Gage R&R is all about the integrity of your measurement system and its ability to capture the true values of what you are measuring. A few quotes on measurement and data might be appropriate:
“In God we trust. All others must bring data.” — W. Edwards Deming
“One accurate measurement is worth one thousand expert opinions.” — Grace Hopper
“It is really just as bad a technique to make a measurement more accurately than is necessary as it is to make it not accurately enough.” — Arthur David Ritchie
This article will present what is a Gage R&R study, some definitions and underlying statistics, the benefits of doing a Gage R&R study, and some best practices for successfully doing one.
Overview: What is Gage R&R?
A Gage R&R is a scientific study that falls under the larger topic of Measurement System Analysis (MSA). When you observe any measurement, it will typically vary from time to time. That variation in the observed measurement will be the result of the variation in the object that you are measuring plus the variation of the measuring device.
For example, when you get on the bathroom scale each morning, the scale reading is usually different. That variation will be the variation in your weight plus the variation in the scale mechanism. The Gage R&R helps you distinguish how much of the variation is due to the scale mechanism.
If the proportion of scale mechanism variation is a large part of the overall variation, then you won’t really know your true weight — or whether your new diet is working. The Gage R&R study specifically measures the repeatability and reproducibility of the measurement system.
Repeatability is your ability to measure the same thing multiple times with the same measuring device, under the same circumstances and get the same answer. Reproducibility is the ability of you and your friend Charles to measure the same thing multiple times with the same measuring device, under the same circumstances and both of you getting the same answer.
The Gage R&R study is a carefully planned process by which you select a number of “operators,” or people to do the study, and a number of objects to measure that are known to have inherent variation. You will also choose a measuring device such as a ruler, timer, or scale. Usually you’ll select 2-3 people and 5-10 objects. Each person will measure each object multiple times, usually 2-3 times. You are not interested in the actual values but in the variation that exists when a person measures an object multiple times and the variation that will exist between the different people who are measuring the same object.
The purpose of the same person measuring each object multiple times is to assess the repeatability of the measurement system. Using multiple people to measure the same object multiple times will assess the reproducibility of the measurement system. Calculations of the variance are made of the overall variation of the process plus the variation due to the repeatability and reproducibility. The results are then presented as a ratio of the repeatability and reproducibility and the overall variation. This is expressed as a percentage contribution of the R&R as a function of the overall variation.
Generally recognized criteria for gage acceptability is when Gage R&R variability to overall process variability is:
- Under 10%: Acceptable gage
- 10% to 30%: May be acceptable
- Over 30%: Gage is unacceptable and should be corrected or replaced
3 benefits of Gage R&R
If you want to make good data-driven decisions, you must have faith that your measurement system is providing good data. Here are a few of the benefits of doing a Gage R&R study.
1. Provides a good measure of the process precision
While you want your measurement system to provide both accuracy and precision, the Gage R&R deals strictly with the precision component of measurement variation and provides a simple percentage contribution of measurement variation as a function of overall variation. You want as little contribution from measurement issues versus actual process variation as possible.
2. Extensive computations can be done using statistical software
To fully compute the contribution of R&R, there are a number of detailed statistical computations you must do. Fortunately, there are a number of statistical software applications that can do the calculations for you. But, you must properly design and execute the study, otherwise the results will be meaningless.
3. Efficiency of design
A properly designed Gage R&R study will allow you to understand the variation of your process with a minimal use of resources and in a relatively short period of time. Previously, we discussed using 2-3 people, 5-10 objects and 2-3 repeated measurements. With good planning, this can be accomplished in a reasonable amount of time.
Why is Gage R&R important to understand?
Your process varies whether you like it or not. You must understand how much of that variation is due to the process that you are measuring or due to the variation in the measurement system or device.
Better data-driven decisions
If you are making decisions based on data captured from a flawed measurement system, your decisions may not be the right ones. The Gage R&R will help you make better decisions.
Most problems with measurement systems have a simple solution
Knowing whether your measurement system issue is due to repeatability or reproducibility will point you in the proper direction for correcting and improving your measurement system. Many times, there may be an easy solution to your problem.
Variation is often called Evil
You must always strive to reduce variation in your process. You can’t change what you don’t understand, and you can’t understand what you can’t measure. Having a reliable and trustworthy measurement system will allow you to understand your process better and, therefore, reduce the variation.
Completing a Gage R&R study with a real-world example
While many people often associate Gage R&R and MSA with a manufacturing process, it is just as applicable to non-manufacturing processes. A finance department was getting complaints from its customers that the process of applying credits on invoices was very variable and unpredictable. The CFO approached one of the company’s MBBs, Lora, and asked her to help. Lora knew immediately that she would need to do a Gage R&R study.
Three clerks were chosen to participate in the study. Ten random customer invoices were selected. Lora decided to have each clerk measure the 10 invoices three times to determine the repeatability and reproducibility of the measurement system. The senior finance manager had already evaluated each invoice and determined the correct amount that should have been credited. The value that the study was measuring was how many dollars the clerks credited to each invoice.
At the conclusion of the study, Lora ran all the data through a statistical software program and found the issues. Unfortunately, the percentage contribution of R&R was 25%, which put it in the range of barely acceptable. She broke that down and found that while the repeatability was acceptable, the reproducibility seemed to be the issue. Each clerk was interpreting the correct credits in a different way. Lora also found out there was a bias — the clerks were consistently applying a higher credit than the finance manager was.
Working with the finance manager, Lora developed some better operational definitions and training for the clerks. It took a few weeks for the changes to really impact the invoice process, but then things got more consistent, and the customer complaints stopped.
3 best practices when thinking about Gage R&R
Planning and executing a Gage R&R study is critical to getting meaningful results. Here are a few tips for conducting one.
1. Select people for the study that do the process
Don’t add an additional source of variation by selecting study participants who are not experienced or don’t do the process as a natural part of their job.
2. Remove any bias by having the participants measure the objects in random order
You don’t want people to bias the study by “remembering” previous measurement values. Randomize the objects so that they really have to measure the object rather than being able to rely on memory.
3. Do the study as part of the natural process
Don’t insert an artificial feeling to the study. Run the process as normally as you would. People will be nervous participating as it is. Don’t contaminate the study with any extraneous variation or noise.
Frequently Asked Questions (FAQ) about Gage R&R
What does the R&R stand for in Gage R&R?
Repeatability & Reproducibility
What exactly does a Gage R&R measure?
Gage R&R measures the precision of the process. It does not measure accuracy, bias, discrimination, or other characteristics of a measurement system.
What type of data is required for a Gage R&R?
A Gage R&R study is designed for continuous data. There are similar measurement system studies for analyzing the variation of the process using attribute data.
Wrapping up Gage R&R
The goal of a Gage R&R study is simple. It is designed to allow you to identify what proportion of the variation for your observed data is caused by the actual variation of what you are measuring and the variation due to the measuring device. Specifically, you are looking at the repeatability and reproducibility of the system.
If your measurement system has a problem with being repeatable or reproducible, you need to take action to mitigate or eliminate the reasons so you can trust your data. To paraphrase Dr. W. Edwards Deming, “Without (trustworthy) data, you’re just another person with an opinion.”