Reproduce results from: Cai et al. (2023)
A FAIR-compliant Snakemake workflow reproducing the simulation study of Cai et al. (2023) comparing Fully Conditional Specification (FCS) and Joint Modelling (JM) multiple imputation on multivariate data with missing values. The workflow runs 500 Monte Carlo replications on a trivariate normal dataset with 50% MCAR missingness, implemented in R using the mice package for FCS and a custom Gibbs sampler for JM. Seeds are fixed per replication to ensure exact numerical reproducibility across environments.
Main steps
compare_posterior_distributions — runs MC replications and generates the Q-Q plot.
evaluate_fcs_imputation — evaluates FCS coverage and bias.
evaluate_jm_imputation — evaluates JM coverage and bias.
evaluate_order_effect — tests ordering sensitivity. All four rules are independent and can run in parallel.
No data
What is a RO-Crate?
A RO-Crate is a standardized research object package that bundles a workflow together with rich machine-readable metadata. Each RO-Crate contains:
- the workflow definition file (e.g. using Snakemake)
- a ro-crate-metadata.json file describing the workflow's purpose, inputs, outputs, and provenance
- references to tool dependencies, software versions, and execution environments
- persistent identifiers linking to related resources such as publications, datasets, or other workflow registries
This ensures that the workflow can be reliably shared, cited, validated, and re-executed in a reproducible manner. RO-Crates follow the RO-Crate specification and are natively supported by various platforms.
More information can be found here.
This page was built for workflow: Reproduce results from: Cai et al. (2023)
