About this calculator
This calculator evaluates measurement uncertainty when variables share covariance. Many real measurements are not independent. A temperature probe, scale, calibration curve, and reference standard can move together. Ignoring correlation can understate risk. It can also overstate uncertainty when variables move in opposite directions. The tool handles both cases. It accepts measured values, standard uncertainties, sensitivity terms, degrees of freedom, and correlation coefficients. It then builds a covariance matrix. The final combined uncertainty follows the law of propagation of uncertainty.
Why covariance matters
Covariance describes linked variation. Positive correlation increases the total result when sensitivities share the same sign. Negative correlation may reduce the combined value. A mixed sensitivity pattern can reverse those effects. This is useful for laboratories, finance checks, process studies, and engineering estimates. The calculator displays variance parts so every term stays visible. It also reports expanded uncertainty with a selected coverage factor.
Advanced measurement workflow
Start by choosing the model type. Direct sensitivity mode works for any formula when partial derivatives are known. Linear mode calculates a result from coefficients and measured values. Product power mode supports multiplication, division, and exponent relationships. Enter standard uncertainties in the same units as each variable. Add correlations between minus one and one. Leave unused variables blank. For traceability, enter degrees of freedom when they are known. The effective degree of freedom is approximate when covariance terms are included.
Reading the results
The result block gives the calculated output, combined standard uncertainty, expanded uncertainty, and relative uncertainty. The covariance table shows variance contributions. Large positive terms point to paired inputs that increase risk. Large negative terms show cancellation. Review these terms before reporting a final interval. A compact report can be downloaded as CSV or PDF. Use the files for worksheets, audit trails, or quick review notes.
Practical reporting tips
Use realistic input uncertainties. Check signs for sensitivity coefficients. Confirm that correlations are justified by data, calibration history, or expert judgment. Do not add correlation only to make results smaller. Document every assumption. Repeat the calculation when data changes. For formal work, compare the output with your quality system, measurement guide, or statistical procedure. Keep units consistent. Save inputs before sharing results with reviewers and clients later too.