Measurement System Analysis: Instruction Manual

Measurement System Analysis
Instruction Manual
Febraury 2009
Thank you for selecting our software package. This program is owned by BPI Consulting, LLC. This program cannot be
copied or used unless under license with BPI Consulting, LLC. BPI Consulting, LLC is not liable for any decisions made
based on the use of this software package.
Requirements: This program is a Microsoft Excel add-in. You must Microsoft Excel for this program to work. This
program supports any version of Excel from 2000 on.
BPI Consulting, LLC
20314 Lakeland Falls
Cypress, TX 77433
281-304-9504
www.spcforexcel.com
BPI Consulting, LLC, 2009
Table of Contents
SPC for Excel Help Files
Measurement Systems Analysis
Crossed Gage R_R: Average/Range Method
Crossed Gage R_R: ANOVA Method
Nested Gage R_R
Range Gage R_R Method
Bias - Independent Sample Method
Bias - Control Chart Method
Linearity
Attribute Gage R_R
p2
3
3
4
9
11
15
17
20
24
26
SPC for Excel Help Files

Measurement Systems Analysis
The Measurement Systems Analysis techniques used by the program are primarily based on the following two
sources:
1. Measurement Systems Analysis, Third Edition, AIAG, May 2003 (www.aiag.org)
2. Evaluating the Measurement System by Donald Wheeler and Richard Lyday, SPC Press, Knoxville, TN, 1989 (
www.spcpress.com)
Both are excellent references for developing a better understanding of the measurement system. The program has
the following components:
Crossed Gage R&R: Average and Range Method

Gage R&R Report (AIAG format)
Variation by Components Summary (variance, standard deviation, tolerance and process standard deviation)
Variance components chart
Averages charts stacked and unstacked
Range charts stacked and unstacked
Run chart by part
Scatter plots
Whiskers charts
Error charts
Normalized histograms
X-Y Plots
Range charts for each operator
Operator bias chart
Operator consistency chart
Crossed Gage R&R: ANOVA Method
Gage R&R ANOVA Report (AIAG format)
ANOVA table and residuals plot (AIAG format)
Variation by Components Summary (variance, standard deviation, tolerance and process standard deviation)
Nested Gage R&R
ANOVA Table
Variation by Components (variance, standard deviation, tolerance and process standard deviation)
Operator by batch
Operator averages chart
Operator range chart
Range Method for Gage R&R
Bias Method Independent Sample Method
Bias Method Control Chart Method (checks Stability also)
p3
Linearity Method
Attribute Gage R&R
Effectiveness table (attribute Gage R&R report, AIAG format)
Crosstabulations
Kappa values
Crossed Gage R_R: Average/Range Method

The example below uses the data from page 101 of Measurement System Analysis, Third Edition. The data is
available in the workbook Gage R&R Example Workbook.xls (default installation folder: My Documents/SPC for MS
Excel). In this example, three operators are testing 10 different parts three times. For more information on the
average/range method, please click here to see our newsletter on this topic.
The first step in using the Gage R&R program is to setup the data entry page.
1. Select the MSA icon (
) on the SPC toolbar.
2. Select the Setup for the classical (crossed) Gage R&R option and OK.
3. Fill in the information in the form below and select OK. You must enter all the information. The number of
decimal places in the measurement is used for rounding in the titles of the control charts. The number of trials and
parts can range from 2 to 20; the number of operators can range from 1 to 25 (with one operator you will only get a
Gage R&R report based on the average and range method).
4. Fill in the blank template that was generated on a new worksheet. A completed template is shown below. Enter
the operator names in the upper right hand corner. The names will automatically appear in the first column. NOTE:
The program uses Microsoft Excels naming function to run. You cannot have spaces or certain characters (e.g. /).
Instead of using John Smith, use John_Smith.
Enter the rest of the information for Date, Gage Name, Gage Number, Gage Type, Product, Characteristic, Process
Standard Deviation, Upper Specification Limit, Lower Specification Limit, and Performed By. None of this information
is required to run the program with the possible exception of the specification limits and process standard deviation.
p4
These are required if you are basing the acceptability of the measurement system on one of these two. You then
enter the data from the operators for each trial and each part. A completed template is shown below.
6. Select the Run the analysis (have entered data) option under Average/Range Method.
7. Select an option under Base Acceptability on:
i. Tolerances: Use this option if your parts have very little variation or not representative of the total variation
in your production process. To use this option, you must enter one or both of the specifications in the
template as shown above.
ii. Total Variation Based on Parts: Use this option if your parts are representative of the total variation in your
process. This is the default option.
iii. Process Standard Deviation: Use this if your parts are not representative of the total variation in your
process and if you have a good estimate of the process standard deviation (e.g., from a control chart kept on
the process or a historical standard deviation). To use this option, you must enter the process standard
deviation in the template above.
8. Select an option under Charts to Include:
i. All Charts: this will generate all the charts associated with the study; this is the default setting
ii. No Charts: only the Gage R&R report will be generated
iii. Select Charts: only the charts you select will be generated; if you select this option, you will see the form
below; select the charts to include and then OK.
p5
9. Select the sigma multiplier. This is the value by which the standard deviation is multiplied. The default value is 6.
Six standard deviations capture 99.73% of the process data. The AIAG reference used 5.15 for some time which
captures 99% of the process data.
10. Select OK to run the program and generate the results.
Crossed Gage R&R Average/Range Output

The following worksheets are generated regardless of the option selected for additional charts:
Gage R&R Report (named Report followed the name of your study) - AIAG format is followed.
R&R Summary breaks down the variation into components: variance, standard deviation, tolerance (if one or
both the specifications were entered in the template) and process standard deviation (if entered in the
template).
Variance Component Chart plots the percentages in the R&R summary to allow comparisons.
The output for the AIAG Gage R&R report is shown below using the acceptance criteria based on the total variation
of the parts. The top part of the worksheet gives the gage information. The middle part of the worksheet shows the
calculations and results for the repeatability, reproducibility, Gage R&R, part variation and total variation as well as
the percentages based on total variation in the parts and tolerances (if the specifications were entered into the
template). The bottom part of the worksheet contains the conclusions based on the option selected (tolerances,
part variation or historical standard deviation). The sigma multiplier is used to determine the % tolerance values.
p6
Note: if you select the option of basing the results on the process standard deviation, that standard deviation is used
in place of the calculation for TV in the table above.
The output from the R&R summary sheet is given below. The top part of the worksheet contains the results for the
variances and the % contribution due to the variances. The total in this column will add to 100. The bottom part of
the worksheet contains the results for the standard deviations and % contributions to the sigma multiplier times the
standard deviation as well as % tolerance and % process (if entered into the template).
The other worksheet added is a plot of the components of variation. The chart for this data is shown below. It
contains the percentages from the R&R worksheet for % contributions to variance, standard deviation, tolerance,
and process for repeatability, reproducibility, total gage R&R and part to part variations.
p7
Any charts that were selected are generated on separate worksheets in the workbook. A summary of each chart is
given below.
Stacked Averages Chart the average of each operator on each part is plotted by operator using the part
number as the index. There is one line for each operator. This helps determine how consistent the operators
are. The overall average and control limits are also plotted. If the parts represent the total (true) variation in
the process, at least half of theses points should be out of control. If this not the case, the measurement system
does not have the ability to distinguish between samples (poor resolution) or the parts do not reflect the total
variation in the process.
Unstacked Averages Chart same as the stacked chart but the operators are plotted together, not separately.
Stacked Range Chart used to show the range of each operators trials on a part and includes the average range
and control limits. There is one line for each operator. The chart is used to determine if the process is in control.
If there are out of control points, the special causes need to be found and eliminated. Care should be taken with
interpreting the Gage R&R results if there are special causes present. Special causes occur if there are points
beyond the control limits.
Unstacked Range Chart same as the stacked range chart but the operators are plotted as one line.
Run Chart by Part plots the individual readings by part for all operators to help see if there are any outliers and
to see the variation in the individual parts.
Scatter Plot plots the individual readings by part-by-operator to examine how consistent the operators are, to
look for part-operator interactions, and to look for outliers.
Whiskers Charts plots the high, average, and low value by part for each operator to examine how consistent
the operators are, to look for part-operator interactions, and to look for outliers (same items as for the scatter
plot).
Error Charts plots the error (observed value average measurement of the part) by part-operator to
determine which operator may have bias and which operator has the most variability.
Normalized Histograms plots the normalized value (observed value average measurement of part) as a
histogram to determine how the error is distributed by operator.
X-Y Plot plots the average of the readings by each operator against the overall part averages to examine
consistency in linearity between operators.
Appraiser Charts consists of three charts:
* Range charts for each operator to determine if each is in control
* Bias chart for all operators to determine if different operators display detectably different average
values for the parts.
* Consistency chart for all operators to determine if different operators display detectably different
standard deviations for the parts.
p8
Crossed Gage R_R: ANOVA Method

The set-up for the ANOVA method is the same as for the Average and Range Method. Please follow the instructions
for steps 1 through 4 for the Average and Range Method. Then follow the steps below.
2. Select the Run the analysis (have entered data) option under the ANOVA Method.

Crossed Gage R&R ANOVA Method Output
The following worksheets are generated:
ANOVA Report (named ANOVA Report followed by the name of your study) - AIAG format is followed.
GRR ANOVA contains the ANOVA table and residuals plots.
ANOVA R&R Summary breaks down the variation into components: variance, standard deviation, tolerance (if
specs were entered in the template) and process standard deviation (if entered in the template).
ANOVA Variance Component Chart plots the percentages in the ANOVA R&R summary to allow comparisons.
The ANOVA Report output is shown below. It contains the Gage Repeatability and Reproducibility ANOVA Method
Report as outlined by AIAG. The top part of the report contains the gage information. The middle part of the
worksheet contains the results of the analysis. The acceptability of the measurement system is based on %
contribution, not on % total variation. The method is acceptable if the % contribution is less than 10%.
p9
The bottom part of the worksheet contains the number of distinct categories as well as the total variation and %
tolerance information if the specifications were entered on the template.
The GRR ANOVA worksheet contains the ANOVA table, a breakdown by source of variation and the residuals plot
as shown below. The residuals chart plots the residual versus the average for each operator for each part. The
residual is the result minus that average. The points should be randomly scattered above and below zero.
p10
The other two worksheets are titled ANOVA R&R Summary and ANOVA Var Comp. These are the same as those
generated for the Average/Range method. Please see the information in the Average/Range method section of this
manual for more information.
Nested Gage R_R

A nested Gage R&R is used when you can't have each operator test the same part. It may be that the part is
destroyed during the test (destructive testing).
The data for this example is in the workbook Gage R&R Example Workbook.xls (default installation folder: My
Documents/SPC for MS Excel). This example involves testing the hardness of steel tubes. The hardness test involves
a small piece of the steel. However, the same piece cant be retested by another operator. The test will be done by
taking two samples from each tube. One operator will run those two samples. Each operator will do this for 5
tubes.
2. Select the Setup for nested Gage R&R option and OK.
p11
3. Fill in the information in the form above and select OK. You must enter all the information. The number of
decimal places in the measurement is used to round the titles in the control charts.
4. Fill in the blank template that was generated on a new worksheet. A partial completed template is shown below.
Enter the operator names in the upper right hand corner. The names will automatically appear in the first column.
NOTE: The program uses Microsoft Excels naming function to run. You cannot have spaces or certain characters
(e.g. /). Instead of using John Smith, use John_Smith.
Enter the rest of the information for Date, Gage Name, Gage Number, Gage Type, Product, Characteristic, Process
Standard Deviation, Upper Specification Limit, Lower Specification Limit, and Performed By. None of this information
is required to run the program with the possible exception of the specification limits and process standard deviation.
These are required if you are basing the acceptability of the measurement system on one of these two. You then
enter the data from the operators for each trial and each part. A completed data entry screen is shown below.
6. Select the Run the analysis (have entered data) under the Nested Gage R&R Method and then select OK.
p12
Nested Gage R&R Output

The following worksheets are generated:
Nested GRR contains the ANOVA table and variation by components (variance, standard deviation, tolerance
and process)
Variance components chart plots the percentages in the Nested GRR worksheet
Operator by batch chart plots the results for each operator by batch
Operator averages chart determines if any results are out of control for the average results for each batch; the
more out of control points the better the measurement system
Operator range chart determines if each operators variation is the same for all operators (range chart will be
in control)
The Nested GRR worksheet is shown below. The top part of the worksheet contains the gage information. The
next part of the worksheet contains the ANOVA table. The ANOVA table has four sources of variation: operator,
batch, repeatability and total. With a nested design, you are looking to see if either operator or batch is significant.
This is determined by the p value. A p value of less than 0.05 means that there is significant. In this example, the p
value for batch is less than 0.05, which means it is significant. This is where you want to have the largest variation
(sum of squares). This means that the measurement system can tell the difference between batches within each
operator.
The bottom portion of the worksheet breaks out the components of variation. This includes the results for the
variances and the % contribution due to the variance. The total in this column will add to 100. Next are the results
p13
for the standard deviations and % contributions to the sigma multiplier times the standard deviation as well as %
tolerance and % process (if entered into the template).
The Variance Components sheet is shown below. It is a graph of the % contributions shown in the Nested GRR
worksheet.
The Operator by Batch chart for this data is shown below. This is a plot each operators results. The straight line
connects the average of the samples for one batch per operator.
The Operator Xbar chart is for this data is shown below. If the measurement system can tell the difference between
samples, most of the points should be out of control. This is the case for this example.
p14
The Operator Range chart for this data is shown below. This chart determines if each operator is repeatable
compared to each other. If the range chart is in control, it means that the within batch variation is the same for all
operators.
Range Gage R_R Method

This method provides a quick look at measurement variability and requires that you have an estimate of the process
standard deviation. This method typically uses two operators and five parts. Each operator measures the part one
time only. An example of the data input required is shown in the figure below and is based on the data from page 98
of Measurement Systems Analysis, Third Edition. The data is available in the workbook Gage R&R Example
Workbook.xls (default installation folder: My Documents/SPC for MS Excel).
1. Enter the data into a worksheet as shown below. The data can be anywhere in the worksheet.
p15
Part
Operator A Operator B
0.85
0.80
0.75
0.70
1.00
0.95
0.45
0.55
0.50
0.60
2. Select the data highlighted above.

) from the SPC toolbar.
4. Select the Range Gage R&R Method option from the second tab "Range/Bias/Linearity/Attribute Gage R&R" and
then select OK.
5. Enter the information into the Range method form. The default value for the "Range containing part numbers and
operator results" is the selected range on the worksheet. You may change it if that is not correct. The process
standard deviation must be entered. In this example, the standard deviation is 0.0777.
Range Gage R&R Method Output

A new worksheet is generated in the workbook. The top part of the worksheet contains the numerical results as
shown below. The conclusion is also given based on the following:
If % GRR is less than 10%, the measurement system is acceptable
If % GRR is between 10% and 30%, the measurement system may be acceptable based on the application
If % GRR is greater than 30%, the measurement system needs improvement
p16
The worksheet also contains the range control chart. Any out of control points will be in red. If there are any out of
control points, the results may not be valid.
Bias - Independent Sample Method

This method determines if the measurement system is biased. It is done by using one sample and determining its
reference value. Then one operator measures the sample ten or more times. The data used in this example is from
page 87 in Measurement Systems Analysis, Third Edition. The data is also available in the workbook Gage R&R
Example Workbook.xls (default installation folder: My Documents/SPC for MS Excel). For more details on
determining the bias of measurement systems, please see our newsletter on this topic (click here).
Trial
Value
5.8
5.7
5.9
5.9
6.0
6.1
6.0
6.1
p17
6.4
10
6.3
11
6.0
12
6.1
13
6.2
14
5.6
15
6.0

4. Select the Bias (Independent Sample Method) option from the second tab "Range/Bias/Linearity/Attribute Gage
R&R" and then select OK.
5. Fill in the first tab of the Bias: Independent Sample Method input form. The range that appears in the dialog box is
the range selected on the worksheet. You can change it here if it is not correct. Enter the reference value (which is
6.00 in this example); enter the number of decimal places (1 in this example), and alpha. The default value is 0.05
which gives 95% confidence limits.
6. Fill in the second tab of the Bias: Independent Sample Method input form. This tab contains the output options.
There are two options for plotting the histogram by midpoint or by class width. By midpoint is the default value.
The histogram can be plotted on the same sheet as the analysis (recommended) or as a new sheet in the workbook.
There are two output options for the analysis of bias study. You can place it in the worksheet where the data is or as
a new sheet (recommended). You do not have to go to this tab if you are satisfied with the default options.
p18
7, Select OK to run the program and generate the results.
Bias Independent Sample Method Output
A new worksheet is added with the results or the results are placed in an existing worksheet depending on which
option you selected. The first part of the output contains the numerical calculations. The numerical values include:
n = number of readings
Mean = average of the readings
Reference Value = reference value of the sample
Standard Deviation (s) = standard deviation of the readings
t statistic = the t value based on the degrees of freedom
df = degrees of freedom
t value (2 tailed) = t value from the t tables
Bias = average reference value
Lower = lower confidence interval (based on alpha)
Upper = upper confidence interval (based on alpha)
A conclusion is also presented that states if you can assume the bias is zero.
A histogram of the readings is also given. In this example, you can see that the histogram does contain values on
both sides of the reference value.
p19
Bias - Control Chart Method

This method uses the control charts to determine if the measurement system is biased. It can also be used to check
the stability of the measurement system. For more details on checking the stability of the measurement, click here
to view our newsletter on this topic. You must have a reference value for the sample. The control chart can be an
individuals chart (X-MR), Xbar-R chart, or Xbar-s chart. The example below uses the X-MR chart. The data is available
in the workbook Gage R&R Example Workbook.xls (default installation folder: My Documents/SPC for MS Excel).
Sample
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Results
100.1
101.9
101.0
100.2
100.6
101.5
100.6
100.2
101.1
101.6
100.8
101.4
101.5
100.8
102.0
100.9
100.5
101.3
101.6
101.3


p20
4. Select the Bias (Control Chart Method) option from the second tab "Range/Bias/Linearity/Attribute Gage R&R" and
then select OK.
5. Fill in top part of the Bias: Control Chart Method input form. The default for the "range containing subgroup
identifiers" is the data in the first column of the selected range on the worksheet. The default for the "Range
containing measurements" is the data to the right of the subgroup identifiers. Both these ranges can be edited in the
input form.
6. Fill in the information on the "Data Input" tab.

Select the type of chart. The program selects what it thinks you have based on the selected area on the
worksheet. You can change it if it is not correct.
Enter the reference value for the sample (in this example, 100.3)
The number of decimal places in the data (in this example, 1)
Enter alpha (default is 0.05)
7. Fill in the information on second tab of the Bias: Control Chart Method input form ("Output Options"). There are
two options for plotting the histogram by midpoint or by class width. By midpoint is the default value. The
histogram can be plotted on the same sheet as the analysis (recommended) or as a new sheet in the workbook.
There are two output options for the analysis of bias study. You can place it in the worksheet where the data is or as
a new sheet (recommended). You do not have to go to this tab if you are satisfied with the default options.
p21
8. Fill in the information on third tab of the Bias: Independent Sample Method input form (Control Chart
Titles/Labels). The default titles depend on the type of control chart selected on the first tab. You do not have to go
to this tab if you are satisfied with the default entries.
Bias Control Chart Method Output

A new worksheet is added with the results or the results are placed in an existing worksheet depending on which
option you selected. The first part of the output contains the numerical calculations. The numerical values include:
n = number of readings
Mean = average of the readings
Reference Value = reference value of the sample
Standard Deviation (s) = standard deviation of the readings
p22
t statistic = the t value based on the degrees of freedom

df = degrees of freedom
t value (2 tailed) = t value from the t tables
Bias = average reference value
Lower = lower confidence interval (based on alpha)
Upper = upper confidence interval (based on alpha)
A conclusion is also presented that states if you can assume the bias is zero. In this example, you cannot assume that
the bias is zero.
The histogram of the measurement readings are shown below. As can be seen, there are just a few readings below
the reference value of 100.3. Most are above.
The control charts are shown below for this data. You want the control charts to demonstrate statistical control
(stability). Any out of control points will be in red. In this example, the measurement system is stable.
p23
Linearity
Linearity is the difference of bias throughout the measurement range. To determine linearity, the samples you select
must cover the expected operating range of the measurement system. You should use at least five samples that
cover this range. One operator should measure each of the parts at least ten times. An example of the required data
input is shown below (page 95, Measurement Systems Analysis, Third Edition). The data is also available in the
workbook Gage R&R Example Workbook.xls (default installation folder: My Documents/SPC for MS Excel). For more
details on linearity, please click here to our newsletter on linearity.
1. Enter the data into a worksheet as shown below. The data can be anywhere in the worksheet. The data must be
in columns. The first row of data contains the part number; the second row contains the reference values for each
part; and the remaining rows contain the measurements.
Part
Ref Value
1
2
1
2
2.7
2.5
2
4
5.1
3.9
3
6
5.8
5.7
p24
4
8
7.6
7.7
5
10
9.1
9.3
3
4
5
6
7
8
9
10
11
12
2.4
2.5
2.7
2.3
2.5
2.5
2.4
2.4
2.6
2.4
4.2
5
3.8
3.9
3.9
3.9
3.9
4
4.1
3.8
5.9
5.9
6
6.1
6
6.1
6.4
6.3
6
6.1
7.8
7.7
7.8
7.8
7.8
7.7
7.8
7.5
7.6
7.7
9.5
9.3
9.4
9.5
9.5
9.5
9.6
9.2
9.3
9.4

4. Select the Bias (Control Chart Method) option from the second tab "Range/Bias/Linearity/Attribute Gage R&R" and
then select OK.
5. Fill in the information in the Linearity input form. The range is the range selected on the worksheet. Enter the
number of decimal places and alpha (default is 0.05). The Rounding in Equations determines how the best fit
equation is displayed. The default value is 3.

Linearity Output
A new chart sheet is added to the workbook. This chart contains all the results. The chart for this example is shown
below.
p25
You will also see two text boxes in the upper left hand part of the chart. These will tell you if the bias is the same and
if the linearity is acceptable. The text boxes for this data (not shown on the chart) are given below.
The bias is NOT the same for all reference values. Tb=10.158, Tcritical=2.002
The linearity is NOT acceptable. Ta=12.043, Tcritical=2.002
In this example, there is a problem with linearity. The bias = 0 line (green on the chart) should be contained by the
upper and lower 95% confidence intervals. The equation is the title is the best fit equation for the individual
readings. The R squared value gives the % of variation in the bias that is explained by the variation in reference
values.
Attribute Gage R_R

An attribute gage R&R study is used when the measurement system has a finite number of categories, such as go/no
go gages. The attribute gage R&R portion of the program will analyze the results when the gage has only two
possible outcomes: pass or fail. The program will generate the effectiveness table, cross tabulations and kappa
scores. The example below uses the data from page 127 of Measurement System Analysis, Third Edition. The data
is also available in the workbook Gage R&R Example Workbook.xls (default installation folder: My Documents/SPC
for MS Excel).
The first step in using the Attribute Gage R&R program is to setup the data entry page.
2. Select the Setup the data entry sheet based on number of operators, parts, and trials" option from the second
tab ("Range/Bias/Linearity/Attribute Gage R&R") on the Measurement System Analysis form.
p26
3. Fill in the information in the Attribute Gage R&R Setup. Enter the number of operators, parts, and trials. Enter
the pass value and fail value (e.g., 1 and 0, pass and fail, etc.). Then enter the name of the Gage R&R study. The
values in the form below are for this example. Select OK to generate the template to enter the results.
4. Fill in the blank template that was generated on a new worksheet. Part of a completed template is shown below.
Enter the operator names in the upper right hand corner. The names will automatically appear in the first column.
NOTE: The program uses Microsoft Excels naming function to run. You cannot have spaces or certain characters
(e.g. /). Instead of using John Smith, use John_Smith.
Enter the rest of the information for Date, Gage Name, Gage Number, Gage Type, Product, Characteristic, Pass Value,
Fail Value, and Performed By. None of this information is required to run the program except the Pass Value and Fail
Value. Then fill in the data as shown below.
p27
6. Select the Run the Attribute Gage R&R analysis (have entered data) option from the second tab
("Range/Bias/Linearity/Attribute Gage R&R") on the Measurement System Analysis form. You have the option to
include the crosstabulations, kappa values and effectiveness table. The default is to include all three.
Attribute Gage R&R

Crosstabulations
The program will generate a crosstabulation table for each combination of operators. The table for A and B is shown
below.
Hal
Fail
Pass
Total
Hal * Beth Crosstabulation

Beth
Fail Pass Total Kappa
Count
44
6
50
0.86
Expected Count
15.7 34.3 50.0
Count
3
97 100
Expected Count
31.3 68.7 100
Count
47
103 150
Expected Count
47.0 103.0
The kappa value is given. If kappa is above 0.75 there is good agreement between the operators. If it is less than
0.40, there is poor agreement. The count information is as follows:
Hal and Beth both rate as fail: 44 times
Hal rates as fail and Beth rates as pass: 6 times
Hal rates as pass and Beth rates as fail: 3
Hal and Beth both rate as pass: 97 times
These tables will help you determine how well the operators agree with one another.
Kappa Values
The output for Kappa values is shown below. It includes the kappa values for the operators as well as the kappa
value compared to the reference values (if there are any).
Kappa Measures for the Operators
p28
Kappa
Hal
Beth
Loa
Hal
0.86
0.78
Beth
0.86
0.79
Loa
0.78
0.79
-
There is good to excellent agreement since all kappa values are greater than 0.75
Kappa Values For Each Operator to Reference
Kappa
Hal
0.88
Beth
0.92
Loa
0.77
There is good to excellent agreement since all kappa values are greater than 0.75
Effectiveness Table
The effectiveness table is shown on the next page. This table determines how effective each operator is.
If the calculated score for each operator falls within the confidence interval of the other operators, the effectiveness
of the operators is the same.
p29

Measurement System Analysis: Instruction Manual

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Measurement System Analysis: Instruction Manual

Uploaded by

Copyright:

Available Formats

Measurement System Analysis

BPI Consulting, LLC, 2009

BPI Consulting, LLC, 2009

SPC for Excel Help Files

Crossed Gage R&R: Average and Range Method

BPI Consulting, LLC, 2009

Crossed Gage R_R: Average/Range Method

) on the SPC toolbar.

BPI Consulting, LLC, 2009

5. Select the MSA icon (

) on the SPC toolbar.

BPI Consulting, LLC, 2009

10. Select OK to run the program and generate the results.

Crossed Gage R&R Average/Range Output

BPI Consulting, LLC, 2009

BPI Consulting, LLC, 2009

BPI Consulting, LLC, 2009

Crossed Gage R_R: ANOVA Method

1. Select the MSA icon (

) on the SPC toolbar.

4. Select OK to run the program and generate the results.

BPI Consulting, LLC, 2009

BPI Consulting, LLC, 2009

Nested Gage R_R

) on the SPC toolbar.

BPI Consulting, LLC, 2009

5. Select the MSA icon (

) on the SPC toolbar.

BPI Consulting, LLC, 2009

8. Select OK to run the program and generate the results.

Nested Gage R&R Output

BPI Consulting, LLC, 2009

BPI Consulting, LLC, 2009

Range Gage R_R Method

BPI Consulting, LLC, 2009

2. Select the data highlighted above.

) from the SPC toolbar.

6. Select OK to run the program and generate the results.

Range Gage R&R Method Output

BPI Consulting, LLC, 2009

Bias - Independent Sample Method

BPI Consulting, LLC, 2009

2. Select the data highlighted above.

) from the SPC toolbar.

BPI Consulting, LLC, 2009

7, Select OK to run the program and generate the results.

Bias Independent Sample Method Output

BPI Consulting, LLC, 2009

Bias - Control Chart Method

2. Select the data highlighted above.

) from the SPC toolbar.

BPI Consulting, LLC, 2009

6. Fill in the information on the "Data Input" tab.

BPI Consulting, LLC, 2009

9. Select OK to run the program and generate the results.

Bias Control Chart Method Output

BPI Consulting, LLC, 2009

t statistic = the t value based on the degrees of freedom

BPI Consulting, LLC, 2009

BPI Consulting, LLC, 2009

2. Select the data highlighted above.

) from the SPC toolbar.

6. Select OK to run the program and generate the results.