This issue of the International Journal of Sport Nutrition and Metabolism (IJSNEM) includes a letter to the editor (Tiller, 2022) that expresses concern regarding a recently published paper by Medeiros et al. (2022). The expression of concern primarily relates to the maximal oxygen uptake ($\dot{\mathrm{V}}{\mathrm{O}}_2\text{max}$) data in the study. Specifically, our reader noticed that the magnitude of change in absolute $\dot{\mathrm{V}}{\mathrm{O}}_2\text{max}$ (L·min⁻¹) after 60 days of vitamin D supplementation (i.e. 28% increase) is both physiologically implausible when considering the known responsivity of that parameter to various interventions and, moreover, is mathematically incompatible with the completely stable relative $\dot{\mathrm{V}}{\mathrm{O}}_2\text{max}$ (mL·kg⁻¹·min⁻¹) values reported in the same paper (unless body mass changed by circa 18 kg).

The authors of the original article were asked to provide an explanation for the apparent inconsistencies, along with their source data, in the hope that an erratum might clarify and/or correct any issues. No coherent explanation or adequate response to the expression of concern could be obtained from the authors, although various revised data files were made available. In summary, those revised data still included unprecedented changes in absolute $\dot{\mathrm{V}}{\mathrm{O}}_2\text{max}$ (L·min⁻¹) at both a group and especially an individual level, with no logical reason to account for that variance; whereas the authors now acknowledge that some of the relative $\dot{\mathrm{V}}{\mathrm{O}}_2\text{max}$ (mL·kg⁻¹·min⁻¹) data in their original manuscript were incorrect (it is uncertain how these errors occurred). Unfortunately, it appears that almost one-third of the original measurements have been lost and so it is now impossible to recover relative $\dot{\mathrm{V}}{\mathrm{O}}_2\text{max}$ data for the complete study population (n.b. relative data cannot be retrospectively derived from absolute values since the authors claim not to have recorded body mass).

It is therefore necessary to retract the original article on the basis that the originally published data are now known to be inaccurate and a complete set of corrected data are not available. Interested readers can consider for themselves the various other issues identified in the associated expression of concern (Tiller, 2022), including the puzzling effect of vitamin D on strength in only one arm (but not the other) and the mis-citation of key references.

In terms of lessons to be learnt from this instance, it is firstly regrettable that these particular issues with the data were not identified during the peer-review process, for which I take primary responsibility as Editor-in-Chief. We are therefore refining our processes at IJSNEM with immediate effect, such that authors will increasingly be asked to provide source data files where relevant when responding to reviewer/editorial comments. Furthermore, authors are encouraged to adhere to best practice for data management (Nightingale, 2020) and certainly to retain source data (ideally in a public repository). One practical suggestion would be for some/all co-authors to be provided with an organised source data file in order to verify all summary statistics reported in the manuscript before it is submitted; the number of authors included on scientific papers in the biomedical sciences continues to increase (i.e. “hyperauthorship”; Cronin, 2001), so there should be no shortage of individuals keen to take responsibility for the data/paper at the submission stage.