Corrections
If something here is wrong, we want to know. A product whose promise is accuracy about evidence has to be correctable.
NatureHQ has no working channel for reports yet, and this page will not pretend otherwise. There is deliberately no form here, because a form that quietly discards what you type — or that appears to file a report nobody receives — would be worse than none at all. There is no published address either, for the same reason: an address that nobody reads is a form with extra steps.
This is a real gap and it is recorded as one. A site whose promise is accuracy about evidence has to be correctable, and NatureHQ is not yet correctable by its readers. When a channel exists it will be named on this page, in the footer, and in privacy — not in one of them.
What is already correctable is the research itself. Every study NatureHQ cites is reachable at its publisher through the links below and on each study page, and a post-publication correction recorded against a paper is shown here as soon as the record is updated. If you find that a summary here misreads a paper, the paper is the place that argument can be had in public in the meantime.
When there is somewhere to send them, the most useful reports will be:
- Factual errors — a wrong figure, date, species or scientific name.
- Misread research — where our summary of a study does not match what the study actually found. This is the most valuable kind of report we can receive.
- Overstatement — where a caveat is missing, or a finding in one species has been written as though it applied to a whole group.
- Broken or wrong references — a DOI pointing at the wrong paper, or a citation that cannot be located.
- Taxonomy — outdated names, wrong ranks, or species confused with one another.
- Accessibility problems — see the accessibility statement.
Please include the page address and, where you can, a source. A correction with a citation can be acted on immediately; one without needs research first.
Where a claim changes materially, the change is recorded rather than applied silently. NatureHQ keeps superseded readings visible: when later research reframes an earlier finding, both stay on the site with the relationship between them stated. The same principle applies to our own mistakes.
Every claim carries the date it was last reviewed, and every study says whether its reference details have been confirmed against an independent catalogue.
Corrections in the research we cite
Retractions, author corrections, failed replications and contested reanalyses recorded against the studies behind our claims.
- Numerical cognition in honeybees enables addition and subtraction
- contested reanalysis — Several researchers published critiques arguing that the training procedure and stimulus design leave simpler explanations open, and that the statistical case for arithmetic is weaker than presented.
- Observational learning in Octopus vulgaris
- failed replication — Subsequent work has not consistently reproduced observational learning in octopuses, and the original result is not treated as settled within the field.
- Positive citation bias and overinterpreted results lead to misinformation on common mycorrhizal networks in forests
- correction — An author correction was issued in the same journal later in 2023. It amends the published record without altering the review’s conclusions; NatureHQ records it because a correction a reader cannot see is a correction they cannot weigh.
- Shaping of hooks in New Caledonian crows
- contested reanalysis — Subsequent fieldwork showed that hook manufacture is widespread in wild New Caledonian crows, so the original "spontaneous invention" reading overstated the case.
- The global tree restoration potential
- contested reanalysis — Multiple technical comments in the same journal challenged the restoration area, the carbon accounting and the classification of grassy biomes as restorable; the authors published a correction to the carbon estimate.