A full report on the precision determination should be given for any ITP. This is a comprehensive report of all ITP details, not the report that each participating laboratory prepares and returns as part of the ITP. This full report should contain information on the details of the organization and execution of the programme as follows:
a) the organization committee, where located, co-ordinator, dates of ITP;
b) the level of precision: level 1 or level 2;
c) the type of precision: type 1, type 2;
d) the number of laboratories, p (list their names without connection to the ITP lab number);
e) the number of materials or target materials, q, plus a description of these;
f) the definition of a test result, the number of replicates, n, and the time span for repeatability;
g) information on the technicians who conducted the testing: one or more, any special details;
h) details of the preparation of the materials and how homogeneity was documented;
i) details on packaging and delivery of materials to the ITP participants;
j) copies of all ITP data reports from each participating lab;
k) the ITP analysis report, with all tables as designated in Annex E, a full description of all analysis steps, the options chosen for outlier rejection, plus all other required comments;
l) the table of precision results, plus any comments on the outcome;
m) a draft of the precision clause for inclusion in the test method standard.
Copyright International Organization for Standardization
--`,,``,`-`-`,,`,,`,`,,`---
© ISO 2005 – All rights reserved 21
NOTE See example of precision calculation in Annex D for tables with data.
Figure 1 — Decision tree diagram for ITP level 1 data analysis
22 © ISO 2005 – All rights reserved
Table 1 — Level 1 precision — Basic dataa Material, M(j)
Laboratory, L(i)
1 2 3 4 ………… q
1
2 Yijk
3 4
………
… p Notation used:
There are a total of p laboratories: L(i) = 1, 2, 3, ... p.
There are a total of q materials or levels: M(j) = 1, 2, 3, ..., q.
There are a total of n replicates per cell: A cell = each combination of L(i) and M(j); normally n = 2.
Yijk = a single test result, where k = 1, 2, … n(ij) and n normally = 2; see cell (2,3) in table for example.
Cell (ij): each cell contains n test result values.
a Table layout for uniform level ITP.
Table 2 — Level 1 precision — Cell averagesa Material, M(j)
Laboratory, L(i)
1 2 3 4 … q
1
2 Avg Yijk
3 4
… p Notation used:
There are a total of p laboratories: L(i) = 1, 2, 3, … p.
There are a total of q materials or levels: M(j) = 1, 2, 3, …, q.
There are a total of n replicates per cell: A cell = each combination of L(i) and M(j); normally n = 2.
Avg Yijk = average of n test results.
a Table layout for uniform level ITP.
Copyright International Organization for Standardization
--`,,``,`-`-`,,`,,`,`,,`---
© ISO 2005 – All rights reserved 23 Table 3 — Level 1 precision — Cell standard deviationsa
Material, M(j) Laboratory, L(i)
1 2 3 4 … q
1
2 Std dev Yijk
3 4
… p Notation used:
There are a total of p laboratories: L(i) = 1, 2, 3, … p.
There are a total of q materials or levels: M(j) = 1, 2, 3, …, q.
There are a total of n replicates per cell: A cell = each combination of L(i) and M(j); normally n = 2.
Std dev Yijk = standard deviation of cell (ij) for n test results.
a Table layout for uniform level ITP.
Table 4 — Initial data format for each material (Level 2 precision — Carbon black testing)
Material, M(j) Date
Test result 1 Test result 2
Operator or technician
Day 1 xxx xxx xxxxx
Day 2 xxx xxx xxxxx
See notes to Table 5.
Table 5 — Format for interlaboratory data (Level 2 precision — Carbon black testing)
Material 1 Material 2 Material q
Lab
number Cell avg Cell
std dev Cell avg Cell
std dev Cell avg Cell std dev
1 xx xx xx xx xx xx
2 xx xx xx xx xx xx
… xx xx xx xx xx xx
p xx xx xx xx xx xx
Note 1 Materials are typically different grades or types of carbon black.
Note 2 The data in Table 4 (for each material) constitutes a “cell”, i.e. avg and std dev are calculated for four data values.
--`,,``,`-`-`,,`,,`,`,,`---
24 © ISO 2005 – All rights reserved
Table 6 — Example of level 1 and 2 precision table organization
(Level 1 or 2 and type 1 or 2a — Precision for ISO XXXXX measured property = xxxxxx, in xxb) Within laboratory Between laboratories
Material Mean
level sr r (r) sR R (R) No. of labsc
A B C D Pooled or avg values Notation used:
sr = within-laboratory standard deviation (in measurement units);
r = repeatability (in measurement units);
(r) = repeatability (in percent of mean level);
sR = between-laboratory standard deviation (for total between-laboratory variation in measurement units);
R = reproducibility (in measurement units);
(R) = reproducibility (in percent of mean level).
See text of precision clause for discussion of precision results given in this table.
a Indicate the level of precision (1 or 2) and the type of precision (1 or 2) in the table heading.
b ISO XXXXX = reference number of test method standard; xxxxxx = property measured; xx = units of property.
c List number of labs in final database. Also list the option chosen: if option 2, indicate number of labs in parentheses ( ).
Copyright International Organization for Standardization
--`,,``,`-`-`,,`,,`,`,,`---
© ISO 2005 – All rights reserved 25
Annex A (normative)
Calculating the h and k consistency statistics