EDITORIAL POLICY

Science of Rowing exists to put rowing science in front of the people who coach and row. That only works if you can trust what you read here. This page sets out how our articles are made and who is accountable for them.

Every article is reviewed before it is published

Nothing goes live on the strength of one person's opinion. Every issue carries three articles: one from Alex Wolf, one from Joe DeLeo, and one from a guest author. Each article is read closely by the other two contributors before it publishes, so every piece is examined by two people with subject knowledge who did not write it.

One of the two editors, Alex Wolf or Joe DeLeo, then records the sign-off. That sign-off confirms the review took place and that the article is fit to publish. It is logged with the editor's name and the time, both taken from the account that signed it rather than typed, and it is completed no later than 14 days before the publication date. An article does not go live without it.

This is closer to academic peer review than to a typical blog, and that is deliberate. If we get something wrong, we would rather find it before you do.

Who is responsible

Alex Wolf and Joe DeLeo hold joint editorial responsibility for everything published on this site. That means final say on what runs, responsibility for the review process itself, and accountability when something needs correcting.

If you think we have got something wrong, tell us at alex@scienceofrowing.com. We would rather hear it than not.

How we use AI

We use AI tools in our work, the same as most publishers now do. Here is exactly where and how.

Writing

AI may assist with drafting, structuring or editing. It never replaces the review process described above. Every article is examined on its substance by two people with subject expertise who did not write it, and a named editor is accountable for it.

Data and charts

Every chart, plot and figure on this site is generated from real data. We do not generate data, and we do not illustrate a point with numbers that were not measured.

Images

Our default is real photography or clearly illustrative graphics such as diagrams and schematics. Where we use an AI-generated image that could be mistaken for a photograph, the label is written on the image itself, so it stays with the picture wherever it is seen. Routine editing of real photographs, cropping, colour correction and the like, is not labelled, because the photograph is still a photograph.

Sorting articles on this site

When you browse the library by research theme, the sorting is done by an AI model. It reads a published article's title and opening lines and picks one theme from a fixed list of nine. It writes nothing and changes nothing, and it only ever sees articles that have already been through the review described above. If it files something under the wrong theme, that is a filing mistake and not a change to what the article says. The provider is named in our Privacy Policy.

Corrections

When we get something wrong, we correct the article and note what changed and when. Substantive corrections are flagged at the top of the piece rather than quietly amended.