Thank you for agreeing to serve as a referee for the Annals of Statistics.

The job of a referee

Please carefully and promptly provide advice to the Associate Editor as to whether or not the paper should be published. The key criteria against which you should judge the paper are:

  1. Interest and importance and novelty as a scientific contribution.Papers should be novel and original in their field and should make an interesting contribution; there is inevitably an element of subjectivity in judging the interest of a paper, and we are confident that our referees have the ability to do this. Part of the case for the importance of a paper is its relation to the existing literature. Technical sophistication alone is not sufficient grounds for publication in the Annals of Statistics.
  2. Quality of writing and presentation. Papers should be well written and their material and achievements clearly presented. They should make proper reference to the existing literature and should have an accurate and appropriate bibliography. The introduction and abstract should give a good entry point into the paper. In judging quality of writing and presentation, please do not be concerned about minor linguistic errors of the kind likely to be incurred by authors whose first language is not English. We do not want to discriminate against such authors, and, in any case, all papers that are ultimately accepted are then copy-edited to correct linguistic errors.
  3. Technical correctness. Please do your best to check that the arguments of the paper are technically correct but remember that, ultimately, the responsibility for this rests with the authors.

The Annals of Statistics makes it a goal to publish new and interesting work in a timely fashion. Thus, submitting your report within one month of receiving the paper will be especially appreciated by the editorial board and the authors. In any case, if you find that you will not be able to submit a report within two months, please notify the Associate Editor immediately. In this latter case, any comments you can make, particularly regarding other possible referees, will be appreciated.

The refereeing system is anonymous, so please do not disclose your own identity as a referee or the identity of the Associate Editor. Manuscripts should be regarded as privileged and confidential.

Presenting your report

Your views and findings should be presented in two parts, an anonymous report suitable for sending to the authors, and a recommendation and confidential comments for the Editors, which will be seen only by the Associate Editor and the Editors.

The report for the authors should contain a reasonable rationale for your view of the paper as judged against the criteria above. It should be written constructively and politely but should not state your recommendation explicitly. Although it will be useful to provide feedback to help the authors make improvements in the manuscript, you are not expected to rewrite the paper or to suggest major avenues for further research.

The recommendation to the Editors must be one of the following:

  1. Accept (possibly subject to minor corrections). In this case author may be asked to make some corrections or very minor revisions (set out in your comments for the author) prior to sending their final manuscript to the publisher. These should be of a nature that does not need to be checked again by an Editor or referee before the final version is accepted.
  2. Requires minor revision. You should make this recommendation if you can see that the paper needs some revision before publication, but you are confident that if the revisions are carried out correctly then the paper will be acceptable. The revisions may need to address presentational shortcomings, to make additional reference to the literature, to correct minor errors, to conduct very straightforward additional work, and so on, but overall they should not entail more than a week’s work for the author. The revised version will need to be checked by the Associate Editor before the Editor’s final decision is made. Please ensure that your comments for the author make clear what revisions are needed.
  3. Requires major revision. Make this recommendation if you think that the paper can be accepted once all the issues clearly indicated in your referee report are resolved satisfactorily, but the amount of work needed is considerably more than for minor revision. The changes may involve fixing errors in the proofs and making substantial improvements in the presentation. The standard time for this category of revision is two months. The revised manuscript will be checked by all the referees.
  4. Reject with resubmission. Make this recommendation if you think that the paper contains interesting material but the contribution is either incomplete or contains major deficiencies (e.g. exhibiting too vague formulations or too narrow focus, or gaps in statistical theory) and it is unclear whether the author will be successful in achieving the required results for publication. In this case, the author has the option to resubmit the paper.
  5. Reject. This decision discourages the author from resubmitting the paper even if it is revised. Please ensure that, taken together, your report for the author and any confidential comments for the Editor provide an explicit rationale for your recommendation.

Most published papers will not exceed 30 pages. Please provide in addition a recommendation whether part of the paper should be placed in the Supplementary Material archive of the Annals of Statistics.

Electronic submission of reports

Please submit your reports electronically through the EJMS system. The report for the authors should be uploaded to the system, preferably as a pdf file. The system will require you to select a recommendation for the Editors and will also allow you, if you wish, to include further confidential comments. You may need to register at the EJMS web site https://www.e-publications.org/ims/submission/.

For confidentiality issues please see Confidentiality and Electronic Documents.

Thanks

The success of the Annals of Statistics depends on the willingness of referees to take the time to ensure the quality and importance of the papers published. We are extremely grateful for your assistance.