Declaration of transparency and scientific rigour: Checklist for Design and Analysis 2022
This checklist for design and analysis provides guidance for transparent reporting and scientific rigour of preclinical research as set out in Experimental design and analysis and their reporting II: Updated and simplified guidance for authors and peer reviewers (Curtis et al., 2022). This checklist is intended as a guide for submission to the British Journal of Pharmacology.
Criteria | Number | Issue | Where to place information |
---|---|---|---|
Experimental design | 1a | Details of prior sample size estimation (e.g., power calculations) should be provided, and, if it was not conducted, a valid scientific justification is provided. | Methods |
1b | The methods declare whether or not randomization was undertaken. If it was not conducted, a valid scientific justification is provided. | Methods | |
1c | The methods declare whether blinding was undertaken, and, if not, a valid scientific justification is provided. | Methods | |
Group sizes | 2a | Group size (n) refers to biological samples and not technical replicates. | Methods |
2b | Inferential statistics (comparisons between groups) are undertaken only if n = 5/group or more. A valid explanation is provided for data with n of less than 5. | Methods/Results | |
2c | The exact group sizes (n) are provided, not a range. The N number of replicates where relevant should be provided as well as the n value for independent experiments. | Methods/Results within figure/table legends | |
Statistical plan | 3a | A data and statistical analysis section is provided giving details of all summary and inferential statistical tests used. | Methods |
3b | Details are provided of any statistical package or programme employed, and details of which tests used in the Data and Statistical Analysis section. | Methods | |
Data and statistical analysis | 4a | If ANOVA is used, a statement is provided indicating that post hoc tests were conducted only if the data are normally distributed, and there is no inhomogeneity of variance. | Methods |
4b |
The presentation and processing of a data set should map to its mathematical distribution. |
Results | |
4c | Any data normalization is explained with a valid scientific justification (i.e., to control for unwanted sources of variation). If normalization generates values with no variance (i.e., control SEM = 0), the data should not be subjected to parametric statistical analysis. | Results | |
4d | ANCOVA is a useful approach that accounts for sources of intervention-unrelated variation and should be used where relevant. | Results | |
Level of probability | 5a | The threshold P value deemed to constitute statistical significance (α) should be defined in the Methods. | Methods/Results |
5b | Authors may elect to depict statistical significance as a single value in the Results and figures usually denoted as *. | Results | |
5c | Authors may elect to report the full value of P, but this approach must be accompanied by a statement in the data and statistical analysis subsection that statistical significance is only stated where P < α. | Methods/Results | |
Outliers/exclusion criteria | 6 | Inclusion and/or exclusion criteria are clearly defined in the methods. | Methods |