Sampling And Non-Sampling Errors

Whenever we estimate a population's characteristics from a sample, errors can creep in. A population of numerical values has two features of interest here: its Central Tendency, measured by the mean, median or mode; and its Dispersion, measured by the standard deviation, mean deviation, range and so on. Because a sample only estimates these, its results may differ from the truth.

Errors are of two broad kinds:

• Sampling errors — the gap between a sample estimate and the true population value, arising simply because we studied a part rather than the whole. …