The average magnitude of deviations from the centre of a normal curve, obtained by squaring all the deviations, calculating their mean, and then finding the square root of the mean. It differs from the *mean deviation
in that it removes difficulties introduced by sign, i.e. it does not matter which side of the centre of the curve any particular deviation is situated. The value so estimated, σ, is the point of maximum slope either side of the central line of the normal curve. If it were possible to make an infinitely large number of observations then the mean deviation would be as good a way of finding the true value of σ as the standard deviation. However with a limited number of samples, as always occurs in practice, the standard deviation provides a better estimate. Often the estimated standard deviation is represented by the Latin letter s while the true standard deviation is represented by its Greek equivalent, σ.