Methods & sources

Show the method. State the assumptions.

All calculations should be checkable; and the way that their results can be interpreted, should match as closely as possible what you were trying to find out about your data.

Who prepares this material

CCHE publishes the explanations and tools on this website. The text and calculator code were developed with AI assistance. Worked examples and numerical functions are checked against the methods described on each page and independent numerical reference cases. Those checks establish how the implementation behaves; they do not amount to independent expert peer review of every guide.

Dr Rob Hayward is credited as site administrator. An author or reviewer credit on an individual page identifies a contribution to that page.

Original explanations

CCHE’s educational guides use original worked examples and explanations with references to primary statistics and research-methods sources. These references support the methods; they do not imply endorsement by the source authors or organisations.

Numerical implementation

Distribution functions used in this package use regularized incomplete beta and gamma functions. The inverse probability function used is based upon a bounded numerical search. The mean test power function is calculated via integration over the non-central t model. The balanced ANOVA power function is calculated via a Poisson mixture for the non-central F distribution. The correlation and proportion power functions use the approximation stated in the software.

Quartiles use linear interpolation at position (n − 1)p under the R type 7 convention. Here n is the observation count and p is the required probability, such as 0.25 or 0.75. Sample variance uses n minus 1; population variance uses n. The result identifies the relevant choice.

Checks and reproducibility

The source includes independent numerical reference cases created with SciPy, alongside calculator validation checks. These check numerical results; they do not establish that a model suits a particular study. Review the assumptions printed on each tool.

Primary references

Editorial status

CCHE is an independent educational source. The example problems and solutions provided in these resources can aid the reader in understanding the use of statistical methodology. Each numerical check evaluates the accuracy of a calculated result given a specified model. However, they do not indicate whether the model is suitable for all research questions. The checks do not provide independent medical review or establish formal institutional affiliation.