Repeatability and Reproducibility Limits: ISO 5725 Critical Difference
What Are Repeatability and Reproducibility Limits?
ISO 5725 characterizes a test method precision with two indicators: the repeatability limit r is the critical difference between two independent results from the same laboratory, operator, equipment and short time interval (r = 2.8 times sr), and the reproducibility limit R is the critical difference between results from different laboratories or conditions (R = 2.8 times sR). The 2.8 factor comes from the 95% confidence critical value for the difference of two results and is the common criterion for method precision assessment.
How to Use It
Whenever you need to judge whether two results agree, in laboratory comparisons, method validation or routine testing: the absolute difference |x1 - x2| of r or less means repeatability is acceptable under the same conditions, R or less means reproducibility is acceptable under different conditions, and exceeding the critical difference means the difference exceeds the method allowed random variation, so investigate causes such as inconsistent operation, sample inhomogeneity, instrument drift or contamination. This judgment is widely applied in result review, retained-sample retesting and interlaboratory comparison evaluation.
What Data to Enter
Enter the method repeatability standard deviation sr and reproducibility standard deviation sR (from the method standard, collaborative study report or your laboratory precision data), and the tool automatically computes r, R and the critical difference for judgment. You can also enter two results directly and the tool judges whether they agree and suggests cause investigation if the difference exceeds the limit. Entering a single result gives a reference critical difference against the mean, and both input modes cross-check each other.
Precautions
The r and R values are approximations at the 95% confidence level (2.8 = 1.96 times the square root of 2). If the method standard gives relative standard deviations (RSD), multiply by the mean to get absolute standard deviations; when comparing a single result with the mean, use correction factors such as r divided by the square root of 2, which the tool handles by scenario. Precision data should come from enough collaborative trials or accumulated data to be representative, and the method applies to chemical, physical and microbiological test methods.