I have previously blogged about the issues that preprints can cause on Google Scholar. Today I was reminded that these issues have real-world implication for junior scientists, and that they may discourage junior scientists from posting preprints.
I had the following conversation with one of my students (paraphrased):
Me: So, do you want to post the paper we just submitted as a preprint?
Student: No, not really.
Me: Are you concerned about keeping your competitive advantage, so you can finish a second paper on the topic before we reveal to the world what we’re up to?
(With this particular paper, I had wondered whether we should submit it as a preprint or not. There are a number of obvious follow-up works we can do relatively quickly, and so could others.)
Student: No, I’m not particularly worried about that. I just don’t want Google Scholar to list my paper as bioRxiv for the next few years, way past the time the paper has actually come out. Several of my current papers are still listed as their preprint version even though they’ve appeared ages ago. Maybe once I have 100 papers and 10,000 citations I won’t care anymore, but at my current stage I can’t afford having Google Scholar obscure my record by listing all my papers in their preprint version only.
Update #1, 12/03/2014: I'm getting a lot of comments to the effect that one can edit the Google Scholar profile. A couple to responses to that:
Yes, there are manual workarounds for most issues. That doesn't mean the default behavior of Scholar is not discouraging.
Merging of articles doesn't work when the preprint shadows the final article, because the final article is simply not visible in the Scholar database.
Even if you fix your own profile, that doesn't fix how the article appears on your co-authors' profile or in a general search for the article, e.g. by title.
It is not clear what happens to citations of shadowed articles. Are they or are they not counted? We don't know. There's certainly the worry that they are not.
The only way I know to fix shadowed articles on your own profile is to manually add the reference, then merge, and then undo all of that a year later when Scholar has finally caught up to the existence of your article. It's cumbersome, prone to errors, and certainly discouraging.
I will not stop posting preprints. But I will also not pretend everything is fine with Google Scholar and preprints when there are some glaring issues. Google Scholar is being used increasingly by departments in hiring and promotion decisions. Scientists should rightfully worry about how their work does or does not appear on Scholar.
Update #2, 12/03/2014: So it turns out manually editing entries doesn't work as expected when articles are shadowed by their preprint. You can add the reference, but you cannot make it link to the correct article. Check out this entry. The title is not clickable, and the Scholar articles listed at the bottom do not include any links to the actual journal version of the article (or even the latest preprint version).*
*My statements are correct as of 12/03/2014. Eventually Google Scholar will catch up and the links will appear.