Analysing data

Abstract

The main purpose of statistical work is to derive summaries of some aspect of data, and in this way to reduce complexity in data without losing structure which is critical. Summary measures have to work hard and we cannot expect that a single index of structure will tell everything about a set of data. Summary measures are specialised tools; they capture one aspect of structure, so that it is generally necessary to produce several measures from the same data set, one for each aspect of structure. Simple examples are the arithmetic mean or median as a summary of location (or centredness) and the standard deviation or interquartile range as a summary of spread around a centre.