Ecology is not a dirty word

I’ve never considered using preprints for my own papers, I’ve never cited one, and I don’t plan to jump on the preprints bandwagon just yet. I read Terry McGlynn’s recent post on why he’s not bothering with preprints, and I agreed with every point. And then I read Sophien Kamoun’s rebuttal post, and I kind of agreed with some of those points too. I started reading the conversations happening on Twitter around these two posts and got a headache. There are some very high-profile scientists that are vocal proponents of preprints. Others are not.

With all this opposing opinion, what should an early career researcher (ECR) do?

Preprints are essentially final drafts of papers that are posted openly online through a preprints server, before they have gone through the formal journal submission and peer review process.

From my perspective as an ECR scientist, there are three main assumptions that need to be met for me to accept the preprints model at this stage of my career.

(1) All my final drafts of papers are written and justified clearly and are without scientific error. This is rarely the case. I’m a pretty good writer, and I’ve got a pretty good grasp on ecological statistics. I’ve published 21 papers so far, 13 as lead author, all of them through peer-reviewed journals. Nearly all of these ‘final drafts’ had major issues of clarity in some sections that affected the reader’s interpretation, and three of the data papers had some analysis issues – peer reviewers pointed this out. One of my favourite papers, a conceptual paper on quantifying net outcomes of animal activity in agroecosystems, overlooked a key property of social-ecological systems – we expanded on this after a reviewer kindly pointed this out and the published paper is a lot better for it. From a career perspective, I’m thankful that only a few people saw these ‘final draft’ versions of my work and pointed out the errors before publication, instead of the embarrassing versions now being freely available online, forever…

(2) Only my actual peers will read my preprints. This is definitely not the case and is one of my key hesitations with the preprints system: its potential influence on the spread of misinformation. I’ve seen people on social media use flawed preprints to argue against solidly established evidence. With all the media coverage claiming that the peer review system is apparently broken (it’s not), it’s hard to convince someone that a flawed preprint is not ‘scientific evidence’.

(3) Posting a preprint will connect me with more (and potentially better quality) peer reviewers than I would have access to through the traditional peer review system. This rarely happens. I did a quick search of ecology preprints at PeerJ Preprints and bioRxiv. Out of 84 preprints I looked at with dates ranging from 2013 to 2016, only 9 had peer review comments left at the bottom (maximum 2 comments on a single preprint). Most authors had been waiting over 3 years for peer review… Is that better than the current journal system? (I also found this set of almost identical papers, none of which had been commented on: 1, 2, 3)

I agree that the hypothetical system of preprints has some hypothetical benefits. There are some seductive arguments for preprints, particularly from an ECR’s perspective:

• ECRs need to publish papers and build a track record relatively quickly to get jobs and grants to survive. Every so often, inappropriate peer reviewers/editors can delay your paper from being published a lot longer than necessary. This is really frustrating, especially when you’re applying for jobs and want to highlight your publications list. Preprints circumvent this problem.
• Preprints are a protest against Big Publishing.
• Preprints can potentially stop you getting scooped if a journal takes too long to review and publish your paper.
• Open science etc.

BUT there are also lots of counterarguments to these arguments. For me personally, the risks outweigh the benefits and the logic just doesn’t make sense.

• Science is driven by community dynamics and collective knowledge, not individuals. I want peer feedback before I share my work publicly. The interpretations and misunderstandings identified by peer reviewers are a good indication of how a broader audience will interpret and misunderstand my paper. I agree that peer discussion should be open, but why not encourage authors to seek pre-submission peer review, and adopt open peer review within the existing journal system?
• Bad peer review is driven by individuals, not the system. Trying to circumvent the system doesn’t solve the problem of rogue reviewers/editors.
• There are other ways to protest against Big Publishing, like publishing in society journals, or boycotting the worst offenders. And besides, a paywall is not a death sentence for an article; there are plenty of ways to share your paper e.g. blogging about your papers, sharing copies via email or on your website, publishing in journals that provide free author sharing links, or that have more lenient online sharing policies etc.
• The spread of misguided opinions and invalid data as ‘facts’ via the internet is already a key challenge to public understanding of science. Preprints potentially add to this problem. It is naïve to argue that preprints are just read by peers. And it’s duplicitous to publicly promote preprints as valid science at the same time as blaming popular media for spreading misinformation.

Here are some points on preprints that I think deserve more discussion before someone on the fence can make up their mind:

• Adoption of preprints as a publishing system should be accompanied by greater commitment to public education on scientific method and communication. Just stating that ‘this paper has not been peer-reviewed’ is not enough in a world where many non-scientists don’t know what the purpose of peer review is, or when media (and some scientists) consistently broadcast that peer review is useless anyway.
• Preprints originated as a system for sharing theoretical physics papers. I asked a few colleagues in humanities disciplines what they thought of preprints – they’d never heard of them and were utterly horrified by the concept. So are preprints more applicable to basic/fundamental sciences, where relevance of results is limited to direct peers, compared to applied or environmental sciences, where claimed results have potentially broader impact beyond specialist boundaries?
• Yes, dodgy papers get published in reputable peer reviewed journals too. Does anyone have any data on how the frequency of dodgy peer-reviewed papers compares to the frequency of dodgy preprints?
• Are preprints more ‘acceptable’ by peers if published by an established researcher than by a junior researcher? Is the system geared toward ‘well-known’ authors with large social media presences who get more exposure and potentially more peer comments for their preprints? As I looked through the bioRxiv ecology preprints, where the tweets that shared the link are listed, preprints by relatively ‘well-known’ authors (subjective, in my general field) appeared to have been tweeted more often than preprints by less well-known authors. Interestingly, many of the tweets also presented the preprint as if it was an actual paper, with no hint that it hadn’t yet been peer reviewed.
• Most scientific communications are screened by others before being publicly shared. Scholarly articles go through the peer review process, popular science articles are edited by the medium’s editor. In contrast, most blogs, just like this one, are not ‘peer reviewed’ before going live. So how do preprints differ from using blogs to share unpublished studies, opinions or negative results?

© Manu Saunders 2017