Policies

Submissions are handled anonymously in both directions. Operations holds author identities and is the journal’s only channel to authors. Editors and referees work from a paper ID; the editorial line learns an author’s identity at publication. Authors are never told who refereed their paper.

Authors disclose at submission which AI tools were used and for what. Language editing and coding assistance are acceptable with disclosure. Text, analysis or results generated by an AI tool and presented as the author’s own are not; undisclosed substantive use is grounds for rejection at any stage.

A paper whose argument rests on data analysis is published with the data and code that produce its tables and figures: the data or access instructions, every script from data to results, and a README with software versions and run order, deposited with a persistent identifier. The Data Editor checks every accepted empirical paper before production. Exceptions for confidential or licensed data are recorded and disclosed.

Anyone who recognises, or believes they recognise, the author of a paper they are handling says so at once and steps aside. Members take no part in handling their own submission. Nobody on the sponsorship side takes part in any editorial decision.

A correction changes a published paper without altering its conclusions and is published as a dated note. A retraction withdraws a paper whose conclusions cannot be relied on, after the author has been heard, with a notice stating the reason; the paper stays in the archive, marked as retracted.

An author may appeal a rejection once, within fourteen days, on procedural error, a demonstrable misreading on a point of fact, data or method, or an undeclared conflict of interest. Disagreement with the editors’ judgement of the contribution is not a ground.