Locate a value within a model

Try

A score of 74 sounds different depending on its reference group. In a group with mean 62 and standard deviation 8, it is 12 points above the mean, or 1.5 standard deviations. State both descriptions. The second makes the distance relative to the group’s variation, which can help compare positions on different scales.

Predict

Predict whether converting every observation to a z score changes a skewed distribution into a bell-shaped one. A shift and positive rescaling preserve shape. Also predict the probability of one exact value in a continuous model: a tall density at a point does not give that point positive area.

Experiment

Use z=(x−μ)/σ, where x is a value, μ the reference mean and σ the positive reference standard deviation. For x=74, μ=62 and σ=8, z=12/8=1.5. Reverse the calculation with x=μ+zσ. A duration model with μ=40 minutes and σ=6 minutes puts z=−2 at 28 minutes.

Now assume a standard normal model Z. A supplied table gives P(Z≤1.2)=0.8849. The upper-tail probability is 1−0.8849=0.1151. Sketch the bell curve, mark 1.2 to the right of zero, and shade the smaller right-hand region. The sketch should agree with the probability being below one half.

Understand

Standardization can describe relative position for any distribution with a finite positive standard deviation. It becomes a probability tool only when combined with an appropriate distributional model. If X is normal with mean μ and standard deviation σ, then Z=(X−μ)/σ is standard normal: mean zero and standard deviation one. The variance parameter in some notation is σ², so read the convention before entering numbers.

A normal density is symmetric and bell-shaped, with total area one. Interval probabilities are areas, not density heights. For a continuous random variable P(X=x)=0 at every single specified x, and including or excluding an interval endpoint does not change its probability. This does not mean that the variable never takes a value; probability zero and logical impossibility are different concepts in continuous models.

The approximate 68–95–99.7 rule describes the fractions within one, two and three standard deviations of the mean for a normal distribution. It is not a universal theorem about all datasets. Strong skew, mixtures, outliers or a nearby physical boundary can make normal tail predictions implausible even when a mean and SD are available.

A percentile asks for a value below which a specified proportion lies. If a supplied standard-normal quantile is z=1.645 for the 95th percentile approximately, a model with μ=100 and σ=10 gives x≈116.45. The arithmetic is exact conditional on the supplied rounded quantile; the resulting probability statement remains a model approximation.

Comparing z scores across examinations compares positions within reference groups, not necessarily knowledge across subjects. Changing the reference population can change a z score without changing the person’s raw score. Document which mean and SD you used, and distinguish population parameters from sample estimates when those are not known.

Apply

Choose one above-mean and one below-mean value and standardize both. Check the signs against their positions on a number line. For the waiting-time counterexample in practice, describe why a right-skewed shape challenges a normal model rather than merely saying that the data are “bad.” Keep model criticism tied to the question, especially when it concerns a rare tail event.

Use the empirical rule, then check its scope

Assume a normal measurement model with mean 80 units and standard deviation 6 units. Predict the approximate proportion between 68 and 92 before calculating: the endpoints are two standard deviations from the mean, so the empirical rule gives about 95%. Symmetry leaves about 2.5% above 92. Between 86 and 92 lies about (95%−68%)/2=13.5%. These are the rule’s rounded percentages, not exact numerical normal-CDF values.

Changed-case check. For a normal model with mean 30 minutes and standard deviation 4 minutes, give the approximate proportion between 26 and 38 minutes and the approximate proportion below 18 minutes. Draw the intervals and explain why the same numbers cannot be inferred from a severely right-skewed dataset merely because it has that mean and standard deviation.

Feedback

The first interval runs from z=−1 to z=2. Its area is approximately 68%+(95%−68%)/2=81.5%. The second cutoff is z=−3, leaving approximately (100%−99.7%)/2=0.15% below it. A correct answer states the normal-model assumption, shows standardization and shaded regions, distinguishes approximate rule values from exact CDF outputs, and refuses to apply the normal percentages automatically to the skewed case.

This explicit action supplements the existing general normal calculations for current AP Statistics identifier 2.11.C (Fall 2026 CED), checked 30 October 2026: official framework.

Continue

Next move from one variable to paired observations. A scatterplot keeps the pairings that separate histograms cannot show.