- Make all raw data available to other scientists
- Hold journalists accountable
- Introduce anonymous publishing
The first recommendation makes a lot of sense. If scientists are serious about allowing for experiments to be repeated in order to try and reproduce results, having raw data widely available makes that easier to do. (I see no need to limit this availability to other scientists, but that’s a separate conversation.) The graphic says that only 44 percent of ‘high-profile’ journals require making raw data available as a condition of publication. I would think that such raw data disclosure could be made part of the increasing number of institutional and other repositories for open access.
The second recommendation is pretty vague. It’s a complaint against the phenomena of articles simply repeating press releases. While I’m sympathetic, and I think this is a problem in more than just science journalism, without a means to achieve such accountability, this is just a nice goal.
The third recommendation strikes me as counterintuitive. With the number of instances where academic misconduct was perpetuated over time, I would think a bit more sunshine on the activities of these researchers would help find these instances. But the graph authors are looking at a different issue. They want to introduce anonymous research to address ‘unpopular discoveries’ – focusing on the findings to the exclusion of the author. Think of it as extending a double-blind review (where the reviewer(s) and author(s) don’t know the identity of the other) to publication and beyond.
Aside from issues of credit, and how that informs promotion and other professional advancement, there’s a question of acceptance. Yes, the review process is supposed to instill confidence in the published findings. But if there’s no name attached, I can easily see some giving such findings less ‘weight’ as a result. I don’t know that it will address the concerns the authors have. That doesn’t mean it couldn’t be tried. It does suggest that it would have limited applicability (such as large enough communities where blinded reviews can actually preserve anonymity.) There are other means to address risky and/or ‘controversial’ research, and having a host of solutions makes trying novel and/or problematic efforts
But would anonymized research really help fight ‘bad science’? Since the graphic writers made the recommendation to address a problem that’s not linked to sloppy research and/or research misconduct, I think they don’t really want to make that suggestion. But it’s hard not to make the inference based on the context of that recommendation. Anonymity has its places in scientific research, but I’m not persuaded fighting bad science is one of them.