Evaluation of analytical performance based on partial order methodology.

Area of Science:

• Analytical Chemistry
• Data Analysis
• Method Validation

Background:

• Traditional performance measurements often rely on linear scales, which may be insufficient for comprehensive analytical performance evaluation.
• Partial order methodology provides an alternative for analyzing data matrices and simplifying relative comparisons based on data profiles.

Purpose of the Study:

• To introduce and illustrate the application of partial order methodology for evaluating analytical performance.
• To demonstrate how this method can be used for interlaboratory comparisons, method comparisons, and method optimization.

Main Methods:

• Partial order analysis of data matrices, treating it as an ordinal analysis.
• Simultaneous evaluation of multiple performance indicators (mean, standard deviation, skewness) without data pretreatment.
• Application to data from interlaboratory comparisons and proficiency testing.

Main Results:

• Establishes a partial ordering of participating laboratories.
• Quantifies the 'distance' of each laboratory from a reference laboratory.
• Classifies laboratories based on the concept of 'peculiar points'.

Conclusions:

• Partial order methodology provides a robust framework for evaluating analytical performance by considering multiple indicators simultaneously.
• This approach offers a "distance" metric from the true value, enhancing the understanding of analytical performance.
• The methodology is versatile, applicable to laboratory performance, method comparison, and optimization without requiring data pre-processing.