Lades et al. (2020)

DECISION: Moderate Errors

that do not affect the core conclusions

Reviewer: Joop Adema, LMU Munich
Author response: Leonhard Lades, University of Stirling
Recommender: Ian Hussey, University of Bern

Decision & Recommendation

Lades, Laffan, Daly, & Delaney (2020) “Daily emotional well-being during the COVID-19 pandemic” was determined to contain Moderate Errors that do not affect the core conclusions. That is, errors that have the benefit of being detectable thanks to the presence and sharing of research materials, whose scope is moderate given that a large proportion of the reported results are affected (likely to a small degree), but which do not affect the article’s core conclusions. A determination of Moderate Errors involves the recommendation that the authors seek a correction. Following the ERROR project’s emergent guidelines, the recommendations associated with a Moderate Errors decision are as follows:

• The report, author response, and recommendation have been posted on the ERROR website (error.reviews/reviews/lades-2020) and as a preprint on PsyArXiv (osf.io/preprints/psyarxiv/e3c7p). Their associated materials have been posted to OSF (see osf.io/sdz5y).
• The authors are asked to recognize these errors in future discussions of the article.
• The authors are recommended to pursue a correction notice with the journal. Specifically, with regard to:
  • The text on p. 905 states that multiple testing corrections were used. However, corrections were in fact only applied to results in the supplementary materials and not any results reported in the article itself. 11 of 42 (26.2%) statistically significant results reported in the article were not significant after multiple testing corrections. Although, these do not alter the main conclusions. The use of multiple testing corrections should either (a) more clearly state that these corrections were applied only to the results reported in the supplementary materials and not any results reported in the article, or (b) all results reported in text should be those with the multiple testing corrections applied.
  • The primary analyses reported in the article did not acknowledge dependencies in the data (i.e., each participant completed the survey multiple times). The authors agreed with the reviewer that they should have employed clustered Standard Errors to acknowledge these dependencies. Applying clustered Standard Errors changes the statistical significance of one result (positive affect ~ schooling children), although this does not change the core conclusions. Corrections should be made to the Standard Errors reported in text, the Confidence Intervals in all panels of Figure 1 and Figure S1, and the results reported in Table S2. The “†” significance bar should be removed from the positive affect ~ schooling children result in Table S2.
  • All analyses and summary statistics should use corrected data that use unique participant IDs. A small number of duplicate participant IDs were found by the reviewer, which the authors argue are distinct participants but are not treated as such in the reported analysis. Correcting this error may produce additional changes in the results of the regressions reported in the text, tables, and plots. These magnitude of any such changes is currently unknown but likely to be very small.

Cite as: ERROR. (2025). Error review of Lades et al. (2020). Version 1. https://osf.io/preprints/psyarxiv/e3c7p

Full Report

Materials

PubPeer Page