Statistical Methods

Transfer gamma parameters without confusing rate and scale

Define the task

A gamma distribution uses a positive shape and either a positive rate or its reciprocal scale. Record both the numerical parameter and its role. Rate has reciprocal outcome units, while scale has outcome units. Use a dimensional check before copying values from a model specification into a simulation call.

Carry out the check

For shape a and scale s, mean is a×s and variance is a×s². With rate b, the same quantities are a/b and a/b². Compute these moments on both sides of a transfer. If the expected center changes, stop and resolve the convention before comparing simulation output.

Worked hypothetical example

For a hypothetical shape of 4 and rate of 2 per hour, the equivalent scale is 0.5 hours. The mean is two hours and variance is one squared hour. Accidentally treating the rate value of 2 as a scale produces mean eight hours and variance sixteen squared hours. Store the intended mean and variance beside the original parameters so a receiving pipeline has a quick reproducibility check. Estimated parameters also require their uncertainty and fitting assumptions.

Method checked 11 November 2026 against R documentation: The Gamma Distribution.

Download the rate and scale transfer ledger.

Related reading

Reading path: Analyze data. Guide 122 of 122. Choose any guide to suit your project.

Statistical Methods

Inspect beta boundary behavior before choosing shape parameters

Distinguish beta shapes that represent different behavior near zero and one.

Statistical Methods

Translate a constant rate into a waiting-time probability

Translate a constant event rate into a waiting-time probability. Includes a worked hypothetical example and checks for parameter-transfer errors.

Browse this reading pathTools for this stageGuides and downloads for this stage