On Computing the Average Deviation from the Mean

In certain computations, notably those involving correlation and regression formulae, the only measure of variability which one uses is the standard deviation. In many other cases, however, the aver age deviation will do as well, but thus far the fate of a certain excellent formula for solving the average deviation has been to go practically unnoticed. By this formula, devised by Truman L. Kelley, the average deviation, which was formerly the hardest of all the measures of variability to reckon, becomes by far the easiest. In the procedure outlined within this article, the same manipulation yields the constants for computing both the mean and the average deviation, respectively a measure of central tendency and a measure of variability. THE FORMULA