The envisioned general strategies are appended below. These guidelines provide a few additional pointers for these strategies. Authors are required to adhere to a number of policies that are also appended below for your information. Authors will be held accountable to adhere to these policies and AEs are encouraged to follow up with the co-editors if any violation is discovered (e.g., during the review process). This will typically lead to a Reject decision and may entail additional consequences.

Initial Screening of New Submissions

When you receive a new paper, it will have passed the initial screening by the co-editors. We anticipate that about 30% of the initial submissions will be desk-rejected before they reach you.

Once you receive a paper, please immediately check for a potential conflict of interest (COI): There is a strict COI with papers where there is an author who has an affiliation at the same university/institution or who is a partner/relative/housemate and also if there is a financial interest of any kind involved. Inform the co-editor as soon as possible if you collaborate currently or in the recent past with one of the authors or if you have been a mentor or mentee of any of the authors at any time, and also if you had a conflict of any kind with one of the authors. Please be aware that in some cases you may be asked to serve as AE for a paper even when the paper falls outside of your area of expertise, which may happen for example when other potential AEs are conflicted or the paper is outside of the expertise of any of the AEs, especially when it is on a niche topic.

You have up to 10 days to screen the paper. For this screening, please follow the general guidelines below. While the AoS emphasizes mathematical statistics and mathematical depth and new theory is a plus, it is not required and papers should not be rejected with comments such as … “the paper lacks mathematical depth” …. “it would be a better fit for JASA.” The main criterion is statistical innovation with forward-looking new ideas. Papers on topics in probability or with a focus on technical improvements are discouraged, regardless of mathematical depth, unless they are substantial and clearly consequential for the field of statistics.

It is expected that you will reject about 20% of the submissions at this stage, with some variations depending on the area. If you reject, write 1-2 paragraphs providing the main reasons why the paper is unsuitable that can be transmitted to the authors. If you find the ideas intriguing but something seems amiss, you may recommend an immediate Reject with Resubmission, for example if the Supplement is not in good shape (See author certifications below).

Recruiting Referees

Good referees are a valuable resource that should not be overused. In recent years it has become harder to recruit referees with the requisite expertise who are willing to put in the time and effort to write a careful and unbiased report, with clear reasons that support the recommendation. Avoid referees who are conflicted with one of the authors under the criteria listed above, and ideally referees will be from geographic areas that differ from the locations of the affiliations of the authors. If you send a paper out for full review, it may be advisable to request reports from 3-4 referees and to discard unresponsive referees after two weeks, ideally replacing them with new ones. Encourage referees to send negative reviews without much delay, ideally within one month.

Timely Recommendations

If you receive one or two negative reviews and find that you concur with the report(s), do not wait for additional reports to send a negative recommendation. Likewise if you receive two reports with similar recommendations and concur with them, do not wait for additional reviewers to make your recommendation. Most importantly, we aim to arrive at a Reject decision very soon, the sooner the better and definitely within the first 3 months. If referees are not responsive, it is important to send a personal message by the AE (especially as the system generated reminders are ignored by many referees). If in spite of personal reminders and referees remain unresponsive and additional referees cannot be found, it could indicate that a paper is of limited interest (and therefore not suitable for the AoS), on the other hand this may also happen if a paper is highly innovative and contains ideas out of the box. At the 4 months point the AE would need to step in, assess the situation, and write a somewhat more detailed report.

Reject and Resubmit: (see also below) It is advisable to recommend this action only if at the core the paper is innovative, has potential and there is a chance that remaining issues can be addressed, requiring more than a major revision. Reject and resubmit should not be recommended if the basic idea of the paper is of limited interest, the innovation is insufficient or when there are other major flaws, as this will only drag out the process and usually will end with a final reject. Chains of reject-resubmit decisions are to be avoided, and if a paper is resubmitted after such a decision the next step normally would be either acceptance, major or minor revision, or rejection.

Major and Minor Revision: (see also below) If a major or minor revision has been recommended, it will often suffice if the AE checks whether the issues that needed addressing have been satisfactorily resolved, without sending the paper back to referees; when it is necessary to send the paper back to specific referees for a check these referees should be held to a tight timeline. For major or minor revisions, the time to decision should be particularly short. Author as Referee: (Co-) authors of submitted papers will be expected to make themselves available as referees (the submitting author will be required to mark a box in the submission web site to this effect).

Feedback

We encourage any feedback about issues and procedures to the co-editors and especially at the annual AoS AE meeting. Our shared goal will be to strengthen the role of the AoS as the premier journal in statistics.

General Strategies

1. Shortening Review Times
The time to first report on a submitted paper needs to be substantially shortened. For example, it is not acceptable to have a time to first report of 6+ months and then hand down a reject decision. Even less acceptable is a reject after 9+ months, which has happened. To address this will require a multi-faceted approach:

Initial Screening: The initial screening process should be strengthened. We as co-editors will do an initial screening of each submission and record immediate rejects for papers that we deem unsuitable for the AoS. We also expect AEs to carefully screen all submissions before sending them to referees and to make a reject recommendation for less promising submissions within two weeks. For this, an AE report will be needed that can be very brief.

Three Reviews: If you do decide to send a paper for review, it is advisable to request at least three reviews and to ask referees to send a brief report within a month if they consider a paper to be unsuitable. A recommendation to reject can be made by the AE even if not all referee reports have been received.

Timely Decisions: A decision on the recommended action should be reached even if not all referee reports have been received within 4 months. The AE may need to step in and act as additional referee. If not at least two referees can be lined up within the first two weeks after sending the paper out for review, additional referees should be recruited. In some cases, difficulties to recruit reviewers could indicate that the paper is of limited interest overall, and in this case the way forward may be for the AE to recommend a reject decision (accompanied by an AE report).

Conflict Disclosure: It is important that referees (and also AEs) are at arm’s length and disclose possible conflicts (eg., current or recent collaborator or student-mentor relationship).

Sending Reminders: If a reviewer who accepted to evaluate the paper does not send a report within 2 months, a personal reminder by the AE is in order, as system-generated reminders usually are often ignored and less effective. As already mentioned, if some referees remain unresponsive, when reaching the 4 month point an AE should proceed to carefully assess the paper and make a recommendation with accompanying AE report.

Reject and Resubmit: It is advisable to recommend this action only if at the core the paper is innovative, has potential and there is a chance that remaining issues can be addressed, requiring more than a major revision. Reject and resubmit should not be recommended if the basic idea of the paper is of limited interest, the innovation is insufficient or when there are other major flaws, as this will only drag out the process and usually will end with a final reject. Chains of reject-resubmit decisions are to be avoided, and if a paper is resubmitted after such a decision the next step normally would be either acceptance, major or minor revision, or rejection.

Major and Minor Revision: If a major revision has been recommended, it will often suffice if the AE checks whether the issues that needed addressing have been satisfactorily resolved, without sending the paper back to referees; when it is necessary to send the paper back to specific referees for a check these referees should be held to a tight timeline. For major or minor revisions, the time to decision should be particularly short. Author as Referee: (Co-)authors of submitted papers will be expected to make themselves available as referees (the submitting author will be required to mark a box in the submission web site to this effect).

2. Emphasize Statistical Innovation
A key requirement is that a paper contains a promising new idea with anticipated impact for statistical theory, methodology, data analysis, applications, computing or statistical learning. Mathematical depth is neither necessary nor sufficient. If the impact is similar, a clean theory with simpler proofs is better than a contrived theory with complex proofs. Papers that emphasize technical improvements of a specialized nature and without evidence for clear impact find a better home in more specialized journals, and similarly papers emphasizing probability in probability journals. A good example of an innovative paper with little mathematical sophistication is ”Projection Pursuit” by Peter Huber, Annals of Statistics 2985, 435-475, which had immense impact. Authors will be asked to state the key innovations in both cover letter and abstract/introduction so that readers (and reviewers) can grasp these quickly.

3. Publish the Best Work in Statistics
The AoS aims to publish the best innovative work in all areas of statistics and statistical aspects of machine learning. Mathematical/theoretical investigations will continue to play a major but not exclusive role. Papers of outstanding quality and innovation with applied and non-mathematical emphasis should also find a home in the AoS, reflecting a broad perspective of the best work in statistics. The IMS mission statement for the AoS provides guidance and is as follows.