Normal distribution explained simply

By Jude Wallis · Published

A normal distribution is a symmetric, bell-shaped, single-peaked distribution that is described completely by its mean μ\mu and standard deviation σ\sigma. About 68% of values fall within 1 standard deviation of the mean, about 95% within 2, and about 99.7% within 3.

AP Statistics: Unit 2 (topics 2.11 The Normal Distribution). The normal distribution is Unit 2, topic 2.11 in the Fall 2026 AP Statistics course, and normal calculations are used again for sampling distributions in topic 2.12 and for inference in Unit 3.

What a normal distribution is

A normal distribution is a model for a quantitative variable whose values pile up around a center and thin out evenly on both sides. Its graph is the normal curve, often called the bell curve.

Every normal curve has the same features:

  • Symmetric. The left half mirrors the right half.
  • Single-peaked. The peak is at the mean, which is also the median.
  • Bell-shaped. The curve falls away from the peak and gets closer and closer to the horizontal axis without touching it.
  • Total area of 1. A normal curve is a density curve, so area under it represents proportion or probability.

Because the variable is continuous, probability belongs to intervals. The area above a single exact value is 0, so P(X<40)P(X < 40) and P(X≤40)P(X \leq 40) are equal.

The two parameters and the z-score formula

A normal distribution is fixed by two numbers:

  • μ\mu ("mu"), the mean, locates the center of the curve.
  • σ\sigma ("sigma"), the standard deviation, sets the spread. A larger σ\sigma gives a wider, flatter curve. A smaller σ\sigma gives a narrower, taller one.

To find areas, convert a value xx to a z-score:

z=x−μσz = \frac{x - \mu}{\sigma}
  • xx is the value of interest.
  • zz is the number of standard deviations xx lies above or below the mean.

The z-scores of a normal variable follow the standard normal distribution, the normal distribution with mean 0 and standard deviation 1. One table or one calculator function then handles every normal distribution. For more on standardizing, see z-score explained.

The 68-95-99.7 rule

In any normal distribution:

IntervalApproximate share of values
Within 1 standard deviation of the mean, μ±σ\mu \pm \sigma68%
Within 2 standard deviations, μ±2σ\mu \pm 2\sigma95%
Within 3 standard deviations, μ±3σ\mu \pm 3\sigma99.7%

This is the empirical rule. If battery life is normal with μ=48\mu = 48 hours and σ=5\sigma = 5 hours, about 95% of batteries last between 48−10=3848 - 10 = 38 and 48+10=5848 + 10 = 58 hours.

Symmetry lets you split the leftovers. About 5% of values lie outside 2 standard deviations, so about 2.5% lie above μ+2σ\mu + 2\sigma and about 2.5% lie below μ−2σ\mu - 2\sigma. The empirical rule guide and the empirical rule calculator cover this in detail.

How to find a normal probability

For cutoffs that are not a whole number of standard deviations from the mean, use z-scores with a table or technology.

  1. Sketch the curve, mark the mean, and shade the region the question asks about.
  2. Standardize each boundary with z=x−μσz = \frac{x - \mu}{\sigma}.
  3. Find the area. The z-table gives the area to the left of a z-score. For an area to the right, subtract from 1. For an area between two values, subtract the smaller left-area from the larger.
  4. Answer in context.

Going the other way, from a percentile to a value: find the z-score with that area to its left, then compute x=μ+zσx = \mu + z\sigma.

On a TI-84 these are normalcdf and invNorm, explained in normalcdf and invNorm on the TI-84. The normal distribution calculator does both directions and shows the shaded region.

When the normal model is used

  • Measurements that cluster symmetrically. Many physical measurements and standardized test scores are modeled as approximately normal.
  • Sampling distributions. For a large enough random sample, the sampling distribution of the sample mean is approximately normal, whatever the shape of the population. That result is the central limit theorem. A sample proportion is approximately normal when the expected counts of successes and failures are both at least 10.
  • Inference. z procedures for proportions rest on the normal model.

A normal model is a choice you have to justify. Incomes, wait times, and house prices are usually right-skewed, and a normal calculation on skewed data gives wrong answers. Check a graph of the data, or look for a statement in the problem that the distribution is approximately normal.

Common mistakes

  1. Assuming every distribution is normal. The 68-95-99.7 rule and the z-table apply only to approximately normal distributions.
  2. Reading the wrong tail. Table values are areas to the left. For "more than" questions, subtract from 1.
  3. Skipping the sketch. A quick sketch catches answers that are on the wrong side of 0.5. An area below a value that is under the mean has to be less than 0.5.
  4. Confusing the z-score with the probability. z=−1.6z = -1.6 is a position. The area to its left, 0.0548, is the probability.
  5. Using variance in place of standard deviation. If a problem gives σ2\sigma^2, take the square root first.
  6. Giving calculator syntax as the work. Write the distribution, its mean and standard deviation, and the boundaries, so a reader can follow the setup without knowing the keystrokes.
  7. Treating "normal" as "typical". In statistics, normal names a specific family of curves.

The normal distribution in AP Statistics

The normal distribution is Unit 2, topic 2.11 in the Fall 2026 AP Statistics course. Unit 2 carries 15-25% of the multiple-choice section.

It also supports later units. Topic 2.12 introduces sampling distributions and the central limit theorem, and the inference procedures for proportions in Unit 3 use normal calculations.

Be ready to find a proportion below, above, or between values, to find the value at a given percentile, to interpret both in context, and to state why a normal model is reasonable. Try the normal curve explorer, then the normal distribution practice set.

Battery life: below, between, and a percentile

The life of a brand of rechargeable battery is approximately normal with mean μ=48\mu = 48 hours and standard deviation σ=5\sigma = 5 hours. (a) What proportion of batteries last less than 40 hours? (b) What proportion last between 45 and 55 hours? (c) How long does a battery at the 90th percentile last?

  1. (a) Standardize 40: z=40−485=−1.6z = \frac{40 - 48}{5} = -1.6.

  2. (a) The area to the left of z=−1.6z = -1.6 is 0.0548. About 5.5% of batteries last less than 40 hours.

  3. (b) Standardize both boundaries: z=45−485=−0.6z = \frac{45 - 48}{5} = -0.6 and z=55−485=1.4z = \frac{55 - 48}{5} = 1.4.

  4. (b) Areas to the left: 0.9192 for z=1.4z = 1.4 and 0.2743 for z=−0.6z = -0.6. Subtract: 0.9192−0.2743=0.64490.9192 - 0.2743 = 0.6449. About 64.5% of batteries last between 45 and 55 hours.

  5. (c) Find the z-score with area 0.90 to its left. From the table, z≈1.28z \approx 1.28.

  6. (c) Convert back: x=μ+zσ=48+1.28(5)=54.4x = \mu + z\sigma = 48 + 1.28(5) = 54.4 hours.

  7. Check with the empirical rule: 40 hours is between 1 and 2 standard deviations below the mean, so the area below it should be between 2.5% and 16%. The answer 5.5% fits.

(a) About 0.0548, or 5.5%. (b) About 0.6449, or 64.5%. (c) About 54.4 hours: 90% of batteries last less than that.

Frequently asked questions

What is a normal distribution in simple terms?

It is a bell-shaped distribution in which values cluster symmetrically around the mean and become steadily rarer the farther they are from it. Two numbers describe it fully: the mean, which sets the center, and the standard deviation, which sets the spread.

What are the mean and standard deviation of the standard normal distribution?

The standard normal distribution has mean 0 and standard deviation 1. It is the distribution of z-scores for any normal variable, which is why one z-table works for every normal distribution.

How do you know if data are normally distributed?

Graph the data. A histogram or dotplot that is roughly symmetric, single-peaked, and bell-shaped with no outliers supports a normal model. You can also compare the data with the 68-95-99.7 rule. In a textbook or exam problem, the statement that a distribution is approximately normal is your justification.

Why is the normal distribution important?

Many measured variables are approximately normal, and, more importantly for inference, the sampling distributions of sample means and sample proportions are approximately normal under conditions that are often met. That is what lets a z-score turn into a probability in confidence intervals and hypothesis tests.