What this tool does
Review a batch of paired results without treating nondetects as zero or applying one percent criterion at every concentration.
When to use it
For already paired laboratory or field duplicates with matching methods, units and reporting bases. Use the same criterion only for an appropriate analyte/method group; run separate batches when limits differ.
Inputs and supported units
Use the labelled laboratory units and a consistent sample basis. Values are not silently interpreted as a regulatory method. Numeric input changes clear earlier results; CSV files are read locally, without uploading. The demonstration values only illustrate the arithmetic.
Calculation or decision method
For two uncensored concentrations with positive mean, RPD = 100 × |a−b| / ((a+b)/2). If either result is below your entered RL multiplier, compare absolute difference with your entered absolute limit instead. Equality to the limit meets that arithmetic comparison.
If either result is written as <number, the tool retains it and marks the pair not assessed; it makes no zero, half-RL or full-RL substitution. Two exact zeros have undefined RPD and can only be assessed by an enabled absolute-difference policy. No automatic qualifier or batch-acceptance flag is assigned.
How to interpret the result
Read the breakdown, exclusions and input record before the headline. Meeting an entered numeric condition is not a complete analytical quality decision. Retain the original bench sheet and resolve differences with the laboratory reviewer.
Worked example
10 and 12 mg/L give an absolute difference of 2 mg/L and an RPD of 18.1818%. For 0.05 and 0.06 at RL 0.02, the demonstration multiplier 5 selects an absolute difference of 0.01 instead. A result written <0.02 is preserved as censored and receives no numeric precision verdict.
Assumptions
The entered measurements describe the same stated sample, method and basis. Inputs must already account for any upstream dilution not explicitly modeled. Criteria come from the governing SOP, not from this website.
Limitations
RPD evaluates duplicate precision, not analytical accuracy or recovery. Do not use percent differences for logarithmic pH values or unmatched units, methods, dilution bases or sample identities. The project's real low-level/nondetect rules may differ; review excluded pairs under those rules rather than editing censor symbols to force a number. Repeated results can agree and still share the same bias.
Sources
- U.S. EPA — Duplicate Relative Percent Difference definition
- USGS — Design, Analysis, and Interpretation of Field Quality-Control Data