Retraction Watch is running a poll on a retraction question. As the rest of this post will pivot off that poll, I’d recommend that if you want to take it, do it now before reading the rest of this post.
SPOILER SPACE (I suppose for those who want to take that poll without bias.)
Nothing to see here….
Yet more nothing….
Just a bit more….
The poll asks what should happen to research that is proven wrong. This might have an obvious answer, but the specific example is Linus Pauling’s model of DNA. You probably haven’t heard about it, as it’s a triple helix. While there were some basic problems with the paper, it was written while the structure of DNA was not yet determined. So it wasn’t an issue of scientific misconduct, it was being on the wrong side of history.
So, what should be done? Should it be retracted? If so, should it be done just like papers retracted for other reasons? Or is a different kind of notation needed?
I’m inclined not to retract, but to annotate. There are editions of the U.S. Code (federal law) and U.S. Constitution that are annotated or at least show the edits made from subsequent decisions and/or amendments. Such documents show how they were changed over time.
The challenge here is that the scientific literature is not collected or organized in a similar fashion. There are textbooks, but not necessarily definitive collections of original literature. But, if the current drives towards open access continue, I would like to see the metadata connected to published resorts made more accessible. That way, if one looks up an article that has been retracted, or one that has been downplayed because subsequent research disproved it (or just failed to replicate results), the history of that document is available.
This may be a project of the easier said than done variety, but the growth of open access journals with strong online presences could demonstrate how it could be done. Or is this already being done and I just don’t read those journals?