Tuesday, February 10, 2009

Measuring dispersion

Yesterday we talked about measuring central tendency. This is a good statistic, but it is even better when it is partnered with its side kick: dispersion.

Consider the diagram:

As we can see, the two distributions of data have similar central tendencies, but different amounts of dispersion.

One way to measure dispersion is to take the average of each value in the data set subtracted from the mean, using the formula shown in the picture.

Notice the symbol for summation of the absolute value of all observations subtracted from the mean.

No comments: