypothesis tests later in the course.

**Worked example:** For the N(500, 100²) exam distribution, find the 90th percentile score, the value where 90% of all test takers score lower.

1. Step 1: Confirm you are given cumulative left area = 0.9, μ=500, σ=100
2. Step 2: Run inverse normal on your graphing calculator with these parameters
3. Step 3: The output returns x ≈ 628.16
4. Final answer: The 90th percentile exam score is approximately 628.

> **warning**
>
> Never input upper tail areas directly into inverse normal. Always convert upper tail p-values to 1-p first to get the correct x value.

*Calculator:* graphing_only

## Common pitfalls

- **Wrong:** Confusing normal notation $N(\mu, \sigma^2)$ as mean followed by standard deviation
  - Why it fails: This leads you to use the full standard deviation value as the variance, producing z-scores that are half their correct size
  - Correct: Explicitly write out the mean and standard deviation separately before starting any calculation to avoid parameter mixups
- **Wrong:** Applying the normal model to strongly skewed discrete data
  - Why it fails: Normal distributions are symmetric and continuous, so probability results will be wildly inaccurate for skewed datasets
  - Correct: Always confirm the data is approximately symmetric with no extreme outliers before using a normal model
- **Wrong:** Forgetting to subtract left cumulative area from 1 for upper tail probabilities
  - Why it fails: Default normal CDF functions return left-side area, so unadjusted upper tail results will be greater than 0.5 and completely wrong
  - Correct: Label your target area as left, right, or interval before running any CDF calculation
- **Wrong:** Rounding z-scores to 1 decimal place
  - Why it fails: This introduces 10-15% error in final probability values, leading to lost points on AP FRQs
  - Correct: Keep a minimum of 2 decimal places for all z-scores, ideally 3, for intermediate steps
- **Wrong:** Using sample standard deviation s for population σ with no justification
  - Why it fails: College Board rubrics deduct partial credit for unstated assumptions about population parameters
  - Correct: Explicitly state that you are assuming the given standard deviation is the true population value in your solution

## Cheatsheet

| Concept | Formula / Rule | AP Exam Use Case |
| --- | --- | --- |
| Empirical Rule | 68% within 1σ, 95% within 2σ, 99.7% within 3σ of μ | Fast MCQ probability estimates |
| Z-Score | $z = \frac{x - \mu}{\sigma}$ | Standardize raw values to the standard normal distribution |
| Normal CDF | $P(a < X < b) = P(X < b) - P(X < a)$ | Find probability a normal variable falls between two values |
| Inverse Normal | Given left area p, find x where $P(X < x) = p$ | Find percentile cutoff values for specified probabilities |

## What's next

Mastery of the normal distribution is the single most important foundational skill for AP Statistics, as it underpins over 50% of the exam content including sampling distributions, confidence intervals, and hypothesis testing. The z-score logic you learned here will be reused directly to calculate test statistics for z-tests and construct confidence intervals for population means. Before moving forward, ensure you can reliably distinguish between when to use normal CDF vs inverse normal, as this distinction is the most common source of avoidable FRQ point loss. Next, you can extend this knowledge to related core topics that build directly on normal distribution properties.

---

From [OwlsPrep](https://www.owlsprep.com) — free study guides for A-Level, IB, AP and IGCSE, written against the official syllabus. Canonical page: https://www.owlsprep.com/study/ap-statistics-u11-the-normal-distribution/