Thursday, April 02, 2009

Academic journals are killing science

In an interview the astrophysicist and all round clever chap was asked a question that was outside the current scientific data. After Sagan repeatedly told this to the interviewer the interviewer asked him to give a gut answer. Sagan famously replied "but I prefer not to think with my gut."

I don't particularly like thinking with my gut either, but in this blog I am going to give my gut its voice. There is doubtless much research and argument relating to what I am going to say, but right now I'm too busy or lazy to look it up. So without further ado, I can hand you over to my esteemed colleague Mr Gut.

It can't just be me that is becoming increasingly frustrated with the whole process of submitting papers to academic journals. You faff about putting in the correct format (which always seems to be different for different journals, even those within the same topic area). You put the figures at the end, or in the text, you save the figures as a .tiff, or just as word files, you anonymise the paper (or not), etc. etc. Then you send it off (thankfully via email or upload now -- some things have improved) and then you wait. You wait until the editor finds some people to review the paper, gets agreement from potential reviewers that they are happy to do and then sends out copies (or more likely emails a link to a pdf) to the reviewers.

Sometime between 3 and 6 months after you originally submitted the article you get back the reviews and are told whether or not the paper is accepted, and if it is accepted what revisions are required. If it is accepted it might come out a year later.

So lets summarise. If all goes well it might take a minimum of a year between actually doing the research and getting the paper into print. If it all goes less well it can take much longer. For example myself and a colleague submitted a paper in 2001 (the research was done in 2000) to Cognitive Science. They rejected without review because it wasn't interdisciplinary enough. It then went off to Journal of Educational Psychology who required too many fundamental changes. We were then invited to submit it to a journal called Discource Processes by the editor himself. This we did in 2003. They wanted changes, we made the changes, they rejected the manuscript, we submitted it to International Journal of Human Computer Studies. They required revisions, we did them, they accepted, job done.

The paper was published in 2007, six years after we submitted it to the first journal and seven years after we did the experiments. Fortunately for us the paper was rather theoretical and wasn't something that dated, but can you imagine that it was a paper on social media? We would have had a paper on discussion lists and MUDs published in the age of Twitter and Facebook, potentially still relevant but hardly current.

Not all paper take this long of course, but even a two year gap between study and publication is unacceptable, this is probably one reason why many academics are turning to blogs and the like to get their ideas in the public domain. I am fine with this. If it is in an area I know well the absence of peer review causes me no problem at all. I can tell for myself whether the arguments and data are good or bad. But it is important for those who cannot do this that the article has independent verification of quality and it has to be independent. Anyone can get their academic chums to give their blog-paper the thumbs-up and therefore the specious patina of respectability.

So there must be a way of speeding up the review process whilst still offering quality control.

In Wikinomics the authors discuss the case of particle physicists who use upload their manuscript to a wiki which is then edited by collaborators and finally published, a process that takes weeks if not days to complete. This is particularly imporant in some of the hard sciences (high-energy physics, genetics, etc.) where things move so quickly, but I also think it is important in many other academic disciplines (such as social media). The question is how to movivate the 'reviewers'? They could be rewarded by becoming a named author on the paper but then there is the problem that people might develop a pro-publication bias to get a publication. But the motivation should really be that participating in the reviewing process allows you to submit your own articles to the journal: everyone would surely benefit from their getting their papers turned round in 1/10 or so of the time it would do normally so academics should be falling over themselves to obtain membership of this club by performing reviews.

Now that my gut has been given its head (as it were), the rational part of me (to committ that egregious Cartesian fallacy) would like to ask anyone reading this. What do you know about attempts to do this, especially in the non-physical sciences. Are there any problems (one can imagine all kinds of game theoretic problems occuring). But does it work? It certainly should.

No comments:

Post a Comment