The Normal Distribution
Turn a bell curve into useful answers. Learn to shade the right region, find probabilities and percentile cutoffs, and explain what a result means for observation.
By the end of this lesson, you should be able to:
- Describe a normal model using its shape, mean and standard deviation.
- Estimate proportions using the rule.
- Standardize a value and interpret its -score.
- Calculate probabilities for tails and intervals.
- Find percentile cutoffs, central intervals and extreme-tail boundaries.
- Compare relative positions and recover parameters from given percentiles.
Before you start: Know that probability distributions describe possible outcomes and their chances. Review Topic 2.10: The Binomial Distribution to contrast discrete counts with continuous measurements.
First time learning this? Follow the practice-time model through its curve, probabilities and cutoffs.
Here to revise? Use the method checklist, then try the practice before opening the solutions.
The concept in 60 seconds
A normal distribution is a continuous, symmetric, single-peaked bell curve. Its mean tells you where the center is; its standard deviation tells you how spread out the values are.
For a continuous model, probability is an area under the curve over a region. A taller point on the curve does not itself give the probability of that exact value.
Two directions, two questions:
“What proportion finishes before minutes?” gives a value and asks for an area.
“How quickly do the fastest finish?” gives an area and asks for a value.
We will use an original fictional teaching model: the time , in minutes, for student to finish a practice activity is modeled as normal with mean minutes and SD minutes. The normal assumption is supplied for this lesson; it is not a claim about real student records.
Quick check: is every symmetric distribution normal?
No. A symmetric uniform distribution is flat, and some symmetric distributions have multiple peaks. Symmetry alone does not justify a normal model. Check the shape and the information supplied by the question.
Shape, mean and standard deviation
Continuous random variable
A measurement can take any value within its modeled range, including decimals. Completion times such as and minutes are possible.
Contrast this with a binomial count, which takes integer values.
Mean : the center
locates the center of the normal curve. Mean, median and mode all coincide there.
Here minutes; is on each side.
SD : the spread
is positive and uses the same units as . A larger SD gives a wider, lower curve; a smaller SD gives a narrower, taller curve.
Here minutes. Variance .
All possible outcomes together have probability . Increasing the curve’s width does not increase its total probability.
Curve height measures density; region area measures probability.
Write the model clearly as . If your course uses , this is . Some texts use , so always label whether the second number is the SD or the variance.
Solid and dashed curves distinguish the models. Both panels use the same numerical scales. Every complete normal curve has area ; the display shows a finite portion of its tails.
What can justify a normal model?
Use the normal assumption when a problem gives it. When you are assessing data, a roughly symmetric, unimodal bell-shaped distribution without strong skew or extreme outliers can support an approximately normal model. A small data set can look irregular even when a normal model is reasonable; shape is evidence to assess, not automatic proof.
A normal curve has tails that extend indefinitely. A real measurement may have physical limits, so a normal model is an approximation to reality and should be suitable for the range being studied.
Keep the variable straight: This lesson models student’s completion time. Its SD is minutes. Do not replace it with (where is the sample size); the distribution of a sample mean is a different topic.
Quick check: what are the mean and SD of the standard normal distribution?
The standard normal distribution has mean and SD . It describes standardized values rather than completion times in minutes.
The rule
For a normal distribution, about of observations lie within SD of the mean, about within SDs, and about within SDs. These are useful approximations.