Reviewer Guidelines

The journal depends on the judgement of its reviewers and takes seriously the time that reviewing costs. These guidelines explain what the journal asks of a reviewer and what a useful review looks like.

Before Accepting an Invitation

Accept only if all three are true:

•      The manuscript falls within your expertise. It is fine — and helpful — to say which parts you can and cannot judge.

•      You have no competing interest: no collaboration with the likely authors in recent years, no personal or financial stake in the work, building, practice or product discussed, and no strong prior dispute with the line of research.

•      You can deliver the review within the agreed period, normally four weeks. If you cannot, decline promptly — a fast no is far more useful than a slow yes.

If you suspect you know who the authors are, tell the editor and let the editor decide whether you should proceed.

Confidentiality

The manuscript is a confidential document. Do not share it, discuss it, cite it, or use its content in your own work. Do not upload the manuscript, in whole or in part, to any generative artificial intelligence tool or external service — this is a breach of confidentiality and of the journal’s policy. The review must be your own work. Delete your copies when the review is complete.

What to Assess

•      The research question — is one stated, and is it worth asking?

•      The method — is it explicit, appropriate to the question, and described in enough detail to be examined? Software-based methods should name the software and version.

•      The evidence — do the drawings, photographs, data or fieldwork actually support the claims? In this journal, drawn and photographic evidence carries the same weight as numerical evidence and deserves the same scrutiny.

•      The argument — does the conclusion follow from what is shown, and are limitations acknowledged?

•      The literature — is relevant prior work engaged fairly? Recommend additional references only where they are genuinely necessary, and never primarily your own.

•      Research ethics — where the work involves human participants or culturally significant sites, are approval, consent and permissions adequately described?

Please also tell the editor — in the confidential comments — if you notice substantial overlap with published work, signs that figures have been manipulated, or indications that the manuscript was prepared with generative artificial intelligence.

What a Useful Review Looks Like

•      It is specific. “Section 3 does not explain how the sample of buildings was selected” can be acted on; “the method is weak” cannot.

•      It is addressed to the work, not the authors. Assume competence and good faith; many of the journal’s authors are early-career researchers and the review may shape how they work for years.

•      It distinguishes the essential from the desirable. Separate what must change for the manuscript to be publishable from what would merely improve it.

•      It is honest about the manuscript’s strengths. Saying what works is information the editor needs, not politeness.

A review of two focused pages is worth more than ten pages of line edits. Copy-editing is not the reviewer’s job; note that the language needs attention and move on.

Recommendations

Reviewers recommend one of: accept, minor revisions, major revisions, or decline. The recommendation is advice — the handling editor makes the decision, weighing all reviews. Where reviews disagree, the editor may seek a further opinion.

After the Review

Reviewers may see the other reviews of the same manuscript once a decision is made, where the system permits. Reviewers who wish to have their contribution recorded may claim their reviews on ORCID; the journal supports this on request.