Performance values are calculated in three steps to get rid of outliers or typing errors. 

Step 1

All the data are used to calculate the mean, standard deviation (SD) and standard normal deviate (Z). 

Z = the number of SDs a value is away from the mean, e.g. a pH result of 4.5 will score a Z-value of 0.5 if the mean for the data set is 5 and the SD 1

A performance interpretation is then assigned to the data from the range:

  A for Z-values between -1 and 1
  B for Z-values between -1 and -2 or between 1 and 2
  C for Z-values between -2 and -3 or between 2 and 3 
  D for Z-values between -3 and -4 or between 3 and 4 
  E for Z-values between -4 and -5 or between 4 and 5 
  >E for Z-values less than -5  or greater than 5 

Step 2

Only the data in the A performance range are used to re-calculate the mean, standard normal deviate and performance.  These are reported as the 'Adjusted' values.

Step 3

The final calculation of the mean, standard deviation from data in the A performance range in the adjusted data. These values are reported as 'Final'

Interpretation

Performances not in the A range, indicates a result out of line with the other participants, maybe something wrong with their method, apparatus or reagents, etc.

The standardised normal deviate (Z) is unitless and results can be graphed over time for trend analysis purposes

Labs should strive for performances between -1 and +1 and to keep the fitted performance line parallel to the base. Upwards or downwards trends indicate change over time and should be corrected