Dear Editor
Elizabeth Loder is right that ‘among the failures, credit is due’, but perhaps gives a little too much credit to journals that require authors to fill out data availability statements. Loder points to evidence from systematic reviews showing that publications in journals that mandate data sharing or inclusion of data availability statements were more likely to share data (16/87 (18%) v 4/213 (2%)).(1) In research more broadly, these statements commonly indicate that data is available upon reasonable request. Unfortunately, such data is only rarely made available when requested.(2)
The current regime of data availability statements probably leads authors to issue a perfunctory declaration of what they perceive sounds acceptable, out of eagerness to achieve publication. Data availability statements might be better than nothing, but only a modicum of credit is due for inducing authors to incant ‘available upon request’ before clicking submit.
Journals could easily clarify data availability statements by asking authors to simply state either whether or not data has been shared or, if data is declared as available upon request, to commit them to specific criteria whereby data will be shared. When informed that that data is not available as promised, journals could routinely amend or redact data availability statements to correct the record. More credit still would be due if journals institute an expectation of sharing basic documentation such as protocol and code on open repositories for all studies (3) and committ to targets for the proportion of publications which comply with such transparency as well as for sharing of types of study data which can be responsibly be made opely available.
1. Nguyen P, Kanukula R, McKenzie J et al. Changing patterns in reporting and sharing of review data in systematic reviews with meta-analysis of the effects of interventions: cross sectional meta-research study. BMJ 2022;379:e072428
2. Gabelica M, Bojčić R & Puljak L. Many researchers were not compliant with their published data sharing statement: a mixed-methods study. Journal of Clinical Epidemiology 2022; 150:33-41
3. www.OSF.io
A little credit, but not much