MSSD (Mean of the squared successive differences)

Used as an estimate of variance. It is calculated by taking the sum of the differences between consecutive observations squared, then taking the mean of that sum and dividing by two.

Two common applications are:

·    Basic Statistics - A common application for the mean square successive difference is a test to determine whether a sequence of observations is random. In this test, the estimated population variance is compared with MSSD.

·    Quality Control - Another application is as an estimate of variance when the subgroup size is 1. For the Individuals, CUSUM, EWMA, and Moving Average Control Charts, Minitab estimates standard deviation using the average moving range method by default. For cases when you can't assume that two successive points form a rational subgroup and use the moving range methods, the MSSD method provides an alternative. To use as an estimate of standard deviation, take the square root of MSSD.

Calculating MSSD

For example, suppose you are collecting data on a machine filling vials of the MMR vaccination. You want to make sure the machine dispenses randomly, that is without any special cause of variation.

The fill volumes of 12 vials are:

0.50ml

0.48ml

0.49ml

0.50ml

 

 

 

 

0.505ml

0.50ml

0.49ml

0.498ml

 

 

 

 

0.50ml

0.479ml

0.49ml

0.51ml

 

 

 

 

 

MSSD =

S (Xi + 1 - Xi)2

= 0.00008

 2 (n - 1)

To perform this calculation by hand, subtract 0.48ml from 0.50 ml for the first difference (0.02). Subtract 0.49ml from 0.48ml for the second difference (-0.01). Continue until you have 11 differences. Square each of the differences then add them all up. Divide this sum by 22, or 2 times (n-1).