This repository contains the code and TeX sources for the manuscript
`A review of Bayesian perspectives on sample size calculation for confirmatory trials'.
The R code to reproduce the figures is contained in the Jupyter notebook
sample-size-calculation-under-uncertainty.ipynb
.
The notebook uses R and the tidyverse metapackage.
All software dependencies are documented in the .binder
subfolder.
The repository is compatible with the Python tool repo2docker.
This means that the repository can be turned into a Docker container that contains all
required software using repo2docker.
Using this container to execute the code then ensures a maximum level of
reproducibility.
We also use the (free) service mybinder.org to provide an interactive version of the repository online that can be explored without installing any software on your local system.
If the command nbconvert
is available on the system,
the figures can also be created non-interactively by executing the command
jupyter nbconvert --to notebook --execute --ExecutePreprocessor.timeout=600 sample-size-calculation-under-uncertainty.ipynb
We also provide a simple Shiny app that enables exploring the sample size calculation rules interactively. The most recent development version can be accessed by clicking on the badge below.
Versioned releases
Older releases of the repository are available via the badges listed below.