Research knowledge base
Make your analysis reproducible
Plan the analysis so that another authorised colleague can trace a table or figure back to the data, decisions and software that produced it.
01
Agree the question and the analysis unit
Before choosing a model, record the research question, outcomes, predictors, sampling structure and analysis population. Distinguish a planned analysis from exploratory work. The data-preparation guide describes the materials that make an initial statistical consultation useful.
02
Preserve the source and document transformations
Keep source data read-only and create a separate working dataset. Document variable definitions, coding, missing values, exclusions and derived measures. Where data cannot leave an approved environment, arrange an authorised review process rather than placing restricted material in a public project.
03
Choose an analysis environment you can explain
R Project supports scripted statistical work and graphics. jamovi provides a statistical interface that can be useful for teams working with saved analyses. Choose according to the method, team skills, review needs and export requirements. Neither tool determines whether the model answers the research question.
04
Keep outputs connected to decisions
Retain the code or analysis file, software versions and relevant settings. Log changes to the analysis population and explain sensitivity analyses. Where random procedures are involved, retain the appropriate reproducibility settings and recognise that results can still depend on software or numerical differences.
05
Run a colleague handover check
Ask an authorised colleague to start from the supplied instructions and reproduce one important table and figure. Log missing files, ambiguous steps and dependencies. Resolve the gaps, rerun the affected work and retain the reviewed version. This is a check of traceability, not independent validation of the research conclusions.
The worked analysis-handover case study shows a practical acceptance record. When preparing public outputs, follow the data-sharing workflow and check the intended journal’s requirements for data, code and supplementary files.
07
Continue with your research
External resources reviewed 22 October 2026. Check the current instructions for the selected journal, tool or institution before acting. These links identify independent resources and do not imply an affiliation with or endorsement by their providers.