ACL Conference Conflict-of-interest policy

Conference Handbook - ACL Conference Conflicts-of-interest Policy

Conference Handbook - ACL Conference Conflict-of-interest Policy

To ensure fair and unbiased evaluation of conference submissions, it is crucial to avoid conflicts of interest throughout the paper review process, as described below.

Definition of a Conflict of Interest:

A person has a conflict of interest with a submitted paper if that person:

1. is a co-author of the paper, or
2. has been a student or supervisor of one of the authors in the previous five years, or
3. has co-authored a paper or collaborated with one of the authors in the previous five years, or
4. is employed at the same company or institution as an author, or
5. has any other circumstances that could cause a bias in evaluating the paper.

ACL Policy on Conflicts of Interest:

To avoid conflicts of interest, participants at all levels -- reviewers, area chairs, and program chairs -- must:

• Disclose potential conflicts of interest: All participants (reviewers, area chairs, and program chairs) are responsible to determine and disclose potential conflicts of interest as soon as they have access to the conference submissions.
• Restrict decision making: No person shall review or participate in the acceptance decision of any paper for which they have a conflict of interest.
• Protect reviewer identities: The identity of the reviewers of a conflicted paper shall be withheld from all people who have a conflict of interest with that paper.

Recommended Procedural Guidelines:

As the community has grown increasingly collaborative, it has become a challenge to organize the reviewing and decision procedures to strictly enforce the three criteria listed above for all types of conflicts of interest (COIs). The recommended procedural guidelines below, updated in May 2016, are current best practices given the recent community size and the conference management software. With the consent of the ACL Executive Board, future program chairs are empowered to add to and improve these best practices to better support the ACL policy on conflicts of interest.

Definition of Senior Program Committee (SPC):

In what follows, SPC refers to people who are in charge of review assignment and management. Thus, SPC refers to area chairs for large conferences such as ACL, NAACL, EMNLP, and SPC refers to program chairs for workshops or small conferences where program chairs handle review assignment directly without area chairs.

Handling of a Reviewer's COI:

• The SPCs (area chairs or program chairs, as defined above) must not assign a paper to a reviewer who has a conflict of interest.
• If a reviewer determines that they may have been inadvertently assigned to a paper with a conflict of interest, they must immediately disclose the fact to the SPC. If there is indeed a conflict of interest, the SPC will reassign the paper to another reviewer. It is the reviewer’s responsibility to examine potential conflicts of interests as soon as possible after the papers are assigned to them.
• The SPCs must ensure that the reviewers know they must examine the potential conflict of interest as soon as possible after the papers have been assigned to them.

Handling of an area chair's COI:

The following guidelines are applicable to conferences that operate with area chairs.

• Papers co-authored by area chairs must be assigned to areas other than the one they chair. If no adequate area is available, the program chair must create a special COI track and directly manage the reviewing and acceptance decisions.
• If an area chair has a COI with a paper (but is not an author), there are two options:
  • Option A: the paper will be assigned to another area. If no adequate area is available, the program chair will create a special COI track and handle the management of reviewing and acceptance directly. This is the desired option as it completely separates the area chair and any co-area chairs from the decision making. However, as the community has grown with increasing collaboration, a substantial portion of the papers might need to be re-routed to the special COI track. Therefore, Option B is also acceptable.
  • Option B: the paper may remain in the area provided that at least one other area co-chair does not have a COI with that paper. If the paper remains in the area, the conflicted co-chair(s) must not have access to the reviewer assignment and the review content. (This effect can be achieved by setting the COI flag in the START conference management system.) The conflicted co-chair(s) must also withdraw from all aspects of decision making for that paper.

Handling of a Program Chair's COI:

• If a program co-chair has a conflict of interest in a paper, a different co-chair must handle the submission. If there is only a single program chair, the general chair of the conference shall assign it.
• It is strongly recommended that the program co-chairs and their students seek alternative venues for publications, e.g. TACL, where the paper can still be presented at the respective conference without a COI. In justified cases, the program co-chairs for conferences may seek agreement with the GC of their respective conference and still submit their papers to the conference that they are chairing, e.g. through ARR-based reviewing and committing to the conference. In any case, the program co-chairs are responsible that the reviewing of their papers is managed by a co-chair without a COI and they don’t have access to any information about their paper. The general or local chairs need not seek alternative venues provided that they are not sharing the role of program chairs to oversee reviewing and making decisions on the conference submissions. The program chairs of workshops may submit their papers provided that there will be another program chair without COI who can handle those submissions.

Change of Authorship and COI:

• To avoid the potential loophole with the COI management, each paper must declare the complete author list at the time of submission. Adding more authors or switching the authors should be in general disallowed once the paper has been reviewed and accepted.

The reviewers, area chairs, and program chairs will also take care to identify and resolve potentially ambiguous cases under these procedural guidelines given the general guidelines. The ACL executive board may be contacted to discuss specific cases.

(Last Update: Barbara Di Eugenio, Nov. 2020)