# R in Official Stats

Any changes to this abstract?

New:

*Peer-Reviewing Code in Statistical R Packages: Standards, Processes, and Tools*
Mark Padgham and Noam Ross, rOpenSci

> Peer-review of statistical algorithms is challenging and labor intensive, and even software described in peer-reviewed literature rarely has its source code reviewed. Here we describe a peer-review system to support reviewing R code implementing statistical algorithms, developed by rOpenSci, an organization promoting open science through code and data sharing and peer code review. We have created a series of standards, guidance for authors and reviewers, and automated tools to facilitate review of statistical code. These include (1) a series of community-developed standards for documentation, testing, code architecture and programming APIs, of packages, with specialized sub-standards for specific areas such as regression, machine learning and Bayesian algorithms, (2) a series of packages for automated testing and literate programming to document standards compliance as part of code development, and (3) an open peer-review process supported by automated testing, reporting, and editorial handling. Together with our community of editors and reviewers, these tools form the backbone of our new peer-review system, and can also be adopted by individuals and organizations for internal review and validation processes.

## Comments on old abstract:

1) Please revise the title, it is not clear "Peer-Reviewing R Statistical Methods Packages:..." - is it about reviewing statistical methods or statistical packages?
2) Please add a sentence about 'rOpenSci', it is not yet so well-known in the R community.

## Old abstract

*Peer-Reviewing R Statistical Methods Packages: Standards, Process, and Tools*
Mark Padgham and Noam Ross, rOpenSci

> Peer-review of statistical algorithms is challenging and labor intensive, and even software described in peer-reviewed literature rarely has its source code reviewed. Here we describe the extension of rOpenSci's R package peer-review system to support reviewing code implementing statistical algorithms. We have developed a series of standards, guidance for authors and reviewers, and automated tools to facilitate review of statistical code. These include (1) a series of community-developed standards for documentation, testing, code architecture and programming APIs, of packages, with specialized sub-standards for specific areas such as regression, clustering and Bayesian algorithms, (2) a series of packages for automated testing and literate programming to document standards compliance as part of code development, and (3) an open peer-review process supported by automated testing, reporting, and editorial handling. Together with our community of editors and reviewers, these tools form the backbone of our new peer-review system, and can also be adopted by individuals and organizations for internal review and validation processes