Identification of Blunders in PT Results and Determination of Assigned Value
This content discusses the methods for identifying blunders in proficiency testing (PT) results, determining the assigned value, and addressing outliers to ensure accurate calculations. Techniques include using consensus values, expert perception, and models specified in ISO standards. Methods such as Grubbs test are suggested for identifying and removing outliers in participant data to ensure reliable outcomes.
Download Presentation
Please find below an Image/Link to download the presentation.
The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author.If you encounter any issues during the download, it is possible that the publisher has removed the file from their server.
You are allowed to download the files provided on this website for personal or commercial use, subject to the condition that they are used lawfully. All files are the property of their respective owners.
The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author.
E N D
Presentation Transcript
Identification of Blunders in PT results S.SUBRAMANIAN
Identification of Blunders Determination of Assigned Value and its uncertainty as stipulated in chapter 7 of ISO 13528:2022: a) b) By using a CRM as PT item; c) By characterizing the PT item using a valid CRM in one laboratory (also called as Value transfer from a CRM to a closely matched candidate RM); d) By using consensus value from expert laboratories (which are not participants in the PT Scheme); and e) By using consensus value from participant laboratories By formulation (also called as Known Value Scheme); While methods at (a) to (d) are independent of participants (e) is based on participants results only. Any odd value (blunder) present in participants results can Assigned Value and its uncertainty. affect determination of 2 06-03-2025
Identification of Blunders Determination of SDPA as stipulated in chapter 8 of ISO 13528:2022: a) By perception of experts; b) By experience from previous rounds of a PT Scheme ; c) By use of a General Model (Horwitz equation); d) By using repeatability & reproducibility SDs from a previous collaborative study of precision of a measurement method; e) From data obtained in the same round of a PT Scheme While methods at (a) to (d) are independent of participants (e) is based on participants results only. Any odd value (blunder) present in participants results can affect determination of SDPA. 3 06-03-2025
Identification of Blunders Thus, when either Assigned Value, xpt and its uncertainty, u(xpt) and/or SDPA, pt is to be determined using participants results only, then prior to conduct of any statistical analysis of the data, odd values (blunders) have to be identified and removed for the calculation of xpt, u(xpt)as well as pt. 4 06-03-2025
How to identify Blunders We will now identify outliers (blunders) in the participants results by using the following three methods: Grubbs test Using median & MPE Results with Z/Z score 5 and below or 5 and above 5 06-03-2025
How to identify Blunders 1) Initially do Grubbs test to identify outliers in the participants result and remove them for calculating either AV and/or SDPA. 2) Find the median of the remaining participants results . Identify MPE for this parameter in a realistic manner. This is critical. 3) Calculate Median- MPE and Median + MPE 4) Exclude participants results which are beyond these two limits. 5) Using the remaining participants results calculate AV, SDPA and SU by Algorithm A or nIQR method, as per statistical design selected 6) Calculate Z score or Z prime score as applicable 7) Remove those participants results which are 5 and below as well as +5 and above 06-03-2025 6
How to identify Blunders 8) Using the remaining participants results calculate AV, SDPA and SU by Algorithm A or nIQR method, as per statistical design selected Calculate the Z score or Z prime score as applicable for the remaining participants results 10) Also calculate Z score or Z prime score as applicable for all the results which are removed for calculating AV, SDPA and SU 11) The Z/Z score for the removed results should be in the range of -2 to 2. If it is not so do not consider these results as outliers and include them for calculating AV, SDPA and SU. 12) Calculate the Z score or Z prime score as applicable for all the participants results including the outliers 9) 7 06-03-2025
Identify Blunder: ORIGINAL RESULTS Original results Results re- arranged Code 2 5 9 10 17 19 21 24 27 28 30 32 33 37 39 Result 314.7 317.0 327.8 313.8 335.7 330.5 328.8 302.9 385.1 339.6 342.1 320.0 322.0 338.4 248.9 Code 39 24 10 2 5 32 33 9 21 19 17 37 28 30 27 Result 248.9 302.9 313.8 314.7 317.0 320.0 322.0 327.8 328.8 330.5 335.7 338.4 339.6 342.1 385.1 06-03-2025 8
Identify Blunder: Grubbs test -GSM Code Result 39 24 10 2 5 32 33 9 21 19 17 37 28 30 27 G Gcrit 248.92.6882.549 302.9 313.8 314.7 317.0 320.0 322.0 327.8 328.8 330.5 335.7 338.4 339.6 342.1 385.12.1552.549 Code Result 302.91.3832.507 313.8 314.7 317.0 320.0 322.0 327.8 328.8 330.5 335.7 338.4 339.6 342.1 385.12.8302.507 G Gcrit 24 10 2 5 32 33 9 21 19 17 37 28 30 27 Code Result 24 10 2 5 32 33 9 21 19 17 37 28 30 G Gcrit 302.9 1.930 2.462 313.8 314.7 317.0 320.0 322.0 327.8 328.8 330.5 335.7 338.4 339.6 342.1 1.397 2.462 Average 324.49 SD 06-03-2025 329.89 19.509 Average 325.64 SD Average SD 28.119 11.778 9
Identify Blunder: MPE-GSM Code 24 10 2 5 32 33 9 21 19 17 37 28 30 Result 302.9 313.8 314.7 317.0 320.0 322.0 327.8 328.8 330.5 335.7 338.4 339.6 342.1 Median MPE 327.80 19.668@6% Median - MPE= 308.1 Median + MPE= 347.5 06-03-2025 10
Identify Blunder: IzI or IzI >5-GSM z' score -1.14 -1.06 -0.87 -0.62 -0.46 0.02 0.10 0.25 0.68 0.90 1.00 1.21 Code 10 2 5 32 33 9 21 19 17 37 28 30 Result 313.8 314.7 317.0 320.0 322.0 327.8 328.8 330.5 335.7 338.4 339.6 342.1 AV= SDPA= SU= 327.53 11.364 4.101 11 No result has Z' beyond 5 06-03-2025
Performance evaluation for all results AV= SDPA= SU= 327.533 11.3640 4.1006 z' score -1.06 -0.87 0.02 -1.14 0.68 0.25 0.10 -2.04 4.77 1.00 1.21 -0.62 -0.46 0.90 -6.51 Code 2 5 9 10 17 19 21 24 27 28 30 32 33 37 39 Result 314.7 317.0 327.8 313.8 335.7 330.5 328.8 302.9 385.1 339.6 342.1 320.0 322.0 338.4 248.9 24,27 &39 are excluded. But Z' score of 24 is -2.04 & hence can't be treated as "outlier . Include code 24 in main group 12 06-03-2025
Revised Performance evaluation for all results AV= SDPA= SU= 325.951 12.6584 4.3885 z' score -0.84 -0.67 0.14 -0.91 0.73 0.34 0.21 -1.72 4.41 1.02 1.21 -0.44 -0.29 0.93 -5.75 Code 2 5 9 10 17 19 21 24 27 28 30 32 33 37 39 Result 314.7 317.0 327.8 313.8 335.7 330.5 328.8 302.9 385.1 339.6 342.1 320.0 322.0 338.4 248.9 27 & 39 are only removed for calculation. Z' score of both are >3.0 13 06-03-2025
