Aims
The Journal of Reproducible Statistics publishes statistics and applied mathematics whose results are reproducible wherever the data allow. We hold strict standards on method and evidence, and we keep review efficient and transparent: the decision process is significantly faster than at most journals. Every submission ships rerunnable code and data, or a privacy-preserving surrogate where the data cannot be shared.
What we publish
- Methodological work. New statistical methods, theory, estimation and inference procedures.
- Applications. Analyses of real data with substantive insight, not just benchmarks or toy examples.
- Computational statistics and software. Algorithms, implementations, reproducible tooling.
- Replications and null or negative findings. Explicitly welcomed and reviewed on equal terms, not as a residual category.
- Simulation studies with a transparent design and rerunnable code.
Reproducibility, as far as the data allow
Reproducibility is the journal's default expectation and its strongest signal. We pursue it as far as the data allow, and we never treat it as a hurdle that turns away sound work bound by privacy. Every contribution carries a reproducibility package: an executable source (Quarto, R Markdown or Jupyter preferred, LaTeX accepted), the code, and the data. Where the data cannot be shared, a privacy-preserving verifiable surrogate takes their place, anonymised or synthetic, with a pinned runtime environment. We assess reproducibility by re-execution wherever the data allow. Where privacy prevents it even with a surrogate, the work is still welcome, and the article states plainly that its results cannot be rerun on the original data. See the data & code availability policy for what this means in practice.
Out of scope
A few kinds of work fall outside what we can publish:
- Work with no path to reproducibility or verification at all: no code, no reproducibility package, and no willingness to document how the results were produced. Privacy restrictions on the data are not a reason for exclusion; we have a route for those.
- Topics outside statistics and applied mathematics.
- Contributions without their own methodological or empirical substance.
Soundness, not significance
We judge a submission on scientific soundness and reproducibility, not on novelty or how interesting the topic seems. A rigorous, correctly reported, reproducible study on an unfashionable question is in scope and publishable. This is not "accept everything". The bar is rigour plus reproducibility, and that bar is real. What we do not do is gatekeep on how exciting a topic is. Replications and null findings benefit most from this rule, and we welcome them deliberately.
How submissions are assessed, including the disclosed use of AI and the named human editor who signs every decision, is described in the peer-review policy.