Table 1.
Evaluation of journal standards and requirements as of July 2015.
Fig 1.
(A) Selection of articles: Twenty consecutive articles that met the inclusion criteria among those published beginning in January for both 2013 and 2015 in Nature (one that implemented a pre-submission checklist) and Cell (one that did not) journals. This represents articles from periods of time before and after the implementation of the checklist in May 2013. (B) Flow of the analysis: To examine whether quality of reporting has improved over time, the degree of key information reported in 2015 was compared to that in 2013 in both journals combined (Objective 1). To assess whether a checklist is associated with improved quality in reporting, we first compared the changes over time observed in Nature (④ vs. ③). If there was significant difference, we compared time “2015 vs. 2013” in Cell (② vs. ①) and Nature vs. Cell within 2013 (③ vs. ①) and 2015 (④ vs. ②) to adjust for differences between journals and changes over time in reporting (Objective 2).
Table 2.
Data abstraction form to assess the quality and transparency in reporting.
Fig 2.
Distribution of reporting study designs across time.
The distributions of the reporting status are presented in stacked bar graphs. The numbers inside the stacks are the number of articles corresponding to each percentage. The data for 2013 and 2015 are the total numbers of articles assessed from Cell and Nature within a given year. Fisher exact test was performed to assess the difference in reporting each methodological across time. Significant P values (< 0.05) are provided.
Fig 3.
The changes in rigorous reporting of study designs by a checklist.
The numbers inside the pie charts are the number of articles corresponding to each category. P values < 0.10 using Fisher exact test are provided to compare time 2015 vs. 2013 within the intervention (Nature) or the comparison (Cell) group, or to compare intervention vs comparison group within 2013 and 2015, respectively. ≠ is shown where comparisons between the touching two groups are significantly different with P < 0.05.