ABSTRACT
Measures of central tendency, on their own, provide an insufficient summary of a set of data. Distributions may have, for example, the same mean and median, but the spread of the data may differ substantially between data-sets. For example, consider these distributions:
A) −12, −8, −4, 0, 4, 8, 12 B) −60, −40, −20, 0, 20, 40, 60 C) −120, −80, −40, 0, 40, 80, 120 D) −60, −40, −20, 0, 1, 1, 118
Although the median and the mean are the same in the four distributions (i.e., 0), the data are relatively compact around the mean in distribution A, while their spread around the mean is more pronounced in distributions B, C and D. Therefore, to have an accurate summary of a distribution of data, measures of central tendency need to be complemented by measures of dispersions of the data. The description of some of the most common indices of dispersion follows.
