Locked lesson.

Upgrade

About this lesson

Gage R&R studies are the best practice for determining measurement system process capability and maintaining that capability. They quantify the level of precision that the measurement system will deliver. Key principles for conducting effective Gage R&R studies are introduced in this lesson including those for managing the appraisers and for selecting the items for measurement.

Exercise files

Download this lesson’s related exercise files.

Gage R&R Principles.docx
61.2 KB Gage R&R Principles - Solution.docx
61.4 KB

Quick reference

Gage R&R Principles

Gage R&R Studies are used to determine if the common cause variation that effects measurement system precision is acceptable. This lesson introduces Gage R&R Studies and explains the principles used in conducting them.

When to use

A Gage R&R Study should be done on a measurement system when it is first put into service to qualify it. Also, a study should be done after a major change to the system or prior to use in a new or unusual manner. Studies are also done periodically for those systems that are deemed to be vulnerable to high variation.

Instructions

Gage Repeatability and Reproducibility (R&R) Studies are designed to measure the common cause variation of a measurement system. It quantifies the level of measurement error that affects the precision of the system. Gage R&R Studies are often done as part of the qualification of a new measurement system.  They are also done after a major change to the system or the system’s procedures. If the system is to be used in a new or unique manner, such as to conduct special tests for a Lean Six Sigma project, a Gage R&R Study is often done to ensure the data will be meaningful.

 

Some organizations periodically do a Gage R&R study if the measurement system is susceptible to many factors that can cause variation. This is in order to monitor the performance of the system. A Gage R&R Study includes equipment, associates who normally use the equipment, items for inspection, and other factors that will usually be elements of the measurement process. The key is to conduct the measurements in a manner that is as normal as possible to minimize the introduction of special causes.  Several typical characteristics of a study are:

  • Two or three appraisers who are individuals who normally use the measurement system for the application being studied.
  • Multiple items for inspection. Use five to ten items for variable measurements and twenty to thirty for attribute measurements.
  • Items for measurement should be the full range of typical values and include some “bad” items that are outside the normal range of acceptability.
  • Each item is measured by each appraiser multiple times – normally two or three trials.
  • Items are measured in a random order.
  • Items are indistinguishable from each other by the appraisers.
  • Appraisers use the same measurement system equipment.
  • Measurements are done in the actual work location if possible. If not, the setting is as similar as possible.
  • Data is carefully collected so that each data reading is correlated to a specific appraiser, a specific item and a specific trial – although the appraiser cannot know which item is which.

The results of the data are statistically analyzed and repeatability issues, which is variation that an operator has with the system, and reproducibility issues, which is variation between operators, is quantified.

Hints & tips

  • The person conducting the study is not normally one of the appraisers. That way they can manage the data collection form and ensure data is recorded correctly.
  • Use Minitab to create the data collection form. It will randomize the tests and is an easy way to store the data. Plus, it is ready for analysis as soon as the data collection is done.
  • The biggest danger when doing these is that the appraisers know it is a special study, so they change their normal measurement habits or techniques. This introduces a special cause and can invalidate the study findings. That is why we try to do these in the normal workplace. If it is a familiar environment they are less likely to change behavior.
  • The study must be balanced – all appraisers measure all items the same number of times – for the statistics to work correctly.
Login to download
  • 00:05 Hi, I'm Ray Sheen,
  • 00:06 we introduced the idea of a Gage R&R study when we talked about precision.
  • 00:11 The next few lessons will show you how to do a study and
  • 00:15 I'll start in this lesson by laying out some foundational principles.
  • 00:20 I'll start by describing what a Gage R&R Study is, Gage R&R stands for
  • 00:25 repeatability and reproducibility.
  • 00:28 They're statistically-based analyses that quantify the level of
  • 00:32 precision measurement error.
  • 00:34 That means the uncertainty or a normal distribution of the measurement system.
  • 00:38 A Gage R&R study does not address accuracy,
  • 00:41 at least a variable Gage R&R study does not address accuracy.
  • 00:44 An attribute study can address accuracy in addition to precision but
  • 00:50 we'll focus on precision.
  • 00:51 That's why we do not need calibration standards for variable Gage R&R study.
  • 00:56 The Gage R&R studies should be done periodically,
  • 00:59 It definitely should be done when the system is first put in place.
  • 01:03 To make sure the system is capable of rejecting bad items and
  • 01:06 accepting good items.
  • 01:07 This is normally done as part of the measurement system validation.
  • 01:10 But it's also helpful periodically to do a study to make sure that
  • 01:13 the system did not drift.
  • 01:15 In particular, I would advise doing one before relying on a measurement system
  • 01:19 to conduct an improvement project like a Lean Six Sigma project.
  • 01:23 You want to be certain you can trust the data that you'll be using for
  • 01:26 identifying and resolving root causes.
  • 01:29 And just like you would do a Gage R&R study before putting a new measurement
  • 01:32 system in place,
  • 01:33 you should do one whenever a major change is made to the system.
  • 01:37 You should be certain that the new measurement system is capable of
  • 01:39 registering precise measurements.
  • 01:42 So, let's look at what goes into a Gage R&R Study,
  • 01:44 the study includes equipment, associates which we will call appraisers from now on.
  • 01:50 Items for inspection,
  • 01:51 and other factors that will usually be elements of the measurement system.
  • 01:55 The Gage R&R study reveals the impact of common cause variation in the measurement
  • 02:00 process, so everything in the process needs to be there.
  • 02:03 Multiple items are measured by the appraiser in a random sequence and
  • 02:08 normally they're measured several times using the same equipment.
  • 02:11 These results are then statistically analyzed to determine the precision error.
  • 02:15 In order to avoid or at least minimize the introduction of special cause effects,
  • 02:20 the measurements done as realistically as possible.
  • 02:23 For instance,
  • 02:24 the actual operators who normally use the measurement equipment are the appraisers.
  • 02:28 The analysis should be done in the normal work environment so
  • 02:32 that the environmental factors are not creating special cause variance.
  • 02:35 This special cause variance is just as likely to be good as bad if measurements
  • 02:40 for the study are done in a clear, well-lighted laboratory environment
  • 02:45 instead of the normal day-to-day work environment.
  • 02:48 Now if the measurement system is a special one, created say for
  • 02:51 a Lean Six Sigma project,
  • 02:53 then the analysis is done wherever the Lean Six Sigma project
  • 02:55 measurements will be made.
  • 02:58 Once all the data has been collected, it is statistically analyzed,
  • 03:02 we'll go through the analysis in other lessons.
  • 03:05 But based upon this analysis we can check for repeatability
  • 03:09 which focuses on the performance of each appraiser individually.
  • 03:12 And will highlight any equipment variation, and we can check for
  • 03:16 reproduce ability.
  • 03:18 Which focuses on the performance between appraisers or across the appraisers.
  • 03:22 And gives us insight into the differences caused
  • 03:25 by different measurement system operators.
  • 03:29 Let's wrap up this by highlighting some of the best practises in the design
  • 03:33 of a Gage R&R study.
  • 03:35 First, use multiple appraisers in order to identify operator variation,
  • 03:40 however for simplicity, we normally keep it to two or three operators.
  • 03:44 Again, these should be the people who normally use the measurement system.
  • 03:49 Second, you want to be measuring multiple items, not just one, for
  • 03:53 variable data measurements, you should have five to ten items.
  • 03:57 For attribute, by that I mean pass, fail,
  • 04:00 or yes/no data, you should have at least 20 items, and 30 is definitely better.
  • 04:05 Among the items being measured, there should be some that are good and
  • 04:08 some that are bad.
  • 04:10 You want the normal range of the item being represented
  • 04:13 in what's being measured.
  • 04:14 And for attribute data analysis, a 50/50 split is ideal although you will get
  • 04:18 very good results even with a 70/30 split.
  • 04:22 And you don't want the appraiser to be able to visually identify which item is
  • 04:26 which, so don't put numbers or letters on them.
  • 04:30 And definitely don't tag the bad ones with a red tag,
  • 04:34 I saw that happening one time and in a Gage R&R study.
  • 04:37 Another best practise is to balance the study, balancing means that every
  • 04:42 appraiser measures every item the same number of times, that's necessary for
  • 04:47 the statistical analysis.
  • 04:48 And have the items measured in a random sequence each time an appraiser measures
  • 04:53 them so that they can't rely on memory to mark the first one as good or bad.
  • 04:59 Now this can add up to a lot of measurements,
  • 05:01 if you have three appraisers measuring ten items through three runs or trials.
  • 05:06 You're talking about 90 measurement values that need to be collected for analysis.
  • 05:11 The final best practice I wanna highlight
  • 05:14 is the preparation of the data collection form.
  • 05:17 With this form, you can keep track of which measurements go with which items and
  • 05:21 which appraisers, this will be very important for the analysis.
  • 05:25 The randomization of the measurements means that it's very easy in the analysis
  • 05:30 to get these confused or make mistakes.
  • 05:33 If you use Minitab in your studies, it will automatically create the form,
  • 05:37 if you do it manually, you must carefully keep track of the tests and the data.
  • 05:41 So those are the principles of a Gage R&R study, in the next few lessons, we'll get
  • 05:47 into the details of managing a study and finally, we'll walk through some examples.

Lesson notes are only available for subscribers.

PMI, PMP, CAPM and PMBOK are registered marks of the Project Management Institute, Inc.

Gift this course