10.4 The Use of Reference Method
175
10.4 The Use of Reference Method
In cases where the reference method is used, the reference value is determined by
either by a primary method or by a fully validated procedure, including calibration
against a standard, with given traceability to national or international standard. An
assigned value can be thus delivered by a single laboratory, analyzing the ILC/PT
test item, and the target range is derived from the uncertainty of the performed measurement. This scenario is only possible when suitable CRM or chemical standards
are available.
10.5 Consensus Value (Expert Laboratories)
In cases where the consensus value from expert laboratories is utilized, the results of
measurements conducted by a group of expert laboratories are used. The individual
laboratory is recognized as an expert laboratory, if it can prove its performance in
the scope of a given determination (in a given sub-discipline) and if it uses validated
measurement methods. Expert laboratories can be engaged in measurement conduct
before the samples are sent to the participants of the proficiency testing or they can
conduct tests in parallel, on-line with a given round of the PT.
The ILC/PT provider must account for the fact that not all results submitted by the
expert laboratories are reliable, they can also be exposed to an unidentified systematic
error. In that case, engaging a larger number of expert laboratories can limit the risk,
as it is assumed to be highly unlikely that all laboratories would make such a mistake.
It is, however, not always possible to find a larger number of expert laboratories with
high technical competencies in a specific sub-discipline.
10.6 Consensus Value (from Participants)
In cases where the consensus value from participants is utilized, the assigned value
is delivered as a consensus from the results of all participants, using the statistical
methods described in ISO 13528, with consideration of the effects of outliers. The
reference value is calculated as the mean value of the results submitted by the participants, after the rejection of deviating results or with the use of robust statistics. The
most recommended are the Dixon tests (Q test) or Grubbs tests or the use of robust
statistics in which the influence of the extreme values on the mean value is limited.
Determination of the reference value on the basis of the results of all participants
of the given round is a relatively simple and inexpensive method, hence it is readily
used by the providers of PT. The method can be used for any type of test items, which
allows for best fit of the sample type to the given sub-discipline. However, it should
be remembered that the results obtained by the participants should be consistent
175
10.4 The Use of Reference Method
In cases where the reference method is used, the reference value is determined by
either by a primary method or by a fully validated procedure, including calibration
against a standard, with given traceability to national or international standard. An
assigned value can be thus delivered by a single laboratory, analyzing the ILC/PT
test item, and the target range is derived from the uncertainty of the performed measurement. This scenario is only possible when suitable CRM or chemical standards
are available.
10.5 Consensus Value (Expert Laboratories)
In cases where the consensus value from expert laboratories is utilized, the results of
measurements conducted by a group of expert laboratories are used. The individual
laboratory is recognized as an expert laboratory, if it can prove its performance in
the scope of a given determination (in a given sub-discipline) and if it uses validated
measurement methods. Expert laboratories can be engaged in measurement conduct
before the samples are sent to the participants of the proficiency testing or they can
conduct tests in parallel, on-line with a given round of the PT.
The ILC/PT provider must account for the fact that not all results submitted by the
expert laboratories are reliable, they can also be exposed to an unidentified systematic
error. In that case, engaging a larger number of expert laboratories can limit the risk,
as it is assumed to be highly unlikely that all laboratories would make such a mistake.
It is, however, not always possible to find a larger number of expert laboratories with
high technical competencies in a specific sub-discipline.
10.6 Consensus Value (from Participants)
In cases where the consensus value from participants is utilized, the assigned value
is delivered as a consensus from the results of all participants, using the statistical
methods described in ISO 13528, with consideration of the effects of outliers. The
reference value is calculated as the mean value of the results submitted by the participants, after the rejection of deviating results or with the use of robust statistics. The
most recommended are the Dixon tests (Q test) or Grubbs tests or the use of robust
statistics in which the influence of the extreme values on the mean value is limited.
Determination of the reference value on the basis of the results of all participants
of the given round is a relatively simple and inexpensive method, hence it is readily
used by the providers of PT. The method can be used for any type of test items, which
allows for best fit of the sample type to the given sub-discipline. However, it should
be remembered that the results obtained by the participants should be consistent