As a statistical term, this is pretty close to its everyday meaning: how much data items vary from each other – are they all quite close, or do they vary over a wide range of values? There are various measures of variability, from the absolute range (maximum–minimum) to the interquartile range, or the ubiquitous standard deviation (s.d., σ). For a theoretical distribution you can think of this as how widely spread the graph is, or how variable you would expect data that conforms to the distribution to be.
Defined on page 30
Used on pages 15, 19, 21, 24, 29, 30, 31, 32, 36, 45, 107, 108, 109, 110, 118, 122