
Median absolute deviation measures spread from the middle twice
Work through median absolute deviation with five numbers, see why an extreme value may leave it unchanged, and check whether a report uses a scaled value.

Work through median absolute deviation with five numbers, see why an extreme value may leave it unchanged, and check whether a report uses a scaled value.

The interquartile range describes the spread between the lower and upper quartiles. Use it alongside the median, sample size and extremes.

Work through a small winsorization example and learn why capping extreme observations changes an average without reducing the number of rows in the dataset.

A panel survey follows participants over time. Check who remains in later waves before interpreting a headline about how people’s lives have changed.

A trimmed mean removes specified observations from the tails before averaging. Work through a small example and learn what to check in a published statistic.

A replacement television may cost more and offer different features. Hedonic adjustment tries to separate those changes when an official price index is compiled.

Odds and probability use different denominators. A worked example shows why a reported odds ratio should not be repeated as a risk ratio in a headline.

Use a two-item example to see how base-period and current-period quantities can produce different price indices from the same price changes.

A worked example explains why a weighted survey percentage can differ from a raw headcount, and what methodology details a reader should check.