Table 1 Examples of automated tools used to screen preprints, submitted papers or publications
Tool | Screening topics and rationale |
|---|---|
Sciscore [18] | Many factors, including: RRIDs: Unique persistent identifiers that allow readers to determine exactly what resource (e.g., cell line, antibody, model organism, software) was used Ethics & consent statements: Required for legal compliance Blinding & randomization: The failure to blind or randomize experiments is associated with overestimated effect sizes Sample size calculations: Provide information about whether the study was designed and powered to detect an effect of an expected size Sex/gender: Effects may differ according to sex or gender |
ODDPub [19] | Open data, open code: Open data and open code make it easier to reproduce analyses, identify potential errors, and re-use data |
Limitation-recognizer [20] | Author-acknowledged limitations: Every study has limitations. Acknowledging limitations provides essential context that allows readers to interpret the study results |
Barzooka [21] | Bar graphs of continuous data: Many datasets can lead to the same bar graph and the actual data may suggest different conclusions from the summary statistics alone. These graphs should be replaced with dot plots, box plots or violin plots |
Jetfighter [22] | Rainbow color maps: Rainbow color maps are not colorblind accessible, and create visual artifacts for readers with normal color vision |
Trial registration number screener | Clinical trial registration numbers: Clinical trials must be registered in an International Clinical Trials Registry Platform registry, and this number must be reported in publications. This makes it easier to detect practices like outcome switching |
Statcheck [7] | Misreported p-values: p-values that do not match the reported test statistic and degrees of freedom are common and can sometimes alter study conclusions |
Scite reference check | Citation of retracted papers, or papers with corrections or errata: Checking cited papers for editorial notices can help to identify potentially problematic citations |
Seek and blastn (semi-automated) [23] | Confirms that nucleotide sequences were correctly identified: Incorrect identification or use of nucleotide sequences makes it difficult to interpret or reproduce study results. Results from this tool require confirmation from an expert reviewer |