In this Introduction to Statistics module
we will consider two characteristics of
a process, the accuracy and the precision.
It is important to be clear about the
distinction between the two. If the process
mean is close to the target then it is
said to be 'accurate'. If the process
has little variation it is said to be
A process can be accurate without being
precise, and precise without being accurate.
The target shows the results of ten shots
from two different shooters:
the difference between the process
mean and the target value
The shooter on the right is most
'accurate'. Although the shots
are widely scattered, the mean
position is close to the target
the amount of variation, or 'scatter'
The shooter on the left has least
variation between shots and is
thus more 'precise'.
The mean of most processes can be adjusted
fairly easily and most process improvement
effort goes to reducing the amount of
variation. Variation is the cause of most
customer dissatisfaction, a customer who
expects the salinity of a vial to be 50
will probably not be happy if the salinity
is only 40; it may not work as well or
at all. They will not care that the process
average is 50.
Variation also makes a process hard to
manage because it impacts the next step.
A process that relies on the salinity
level being correct may take longer, need
to be reworked or even develop into a
serious problem that requires extraordinary
The Six Sigma process improvement approach,
used by many major companies, is named
after the Greek letter σ
, or 'sigma'. This is used in statistics
to represent the 'standard deviation',
the most common measure of variation.
The Six Sigma approach aims to reduce
variation to the stage where there are
fewer than 3.4 defects for every million
opportunities for that defect.