Shall the authors be required to submit their raw data along with their papers to enhance the reproducibility, credibility and quality of their results?
You can ask to see the data if you wish, some data has to be made available. It all depends on the type of project and data. Personal information, or data that could identify an individual shouldn't be available. I think that all published material should have gone through ethics approval too.
Ethics is a main norm, no questions about that. Allowing a direct access to someone's raw (not personal) data seems to be necessarily considered. It will re-fresh the readers' understanding - of the findings (to be better credible and reproducible) - to be more multi-dimensional.
I'm not sure I understand properly your question. Do you want to see the numerical contents of all bins of a multichannel analyser instead of a single graph with all those points presented as points, eventually with their estimated uncertainties and, perhaps, with a fitted curve? No journal/referee will be eager to print such long tables anyway.
But the data sets could be posted in the website of the journal.... There is no need to print them.
I don't see the problem, you can email the authors and ask for the data. I think that most funders stipulate this as a requirement
I strongly believe that raw data are proprietary materials. An interested reader can request for the data personally like Alan Chamberlain suggested. Credibility and reproduciability of any result should be the work of the editorial team and reviewers. In case of suspected data manipulation there are laid down rules by COPE on dealing with the issue based on who discovered it.
Some governments are putting in new requirements that all publicly funded data be archived, and made available openly (perhaps after a limited embargo period). One example is the US National Science Foundation (NSF). NSF has convened a number of committees to discuss how each discipline proposes to make this happen. In Earth Science there is the Earth Cube effort, for example. I urge interested earth scientists everywhere to get involved through http://earthcube.org .
Despite such efforts, I am strongly opposed to the idea that journals would require the raw data to be open, for an article to be published. Most science research is not publicly funded, and we want to do everything we can to induce industrial researchers to put their results out for peer review and open discussion.
I am a computer scientist, and one of my research interest is , broadly speaking, machine learning. I would love to have access to some data sets of already published research in order to compare my algorithm against theirs. Many researchers do not use propietary data (or any data with privacy restrictions). Some researchers are cooperative and share data if you ask them, but some do not share, because they are still using that data in research (next paper is continuation of previous paper) and do not want to grant any advantage to potential competitors. Or sometimes email adresses are obsolete (for example, my older papers have my email as student, already null because I graduated years ago, my main advisor,'s email is null because he has retired, and my second advisor's mail, null because he got full professorship at another university and moved). How do you get data sets if you can not contact authors?
There are many good reasons NOT to share data (patient personal data, industrial & proprietary research), but it is a great shame that "raw data are proprietary materials" views still dominate public-funded research. The practice is a false economy!
An experiment I am running at the National Insititute of Aquatic Resources (DTU Aqua) in to train researchers that they can archive raw data in long-term data centers and embargo it while they publish. We then explore publishing peer-reviewed data papers (e.g. Earth System Science Data Journal, Geosciences Data Journal, but granted that not all disciplines have them!).
This not only builds best practice for data management, but actually leads to better use of the data long term, visibility for the author. The author also build a research paper based on an already published data paper (or in parallel).
Sharing openly when done correctly feeds directly into the need for personal impacta and building a strong CV. Using a false email alias to prevent collaborators from requesting data is simply taking the easy road, it gives the wrong example to young researchers and is very much a false economy!
- Badges
- Science topic
- Similar topics
- Academic Journals