Depending on the type of kit and results required in the technical instructions, several types on analysis are possible
Quantitative data - measurable information expressed as a numerical value (Ultimate Tensile Strength, Brinell hardness, Aluminium concentration…):
The acceptance classes are based on the participants results submitted before the deadline.We apply algorithm A from ISO 13528 and we obtain an value (x*) and a standard deviation for proficiency assessment (s*).
If we ask for several results per characteristic, we also apply algorithm S from ISO 13528 and obtain a repeatability standard deviation (w*).
The acceptance classes are defined as follow :
Acceptance classes on the mean value
Class 1 : x* +/- 1 s*
Class 2 : x* +/- 2 s*
Class 3 : < x* - 2 s* or > x* + 2 s*
Acceptance classes on the repeatability (if applicable)
Class 1 : 0 to w*
Class 2 : w* to 2 w*
Class 3 : > 2 w*
Qualitative data - descriptive or categorical information that cannot be measured numerically (inclusion type, broken/non broken sample…)
You will find in the technical instructions how each chracteristic is analysed
- Mode analysis (microstructure type for example)
The assigned value is based on the answer given the most by participants
- Expert analysis (inclusion analysis for example)
PTP designates a team of internal and/or external experts to set up assigned values and acceptance classes. The participants results are compared to the experts analysis.
- Median analysis (number of a pits on a corroded sample for example)
The assigned value is based on the median of the participant population.
The acceptance classes are defined as follow :
Class 1 : Percentile 0,2 to 0,8
Class 2 : Percentile 0,1 to 0,9
Class 3 : < Percentile 0,1 et > Percentile 0,9