Distributions (Understanding Shape)

Distributional shape describes how values are spread, including center, dispersion, skew, tails, and possible modes. These features can reveal consequential differences hidden by the same average.

A distribution describes how values or probability are spread across possible outcomes. Center, dispersion, skew, tails, and multiple peaks answer different questions. Two datasets with the same mean can imply very different waiting experiences, resource needs, or chances of an extreme outcome.

Shape is an estimate as well as a concept. Histograms depend on binning, small samples can create apparent peaks, and a bimodal pattern does not by itself establish two distinct populations. Choose the features relevant to the decision and inspect how robust they are rather than force the data into a familiar named distribution.

When to use it

When summary statistics are hiding crucial distributional information; when the shape of the data would change strategy or risk assessment; when fat tails indicate extreme event risks that averages conceal; when bimodal distributions reveal that 'the average' represents no actual group.

How it can help

Inspect the part of the distribution relevant to the decision, check plotting and sampling choices, and compare appropriate quantiles or probabilities alongside averages.