Why this chapter matters
Many measured quantities are continuous, and normal approximations provide practical probability calculations and benchmarks.
What you will learn
- Explain the difference between probability at a point and over an interval for continuous data.
- Standardize values with z-scores and interpret relative position.
- Compute normal probabilities and percentiles from model parameters.
Understand the core ideas
Continuous variables describe measured quantities where values can occur on an interval scale, such as time or temperature. In this setting, probability comes from area under a density curve. The total area is one, and probability over an interval is the area between bounds. A key point is that probability at exactly one point is zero in continuous models. That does not mean the value is impossible, only that single-point probability has no area. This difference from discrete models is essential for correct interpretation.
The normal distribution is a widely used continuous model defined by mean and standard deviation. Standardizing with z-scores lets you compare values from different scales and use common probability tables. Normal approximations are often practical when data are roughly bell-shaped and not dominated by extreme outliers, but they are not automatic. Strong skew, heavy tails, or multiple peaks can weaken fit. Good practice combines numeric calculations with visual checks and context judgment before reporting interval probabilities or percentile cutoffs.
Key terms
- Density curve
- A nonnegative curve whose total area equals one and represents a distribution.
- Normal distribution
- A bell-shaped continuous distribution parameterized by mean and standard deviation.
- Z-score
- Standardized value showing how many standard deviations a point is from mean.
- Percentile
- A cutoff below which a specified percentage of observations fall.
Upper-tail probability from z-score
Scores are normal with mean 70 and standard deviation 8. Find probability of scoring above 82.
- Standardize 82: z = (82 - 70) / 8 = 1.5.
- Look up cumulative probability P(Z <= 1.5) approximately 0.9332.
- Convert to upper tail: 1 - 0.9332 = 0.0668.
- Interpret as expected long run proportion above 82 under this model.
A common misconception
Claim: Any quantitative dataset should be analyzed with a normal model.
Correction: Normal methods need reasonable shape conditions; strong skew or outliers can make normal approximations unreliable.
Lessons in this chapter
- Continuous random variablesInterpret area under a density curve as probability.
- Normal distributions and z-scoresConvert between raw values and standardized scores.
- Normal probability calculationsFind interval probabilities and cutoff values using normal models.
- Model fit and reasonablenessCheck whether a normal model is plausible from data shape and context.
Study task
Chapter checkpoint
What does a z-score of -1.5 mean?
The value is 1.5 standard deviations below the distribution mean.