130
8 Measurement Uncertainty
and the benefits from its broad use. The process of evaluating of the uncertainty of
the measurement results, described in the ISO GUM guide, is a universal method,
possibly to be used regardless of the field of science or technology.
Uncertainty encompasses many factors that influence the result. Some components of the uncertainty can be obtained by statistical analysis of the data from
repeated measurements and can be expressed as the standard deviation; others are
estimated on the basis of the probability distribution. The most often used distributions are the normal (Gaussian), triangular or rectangular distribution.
Providing the result with the assigned uncertainty is required in the ISO/IEC
17025. But it is not the only argument in favor of using uncertainty. The procedure
of evaluating the uncertainty includes a very detailed analysis of the measurement
procedure, which is connected to the necessity to select and assess various factors
that influence the final result. That aspect of uncertainty forces the analyst to consider
the influence of various factors, and that can, in turn, be a good basis for the selection
of the work strategy in the case of a need to improve the analytical parameters of
the method. Therefore, during the analysis of particular parameters, it is possible to
assess which of the included factors is critical. This allows appropriate actions to be
taken that modify the measuring procedure used.
Evaluation of the uncertainty of the measurement result requires, most of all, a
good knowledge of the applied measuring procedure and the infrastructure in a given
laboratory. An essential part of the evaluation of the uncertainty is the selection of all
the parameters that influence the final result. Among the most often included parameters is the quality of the reagents or the reference materials and the performance of
the measuring devices used.
Commonly identified sources of uncertainty of chemical measurements are as
follows:
– Incorrectly defined measurand;
– Sample does not fully represent the feature of the entire object;
– The measurement procedure is not used correctly;
– Systematic errors occur;
– Lack of knowledge of the influence of environmental conditions and their fluctuation on the result of the measurement;
– Uncertainty of the calibration of the measuring device;
– Resolution of the measuring device;
– Uncertainty of the certified value of the reference standards and/or the reference
materials;
– Uncertainty of the physical constants and/or atomic masses used in the calculations
of the final result.
The measurand is a quantity intended to be measured and it depends on the analyzed
object and its feature to be tested. The definition of measurand for a given test is
not always trivial; definitely it should be related to the aim of performing testing. In
respect of the test result, the measurement uncertainty should always be given since
it is considered to be an intrinsic part of the measurement result. The numerical value
8 Measurement Uncertainty
and the benefits from its broad use. The process of evaluating of the uncertainty of
the measurement results, described in the ISO GUM guide, is a universal method,
possibly to be used regardless of the field of science or technology.
Uncertainty encompasses many factors that influence the result. Some components of the uncertainty can be obtained by statistical analysis of the data from
repeated measurements and can be expressed as the standard deviation; others are
estimated on the basis of the probability distribution. The most often used distributions are the normal (Gaussian), triangular or rectangular distribution.
Providing the result with the assigned uncertainty is required in the ISO/IEC
17025. But it is not the only argument in favor of using uncertainty. The procedure
of evaluating the uncertainty includes a very detailed analysis of the measurement
procedure, which is connected to the necessity to select and assess various factors
that influence the final result. That aspect of uncertainty forces the analyst to consider
the influence of various factors, and that can, in turn, be a good basis for the selection
of the work strategy in the case of a need to improve the analytical parameters of
the method. Therefore, during the analysis of particular parameters, it is possible to
assess which of the included factors is critical. This allows appropriate actions to be
taken that modify the measuring procedure used.
Evaluation of the uncertainty of the measurement result requires, most of all, a
good knowledge of the applied measuring procedure and the infrastructure in a given
laboratory. An essential part of the evaluation of the uncertainty is the selection of all
the parameters that influence the final result. Among the most often included parameters is the quality of the reagents or the reference materials and the performance of
the measuring devices used.
Commonly identified sources of uncertainty of chemical measurements are as
follows:
– Incorrectly defined measurand;
– Sample does not fully represent the feature of the entire object;
– The measurement procedure is not used correctly;
– Systematic errors occur;
– Lack of knowledge of the influence of environmental conditions and their fluctuation on the result of the measurement;
– Uncertainty of the calibration of the measuring device;
– Resolution of the measuring device;
– Uncertainty of the certified value of the reference standards and/or the reference
materials;
– Uncertainty of the physical constants and/or atomic masses used in the calculations
of the final result.
The measurand is a quantity intended to be measured and it depends on the analyzed
object and its feature to be tested. The definition of measurand for a given test is
not always trivial; definitely it should be related to the aim of performing testing. In
respect of the test result, the measurement uncertainty should always be given since
it is considered to be an intrinsic part of the measurement result. The numerical value