Reliability of repeatability and reproducibility measures in collaborative trials.

Area of Science:

• Analytical Chemistry
• Method Validation

Background:

• Analytical method validation typically involves assessing repeatability and reproducibility through collaborative studies.
• Acceptance criteria for new methods are often based on these statistical estimates.

Purpose of the Study:

• To present procedures for calculating confidence intervals and operating characteristic curves for method acceptance criteria.
• To compare the reliability of repeatability and reproducibility estimates with varying numbers of collaborators.

Main Methods:

• Utilized collaborative study data to derive repeatability and reproducibility estimates.
• Developed methods for calculating confidence intervals and operating characteristic curves.
• Analyzed the impact of the number of collaborators on estimate reliability.

Main Results:

• Established procedures for evaluating method acceptability based on statistical estimates.
• Demonstrated that a small number of collaborators significantly reduces the reliability of reproducibility estimates.
• Highlighted that method decisions may unduly favor repeatability over reproducibility with limited collaborators.

Conclusions:

• Method acceptability decisions require careful consideration of the number of collaborators in validation studies.
• Ensuring sufficient collaborators is crucial for obtaining reliable reproducibility estimates, which are key for method validation.