Repeatability and Reproducibility Techniques for

Measurement System Analysis

Anshika Raj Yadav (230102111)
Department of Electronics and Electrical Engineering
IIT Guwaha
[Link]@[Link]

Abstract—This report aimed to analyze two important sta c processes observed is due to the variability of the
characteris cs of instruments: Repeatability and Reproducibility. We
will examine the factors that have a considerable effect on the
measurement system.
measured values. To dive into this topic we will use the Gage R&R Gauge R&R can be calculated by using 3 methods:
study. Gage R&R is a process to check that tools, equipment, or
operators are non-conforming. Results of Gage R&R can define the 1. Range method
causes of the measurement errors. A er analyzing the results, the
measurement system will be more precise and more accurate. 2. Average and range method(X-bar and R)
Index Terms—Repeatability, Reproducibility, Gage R&R,
3. ANOVA method
Measurement System Analysis, ANOVA, Sta s cal Process Control
In this report we will only see about ANOVA method.

I. INTRODUCTION
The measurement system is an important factor in III. ANOVA (The Analysis of Variance)
determining the performance of your system. It makes us The ANOVA consists of a series of techniques origina ng
check if the system we are using is capable of making accurate by the inferen al sta s cs theory, that are interested in the
and reliable measurements for decision-making. The quality of comparison of data variability. These methods are used when
a system is inversely propor onal to variability. SPC or there are two or more popula ons to es mate the differences
Sta s cal Process Control is a process of analysis of variability between their sample means; analyzing the respec ve
and it uses some techniques and methods such as Gage RR. variances to achieve its purpose. By the evalua on of two or
The Gage R&R is a study of the variability observed in more different distribu ons, ANOVA allows us to determine if
measurement due to the measurement system itself. R&R
the differences are random or not.
denotes Repeatability and Reproducibility are two crucial
characteris cs of each measurement system. ANOVA is based on a technique that compares the sample
means to es mate the data varia on. It obtains this aim
Repeatability refers to the varia on caused by the decomposing the variability in between and within.
instrumenta on or the varia on observed when the same
ANOVA is a process of sta s cal inference, in par cular it
operator measures the same part more mes with the same
is a technique of parametric sta s cal inference based on an
instrumenta on in the same condi on.
hypothesis test.
Reproducibility is the varia on caused by the measurement In conclusion, the answer to the hypothesis tes ng will be
system or the varia on observed when different operators based on the p-value: observing the p-value rela ve to the
measure the same part with the same instrumenta on in value of F, obtained from sta s cal tables or so ware, and
different condi ons. comparing it with the level of significance α, we can decide
II. GAGE R&R whether to accept or reject the null hypothesis. In par cular,
we accept the null hypothesis if the p-value is greater than the
Gage R&R is a process mainly used to evaluate whether a significance level, while we accept the alterna ve hypothesis
measuring instrument is accurate or not by ensuring its if the p-value is less than α. Refusing, or rather not accept the
measurements are repeatable and reproducible. It involves null hypothesis means that we can conclude that among the
breaking the total measuring variability into two por ons: observed data there may be significant differences. Intui vely
repeatability and reproducibility. The Gage R&R is a study on it can be stated that, if the null hypothesis is true, so the data
the ability of a measuring system. The Gage R&R techniques differ for the effect of random factors, and changing the levels
are used to determine how much of the variability of the of the factor does not change the response, but if the null
hypothesis is false means that the variability total
phenomenon can be a ributed to systema c factors. So the
method allows to value if the difference between the variances
of two distribu ons, namely the total variability, is due to
chance or is significant. Analyzing the table Gage R & R we can pay a en on on the
columns for % Contribu on and %StudyVar. The first indicates
ANOVA in Gage R&R par ons the total measurement which are the percentages of contribu on of each source of
varia on into three sources: varia on of the process, while the second is the standard
1. Repeatability (Equipment Varia on - EV) devia on expressed as a percentage and rela ve to the various
2. Reproducibility (Operator Varia on - AV) sources of process varia on.

3. Part-to-Part Varia on (PV)  for %Contribu on:

The value of Repeatability is 0.02%, so <1% and this

GAGE R&R TABLE means that the measurement system is acceptable , in fact

Metrix of Unacceptable Marginal Acceptable measuring rise and fall mes by the same generator there
Success were no significant differences.
%Contribu on >9% Between <1% The value of reproducibility is instead very high, i.e. 99.98% ,
1% and 9%
that is >9% , so the measurement system is not acceptable.
%study Var >30% Between <10%
10% and This is because measuring rise and fall mes by different
30% generators there are significant differences.
%Tolerance >30% Between <10%
 for %StudyVar we can observe the same results
10% and
(thought the threshold values change):
30%
the value of Repeatability is 1.44%, so <10% and
No. of dis nct <5 Between 5 >10
this means that the measurement system is
categories and 10
acceptable, in fact measuring rise and fall mes by the
same generator there were no significant differences.
The value of reproducibility, as in the previous case, is
IV. AN EXAMPLE instead very high i.e. 99.99%, that is>30%, so the
measurement system is not acceptable.
This graph reveals that the varia on of the
The test phase configura ons use the oscilloscope to measure measurement system is the largest source of
10 values of both rise and fall me by se ng a sampling variability observed in the experiment, with different
frequency. We have used two different waveforms generators operators (reproducibility) as major contributors to
to generate the signal . the varia on.

Components of Variation
V. CONCLUSION
150

100 The Analysis of Variance (ANOVA) and Gage R&R techniques

are essen ally the only ways to iden fy any problem in the
50 measurement system. In a Gage R&R Study, measurements of
one part measured by one operator are analyzed to es mate
0
Gage R&R Repeat Repro Part-to-part the level of varia on, the gage repeatability, and the accuracy
of the measurements. By analyzing the measurements of one
%contribution %studyVar reference part by one operator, it is always possible to
determine whether a measuring device is capable of
measuring a par cular characteris c under condi ons with
rela vely small varia on. Any me we want to measure the
results of a process we will see some varia on. This varia on
comes from two sources:
1) There are always differences between parts made by any
process.
2) Any method of taking measurements is imperfect - thus,
measuring the same part repeatedly does not result in
iden cal measurements.
Even though in a Gage R&R study it is important that the
measurements are done in a random order and that the parts
selected provide a representa ve sample across the possible
range of responses, the ANOVA methods, that form the basis
of Gage R&R studies, determine what por on of the variability
in measurements may be due to the measurement system.

VI. REFERENCES

 Research on Electric and Electronic Measurement for

the Economic Upturn Benevento, Italy, September
15-17, 2014
 Burke, S., & Kissell, W. (2009). Measurement System
Analysis: The Need for a Systema c Approach.
Journal of Quality Technology, 41(1), 1-12.
 Montgomery, D. C. (2020). Introduc on to Sta s cal
Quality Control (8th ed.). John Wiley & Sons
 Case Study: Measurement System Analysis for
Quality Improvement Using Gage R&R Study at
Company XYZ