1. Introduction to PCI and PCI Network Science

PCI Network Science is a community of recommenders, playing the role of editors, who recommend unpublished articles based on peer reviews, thereby converting them into complete, reliable and citable articles, without the need for publication in ‘traditional’ journals. Evaluation and recommendation by PCI Network Science are free of charge. When recommenders decide to recommend an article, they write a recommendation text that is published along with all the editorial correspondence (reviews, recommender's decisions, authors’ replies) by PCI Network Science. The article itself is not published by PCI Network Science; it remains on the preprint server on which it was posted by the authors. PCI Network Science recommenders can also recommend postprints, but this is much less frequent.

PCI Network Science is a community of the parent project Peer Community In, an original idea of Denis Bourguet, Benoit Facon and Thomas Guillemaud.

PCI Network Science is not designed to be a free peer reviewing service for authors aiming to improve their articles before submission to a journal, although, of course, it remains possible for authors to submit their recommended article to a traditional journal.

• PCI Network Science is stimulating: it recommends remarkable articles.
• PCI Network Science is free: there are no fees associated with the evaluation process, and no charge for access to the comments and recommendations. The website is freely accessible.
• PCI Network Science is transparent: reviews and recommendations (for unpublished articles) and recommendations (for published articles) are freely available for consultation. Recommendations are signed by the recommenders. Reviews may also be signed if the reviewers agree to do so.
• PCI Network Science is not exclusive: an article may be recommended by different Peer Communities in X (a feature of particular interest for articles relating to multidisciplinary studies) and may even be published in a traditional journal (although this is not the goal of PCI Network Science).

2. Submission requirements, review policy and workflow

2.1 Scope

No one can deny that we now live in a world whose interconnectedness has increased over time. The analysis of these networks and their complexity has thus become an important issue in understanding the world. In this context, research on complex networks, such as promoted by PCI Network Science, is becoming more important than ever. Understanding the complex networks that exist or that we are building, the local/global and multiscale interactions, requires both a multidisciplinary approach to share tools, methods and concepts, and interdisciplinary approaches that achieve innovative scientific integration. PCI Network Science covers both theoretical and applied research related to complex networks: social networks, technological networks, historical networks, ecological networks, biological networks, animal networks, … PCI Network Science considers both quantitative and qualitative methods as long as they are related to network analysis.

PCI Network Science recommends only preprints of high scientific quality that are methodologically and ethically sound. To this end, PCI Network Science: 
- Requires data, computer codes and mathematical and statistical analysis scripts to be made available to reviewers and recommenders at the time of submission and to readers after recommendation. 
- Welcomes reproductions of studies. 
- Welcomes preprint submissions based on preregistrations (with or without prior review) 
- Welcomes preprints reporting negative results, provided that the questions addressed and the methodology are sound. 
- Does not accept submissions of preprints presenting financial conflicts of interest. Other conflicts of interest must be minimal and declared. 
- Ensures that, as far as possible, the recommenders and referees have no conflict of interest with the content or authors of the study evaluated. 

2.2 Types of articles

The articles recommended may have diverse formats: reviews, comments, opinion papers, research articles, data papers, technical notes, computer notes, etc. No editing, formatting or proofs of the recommended papers are required, but we provide the authors with a template to help them to format their article, if they so wish. We also ask the authors of recommended articles to add a sentence to the acknowledgements stating that their article has been recommended by PCI Network Science.    

PCI Network Science welcomes anonymous submissions (see details). 

Preregistrations should be submitted to PCI RR. 

2.3 Repeatability of science and open science

PCI wishes to promote scientific repeatability and reliability, to improve the overall robustness and integrity of the scientific conclusions drawn. To this end, PCI has established three mandatory rules and makes two suggestions to authors: 

Mandatory rules:

Articles recommended by PCI must provide the readers with: 

-Raw data, made available either in the text or through an open data repository, such as Zenodo, Dryad or another institutional repository (see Directory of Open Access Repositories) with a DOI. Data must be reusable, so the metadata and accompanying text must describe the data carefully and accurately. 

-Details concerning the quantitative analyses (e.g. data treatment and statistical scripts in R, bioinformatic pipeline scripts, etc.) and simulations (scripts, codes), available in the text or through a correctly versioned deposit in an open repository, such as Zenodo, Dryad or some other institutional repository (see Directory of Open Access Repositories) with a DOI or another permanent identifier (such as a SWHID of Software heritage). Note that Git URLs are not permanent. Information on how to issue a DOI for a GitHub repository is given at https://docs.github.com/en/repositories/archiving-a-github-repository/referencing-and-citing-content. Scripts or codes must be carefully described such that another researcher can reuse them.

-Details on experimental procedures must be given in the text.

Suggestions to authors:

-PCI encourages authors to submit preprints based on preregistrations. Authors may post their research questions and analysis plan to an independent registry before observing the research outcomes, and, thus, before writing and submitting their article. This provides a way of clarifying hypotheses, avoid confusing “postdictions” and predictions, and carefully planning appropriate statistical processing of the data (e.g. 10.1073/pnas.1708274114). 

-Preregistrations should be submitted to PCI RR

-PCI welcomes submissions proposing replication studies. All submissions are assessed according to the same criteria, provided that the article is considered interesting by the recommender handling it and the research question is considered scientifically valid.

2.4 Ethics approval

In fields in which research requires approval from an ethics committee or institutional review board, the authors should generally ensure that the proposed research has received all necessary approvals before submission.

2.5 Transparency for data and materials

In general, authors are required to make all study data, digital materials, and computer code publicly available (at submission), to the maximum extent permissible by the relevant legal or ethical restrictions.

2.6 Transparent review

PCI Network Science publishes all reviews of recommended manuscripts, with reviewers retaining the right to choose whether to sign their reviews or remain anonymous. All reviews and recommender decision letters are published on the PCI Network Science platform on recommendation. Reviews of rejected submissions are sent to the authors, but are not published. Reviewers and recommenders are expected to adhere to the PCI code of conduct, avoiding abusive or discriminatory language in their comments. Reviews considered to violate the code of conduct may be edited by recommenders or the Managing Board, returned to the reviewer for editing, or discarded.

2.7 Handling of manuscript by recommenders

A recommender finding a submitted preprint particularly interesting can decide to initiate the process of evaluation for the article concerned. In this case, the authors are notified by e-mail. The recommender invites reviewers, so as to obtain at least two high-quality reviews. Note that the recommender and reviewers must declare that they have no conflict of interest of any kind with the content or the authors of the preprint – see the code of conduct.

If no recommender – including all recommenders in the corresponding field of expertise – has initiated the evaluation of the preprint after 20 days, the authors are notified by e-mail. If this happens, we suggest that the authors cancel their submission.

2.8 Appeals process

The authors of rejected manuscripts can appeal the PCI Network Science decision within 30 days of receiving the decision, by contacting us at contact@networksci.peercommunityin.org. Appeals will be considered by the specialist recommender(s) handing the manuscript and the PCI Network Science Managing Board. Decisions following appeal are final. 

2.9 Word limits and formatting requirements

PCI Network Science imposes no word limits or specific formatting requirements on submissions. Manuscripts should be as concise as possible, but as long as necessary to ensure that the description of the rationale and methods is clear and comprehensive, and that all methods are reproducible. Authors should ensure that they follow established formatting conventions for articles in their discipline. Authors who intend to publish their article in a journal should also take note of any sectional or overall word limits or other formatting requirements that apply to specific journals.

3. Top tips for reviewers

The reviewing process for an article submitted to a PCI is very similar to that for an article submitted to a journal, but with one key difference: peer review process for PCI is transparent, unlike that of journals. The review history is, therefore published by PCI if the article is recommended.

The following guidelines are derived from various services and organizations cited in the section “3.4 Sources”. They comply with the Ethical Guidelines for Peer Reviewers published by the Committee on Publication Ethics (COPE), to which PCI belongs

3.1 General advice

Be constructive, kind and respectful.

Be precise and clear.

Remain open to new approaches.     

If you lack expertise on certain points, state it in your review.

Only agree to review manuscripts that you can assess in a timely manner.

3.2 What you should NOT DO

You should not have a conflict of interest with the authors or with the content of the article (cf the code of conduct). If you have a conflict of interest, you must decline the invitation. 

You do not need to determine whether the article falls within the scope of the PCI. Once a submission has been validated by the managing board, its scope is considered suitable for that PCI.

You do not need to check that scripts, simulation code, software parameters, etc. are available to the reader (e.g. repository link/DOI or appendix); this is a prerequisite for submission to PCI.

You do not need to comment extensively on typographical errors, spelling and grammar (but you can if you want). If language is poor, just mention it, and suggest editing from someone with full professional proficiency in English.

You do not need to check for potential plagiarism. All the articles submitted to PCI are checked with the Ithenticate tool.

You do not need to make recommendations about acceptance or rejection – this is the role of the recommender (the equivalent of an associate editor at a traditional journal) handling this preprint.

You do not need to check the format of references (but you can if you want).

Do not discredit negative results.

Do not ask for new experiments if the study is already well performed and the conclusions are clear enough. You can suggest that the authors perform further experiments, but only if they are really necessary to boost confidence in the interpretation of the results.

Avoid unconstructive, ambiguous, and unsupported comments.

3.3 What you should DO

In general

Promptly accept or decline invitations, to keep the time-to-decision short.

Try to keep to the deadlines: this shows respect towards the authors. Post your review within three weeks after of accepting the invitation. In you anticipate a delay, please inform the PCI recommender and managers.

Check carefully that you do not have any conflict of interest with the content or with the authors of the article. If so, you must not review the article. If you are unsure whether or not you have a conflict of interest, ask the PCI managers via en e-mail sent to contact@networksci.peercommunityin.org.

Provide a detailed, objective report on the merits of the preprint.

Identify flaws (if any) in the design of the research, and in the analysis and interpretation of results.

Expose your concerns (if any) about ethics or scientific misconduct.

State the preprint’s strengths as well as its weaknesses.

Try to consider both the technical merit and the scientific significance of the preprint.

If there is something critically missing, report it. 

Provide specific suggestions for improvements.

Evaluation of the different components of the article

Title

Check that the title clearly reflects the content of the article.

Abstract

Check that the abstract is concise and presents the main findings of the study.

Introduction

Check that the introduction clearly explains the motivation for the study.
Check that the research question/hypothesis/prediction is clearly presented.
Check that the introduction builds on relevant recent and past research performed in the field.

Materials and methods

More generally, check that sufficient details are provided for the methods and analysis to allow replication by other researchers.
Check that the statistical analyses are appropriate.

Results

If possible, evaluate the consistency of raw data and scripts.
If necessary, and if you can, run the data transformations and statistical analyses and check that you get the same results.
In the case of negative results, check that there is a statistical power analysis (or an adequate Bayesian analysis).
Inform the recommender and the managing board if you suspect scientific misconduct.

Tables and figures

Check that figures and tables are understandable without reference to the main body of the article.
Check that figures and tables have a proper caption.

Discussion

Check that the conclusions are adequately supported by the results and that the interpretation of the analysis is not overstated.
Check that the discussion takes account of relevant recent and past research performed in the field.

References

Check that all references are appropriate and that the necessary references are present.
Report any reference cited in the text that does not appear in the reference list.

Peer review of the revised version of a preprint

Evaluate how the authors addressed your comments.
If you have new requests for modifications, please state them clearly. But, in general, try not to start new lines of interrogation if the previous points were addressed adequately.
If you disagree with the other reviewer(s) and agree with the authors' explanation and defense of their original article, help the recommender out by discussing these issues.

3.4 Sources

https://publicationethics.org/files/Peer%20review%20guidelines.pdf

https://publicationethics.org/files/cope-ethical-guidelines-peer-reviewers-v2_0.pdf

Fossils and shit: so you want to do a peer review

https://www.timeshighereducation.com/news/career-advice-how-peer-review-paper#

Plos: how to write a peer review

https://f1000research.com/for-referees/guidelines

https://www.sae.org/publications/journals/reviewers/guide

https://www.britishecologicalsociety.org/wp-content/uploads/Publ_Peer-Review-Booklet.pdf

EASE: how to serve as an effective referee

http://freerangestats.info/blog/2020/06/13/publication-reform

4. The evaluation and recommendation process

1. A PCI Network Science recommender may invite you to review a preprint. If so, you will receive an alert by e-mail containing a link to the PCI Network Science website to allow you to decline or accept this invitation. If you are willing to review the preprint, you will first have to certify that you have no conflict of interest of any kind with the content or with the authors of the preprint.
2. Once you have agreed to review a preprint, you are expected to write your review within three weeks. The format of reviews is similar to that for ‘traditional’ journals. Please bear in mind that your reviews will be published if the article is recommended.
3. Once you have completed your review, you can copy/paste or upload it onto the PCI Network Science website. At this stage, you can indicate whether or not you wish to remain anonymous. 
4. Based on the various reviews obtained, the recommender must come to a decision within about 10 days. This decision and all the reviews are sent to the authors.
5. The recommender can decide to ask the authors to revise their article. The recommender may therefore subsequently contact you again to ask you to evaluate a revised version of the article and the author’s replies to your comments and those of the other reviewers.
6. If, after one or several rounds of review, the recommender eventually decides to reject the article, your review(s) will NOT be published or publicly released by PCI Network Science. They will be safely stored in our database, to which only the Managing Board has access.
7. Alternatively, if the recommender eventually decides to recommend the article, all the editorial correspondence (decisions, authors’ replies, all reviews, including yours) will be published by PCI Network Science, to make the evaluation process transparent.