Instructions:

1) Add reference information to cells highlighted in light yellow:

2) Add required information to all cells highlighted in gold:
Note: Sample data is only required for the number of samples and appraisers actually used - you may leave some gold boxes empty.
3) Results are shown in cells highlighted in light blue:
4) See the results sheet
5) Manually adjust the scale on the chart (if necessary) so thelimits are wide enough to see the variation.

Note: This worksheet will not perform properly if you modify it.

ATTRIBUTE MEASUREMENT SYSTEMS ANALYSIS

REPEATABILITY & REPRODUCIBILITY DATE

GAGE NAME CHARACTERISTIC

GAGE TYPE TEST NUMBER

GAGE NUMBER APPRAISER A NAME

NUMBER OF PARTS (10 to 100) 20 APPRAISERS 2 APPRAISER B NAME

CONFIDENCE LEVEL 0.95 TRIALS 2

BINARY OR ORDINAL
NUMBER OFCATEGORIES
DATA Binary 2 2
Binary 2 1
COMPARE TO
STANDARD? No 2 Ordinal 2
Binary
Category 1 0 1
Category 2 1 Yes
Category 3 2 No
Category 4 3 0
Category 5 4
 0 0 0 0 0 0 0
DATA INCOMPLETE DATA COMPLETE Clear Data

Standard/
Observation Reference Trial 1 Trial 2 Trial 3 Trial 4 Trial 5 Trial 1 Trial 2 Trial 3 Trial 4 Trial 5 Trial 1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
 97
98
99
100
1
2

3
4

5
 0

Trial 2 Trial 3 Trial 4 Trial 5 Trial 1 Trial 2 Trial 3 Trial 4 Trial 5 Trial 1 Trial 2 Trial 3 Trial 4 Trial 5
 ATTRIBUTE MEASUREMENT SYSTEMS ANALYSIS
2
Within Appraiser
Confidence Interval for Percent
Appraiser Number of parts inspected Number of agreements Percent agreements
agreements
20 0 0.00% ( 0.00% 13.91% )
20 0 0.00% ( 0.00% 13.91% )

Between Appraisers
All 20 20 100.00% ( 86.09% 100.00% )

Chart Appraisers
C.I. Limits
Upper limit 13.91% 13.91%
Lower limit 0.00% 0.00%
Estimate 0.00% 0.00%

Confidence Intervals for Percent Agreement Within Appraisers

100%
Percent

80%

Upper
60% limit

40%

20%

0% Appraiser
 0% Appraiser

0
Each Appraiser vs.
Standard
Confidence Interval for Percent
Appraiser Number of parts inspected Number of agreements Percent agreements
agreements
20 #VALUE! ( #VALUE! #VALUE! )
20 #VALUE! ( #VALUE! #VALUE! )

All Appraisers vs.

Standard
All 20 20 100.00% ( 86.09% 100.00% )

Chart Appraisers
C.I. Limits
Upper limit #VALUE! #VALUE!
Lower limit #VALUE! #VALUE!
Estimate #VALUE! #VALUE!

Confidence Intervals for Percent Agreement

Appraisers vs. Standard
100%
Percent

80%

60% Upper
limit

40%

20%

0%
Appraiser
Kappa Statistic click here for guidelines

Kendall Statistic click here for guidelines

Home
Guidelines for Kappa and Kendall
For Interpretation purposes on the Kappa-Kendall output worksheet:
Fleiss’ Kappa:
Use the Kappa statistic when you have nominal data with two (binary) or more levels with no natural
ordering, such as Pass and Fail, or red, blue and green.

Kappa values range between -1 and 1, where 1 is perfect agreement, 0 is what would be expected
by chance, and negative values indicate a potentially systematic disagreement between the
appraisers (however, negative values are very rare). In general, the higher the value of kappa, the
stronger the agreement.

Kendall’s Coefficient of Concordance:

Use this coefficient when the data are ordinal with three or more levels, and the standard is not
known. It can only be calculated when there are multiple appraisers and/or multiple trials per
appraiser. Its values range between 0 and 1. In general, the higher the value of Kendall's coefficient
of concordance, the stronger the association.

Kendall’s Correlation Coefficient:

Kendall's correlation coefficient measures the degree of association between appraisers’ ordinal
ratings and a known standard. It is therefore used when the data are ordinal and the standard is
known. Like any other correlation coefficient, it ranges between -1 and 1. Positive values indicate
positive association; negative values indicate negative association. In general, the higher the
magnitude of the correlation coefficient, the stronger the association.
levels with no natural

would be expected
between the
alue of kappa, the

e standard is not
ltiple trials per
Kendall's coefficient

ppraisers’ ordinal
d the standard is
ve values indicate
he higher the