A Box and Whisker Plot — or box plot — compresses an entire distribution into five summary statistics drawn on a common scale. The central rectangle (the "box") spans the first quartile (Q1) to the third quartile (Q3), enclosing exactly the middle 50% of observations — the interquartile range (IQR). A horizontal line inside the box marks the median (Q2): half the data falls above, half below.
The whiskers extend outward from the box edges to the most extreme values that still fall within 1.5× the IQR. Any observation beyond a whisker tip is a statistical outlier, plotted individually as a dot. This five-number summary — whisker minimum, Q1, median, Q3, whisker maximum — answers the questions every distribution poses: where is the center? How spread is it? Is it skewed? Are there extremes?
These distributions show simulated AI performance scores across five cognitive domains — each sampled from 80 observations with domain-specific means and standard deviations. Pattern Recognition has both the highest median and the narrowest box, indicating consistently strong performance. Embodied Teaching has the lowest median; replicating physical, situated instruction remains among the hardest problems in AI.
Moral Judgment and Creative Synthesis show the widest interquartile ranges — high variability means the middle 50% of outcomes spans a broad range. Performance in these domains is context-sensitive and resists characterisation by a single number. Causal Reasoning sits mid-range on both median and spread, reflecting its partial but inconsistent progress in current systems.