I’m sure all scientists have one – a folder with unfinished manuscripts, or ones that were finished but never found homes. In my case, the folder has manuscripts as old as 6 years and as recent as a few months. But there was one manuscript in particular that I just couldn’t let go. It was work from my dissertation, which is probably one of the reasons I felt so attached to it. I’d invested many hours (some of them while pregnant and far too big to be sitting at an electrophysiology rig) gathering those recordings. In my completely biased opinion, they were beautiful, interesting, and deserved to do more than sit buried in a 300+ page dissertation that few people would ever read.
The problem is there are gaps in the work. We were unable to fully characterize one of the genetic manipulations we were using, leaving doubts about the precise effects on the cells in which it was expressed. There are also a lot of open questions. Some of the results were unexpected and the underlying mechanisms merited further exploration. Additional experiments could have ruled out certain possibilities. But then life happened. I had a baby, graduated, moved away. I wrote up what I had, but with the obvious gaps, the manuscript seemed unpublishable in mainstream journals. I knew what reviewers were going to say and I had no satisfying answers to give them. I no longer had access to the equipment required to do experiments, so collecting more data was out of the question. The manuscript sat in my folder collecting digital dust. For the next two years, I would occasionally read over it again and I always arrived at the same two conclusions: (1) that it included interesting data that deserved to be seen, and (2) that it would not pass peer review in a mainstream journal. So, it was either archive it, or find a non-conventional way to get the work out.
This last week, inspired in part by a recent post by Zen Faulkes @DoctorZen explaining why he published a paper on his blog, I decided that I had to put the work out in some form or another. I considered publishing it on my blog, but since my blog is not well-established, I felt I needed something more visible. I began to debate the pros and cons of posting my manuscript on either arXiv or figshare, and opened it up for discussion on Twitter. Thanks to Bala Iyengar @balapagos, P Desjardins-Proulx @phdpqc, and @figshare who contributed to the discussion and provided information that helped me make my decision. In the end, I’ve decided to go with figshare. Here are just some of the reasons why:
Figshare accepts all file formats. In answer to my query about whether figshare accepts tex files in particular, @figshare tweeted:
I responded that I think this flexibility and utility gives figshare an advantage over competitors like arXiv, which are more restrictive in the file types they accept. According to their website, arXiv accepts only LaTeX, PDF, Postscript, or HTML for text submissions, and only PS/EPS, JPEG, GIF, PNG, or PDF (if using LaTeX) for figures. Now, I’m a LaTeX gal. I prepare almost all my documents, presentations, and posters in LaTeX. So, I have no problem submitting in one of the file formats specified. In addition, @phdpqc tweeted:
I couldn’t agree more on the need to be saved from MS Word, and it is definitely a plus that arXiv provides LaTeX source code. But figshare loses no points here. As already mentioned, they accept tex files, too. I also like the fact that no matter what format I may have for supplementary materials, be it figures in another format, videos, spreadsheets, code, etc., figshare will take it. In addition, figshare allows you to upload all of your research outputs, even if they don’t come as part of a full-fledged manuscript. In contrast, arXiv only accepts materials that are linked to a manuscript submission.
Materials are available on figshare immediately after upload. I think this is what @balapagos had in mind when tweeting:
In contrast, submissions have to go through moderation on arXiv before they are posted. I had the impression from arXiv’s site that this process might take some time, but @phdpqc commented:
Ok, 48 hours. Not so bad. Still, having been on the other side of the lab bench, spinning your wheels for any number of days because you didn’t know the experiment had already been done (or failed) can be frustrating. And even if I didn’t mind the delay, there’s something else about arXiv’s moderation that bothers me. On their site it says, “The arXiv moderators are experts in their fields and in the types of submissions that are appropriate for their subject classifications. They evaluate based on the content of the submission. arXiv moderators may recommend reclassification or removal of a submission”. At the risk of sounding less than humble here, I am an expert in my research and I think I know best how it should be classified. On figshare, there is no moderation process that reclassifies or removes my research materials (unless, of course, they contain something completely inappropriate). I decide what to upload, what categories to place it in, and what tags to put on my submission.
Figshare gives my research a Digital Object Identifier (DOI). This has a number of advantages, including making the work easily citable. arXiv allows you to add a DOI if your article is published elsewhere and given one, but does not automatically give one to your submission. In addition, as mentioned above, arXiv only allows submission of manuscripts, which means ‘orphan’ research outputs such as figures not part of a manuscript, posters, grant applications, etc. cannot get any type of citable link on arXiv. On figshare, all materials receive their own DOI, making them citable units in their own right.
Figshare does not require me to be endorsed to submit. For me, this was the real kicker. arXiv has a system by which authors submitting for the first time, and particularly those who do not have a university affiliation, must be endorsed by another arXiv user before they can submit. The argument is, “endorsement…will verify that arXiv contributors belong [to] the scientific community”. Since I don’t currently have a university affiliation and this would be my first time submitting to arXiv, I would be required to find an endorser. I have friends and colleagues who submit to arXiv and could endorse me. But that’s not the point. First, this process would be yet another delay, in addition to moderation, in the dissemination of my research. True, the delay would likely be negligible, but it is still a delay. Second, I object to this endorsement system on conceptual grounds. Just because I do not have a current university affiliation, does that make me less a part of the scientific community? In the current job market, there are many ‘floating academics’ like myself who do not have an affiliation. In the computational sciences, there are also many researchers who work as programmers in companies and do research in their own time. Is there research less trustable because they don’t work at a university? Furthermore, the growing rate of high-profile retractions shows that simply belonging to the scientific community in the form of a university affiliation is no guarantee that the research is of good quality. arXiv says, “people who fail to get endorsement are still free to post articles on their web site or to submit their publications to peer-reviewed journals”. Or, they can post them to figshare without worrying about an endorsement from another scientist.
Don’t get me wrong, I like arXiv. Any site where people can post their science and people can read it for free is a good thing. But I’ve decided that the better home for my manuscript is figshare. All that remains is to do some final editing before I upload it. In the end, what am I hoping to get out of this? I know that it is unlikely I will get many citations. I’m also under no illusions that this manuscript will help me get that tenure track position I’ve been searching for. Unfortunately, most search committees will not consider this a ‘real’ publication. In fact, there are many who would say I am wasting my time preparing this manuscript (or even writing this post) when I could be working on papers that have a chance of getting published in peer-reviewed journals. Maybe they’re right. But this is important to me. I believe strongly in open science and that research, no matter the gaps or open questions it leaves, does people more good when it gets seen. Ultimately, people will have to evaluate for themselves what contribution, if any, my manuscript makes. I am hoping at the very least that it gives others ideas that they can build on in their own research. And if nothing else, it will be one more manuscript I can move out of the homeless folder.