The key to a good GR&R study lies in the
upfront planning. A GR&R study can be planned
using the following 5 steps:
- Define the objective of the study.
- Define the components of the study.
- Make the measurements.
- Calculate the %GR&R.
- Take action to improve the measurement
Step 1: Define the objective of the GR&R
- This step involves defining both what
instrument (or family of instruments) will be
studied as well as the product or product
features to be studied.
Step 2: Define the components of the study:
The components to consider include the people
(testers or appraisers), the (test) method, the
materials (or samples to test), the (testing)
equipment, and the environment (such as
conditioning requirements and the ambient
Use 2 or 3 people who routinely conduct the
test as the appraisers; do not replace the
typical users with engineers or quality
The test method used should be exactly as
documented. Do not add new steps or precautions.
Regarding test samples, if a non-destructive
test will be used, at least 5 samples (and
typically 10 samples) will be needed.
Ensure that the test equipment has adequate
discrimination for the job. The equipment used
should, of course, be the measuring device
normally used without special repairs or
The environmental aspects of the study include
taking into account special conditioning,
temperature, and/or relative humidity
Step 3: Make the measurements:
It is best for each appraiser to take all the
measurements (including replicate measurements)
for the study before having the next appraiser
start taking measurements.
The samples measured should be randomly
selected and preferably “anonymous” to the
appraiser to reduce bias.
Step 4: Calculate the %GR&R:
This is the heart of the GR&R study. There are
several techniques for calculating the %GR&R.
Part of the planning involves deciding what type
of analysis technique will be used.
R&R Analysis techniques include:
- Analysis for Non-Destructive Measurements
- Analysis using ANOVA
- Analysis for Destructive Measurements
- Analysis for Attribute Measurements
- Analysis Using Graphical Techniques
The %GR&R will tell us if the measurement
system is “good” or if it is “poor” and must be
Step 5: Take action:
If the measurement system is validated as
“good,” additional action may not be necessary.
If it is “poor,” action to improve the
measurement system is needed.