Looking for a project? Analyse open-access involvement by career-stage
October 9, 2013
The LSE Impact blog has a new post, Berlin 11 satellite conference encourages students and early stage researchers to influence shift towards Open Access. Thinking about this, Jon Tennant (@Protohedgehog) just tweeted this important idea:
Would be nice to see a breakdown of OA vs non-OA publications based on career-stage of first author. Might be a wake-up call.
It would be very useful. It makes me think of Zen Faulkes’s important 2011 blog-post, What have you done lately that needed tenure?. We should be seeing the big push towards open access coming from senior academics who are established in their roles don’t need to scrabble around for jobs like early-career researchers. Yet my impression is that in fact early-career researchers are doing a lot of the pro-open heavy lifting.
Is that impression true?
We should find out.
Here’s one possible experimental design: take a random sample of 100 Ph.D students, 100 post-docs, 100 early-career researchers in tenure-track jobs and 100 tenured researchers. For each of them, analyse their last ten years of publications and determine what proportion are paywalled, what proportion are free to read (e,g, on arXiv or in an all-rights-reserved IR), and what proportion are true (BOAI-compliant) open access.
An alternative approach would be to randomly sample 1000 open-access papers (from PLOS and BMC journals, for example), and 1000 paywalled papers (from Elsevier and Springer, say) and find the career-stage of their authors. I’m not sure which approach would be better?
Who is going to do this?
I think it would be a nice, tractable first project for someone who wants to get into academic research but hasn’t previously published. It would be hugely useful, and I’m guessing widely cited. Does anyone fancy it?
Update
Georg Walther has started a hackpad about this nascent project. Since Jon “Protohedgehog” Tennant has now tweeted about it, I assume it’s OK to publicise. If you’re interested, feel free to leap in!
Sauropod Vertebra Picture of the Week 
October 9, 2013 at 8:04 pm
This seems like something that could be organised at a hackday. At one recently, an app was designed (I think by Martin Fenner) that analysed author contribution for the entire PLoS corpus (or a small subset of 1000, I don’t remember). The difficulty is mining the information that relates authors in a published document to their faculty position.
October 9, 2013 at 8:12 pm
In another tweet, Jeroen Bosman points out that Wiley recently conducted a study covering some similar ground, Generation Gap in Authors’ Open Access Views and Experience, Reveals Wiley Survey.
October 9, 2013 at 9:16 pm
Mike, I really like this idea. I tried to get at this kind of data through a survey of faculty at my institution. But with such low response rates to surveys it is difficult to get an accurate picture. I think you’re right, analyzing this from a bibliometric standpoint is the way to go. I might try this on a small scale within my own institution’s researchers where I will be able to determine the status of the authors more easily…
October 9, 2013 at 9:24 pm
Thanks, DeDe. Doing this within a single institution would be a really useful step. If you proceed, please do let us know what results you get. (I see all comments left on this blog, however old the post they’re added to.)
October 10, 2013 at 12:14 am
I agree it’s a great idea and also that a survey is probably not the best approach to data collection. A sum of smaller data for specific institutions might however work as long as they share the basic criteria (classifications etc). Then again, although this area has been a bit more deeply explored, crossing the career stage analysis with another one by disciplines would be really useful
(thinking about the STM vs SSH among other threads here).
October 10, 2013 at 12:36 am
It does seem a good idea, and should complement surveys like
http://exchanges.wiley.com/blog/2013/10/08/mind-the-gap-2013-wiley-open-access-survey/
well. Perhaps it should be advertised to students in social science disciplines – it would be nice to do for one’s own discipline (if not social science) but wouldn’t gain much academic credit, no matter how useful it is.
October 10, 2013 at 9:39 am
Indeed a good idea! Surveying researchers for their papers from the last 10 years might result in a skewed picture, though: 10 years ago, there weren’t many good OA options available yet, and today’s PhD students were still in high-school.
That would make it look like researchers further in their career tend more towards subscription-based publishing, regardless if that was true or not.
To avoid that, I’d suggest to use a time period in which all surveyed researchers were already in positions where they would realistically publish something.
October 10, 2013 at 9:41 am
You’re right. I suggested that last ten years (rather the whole career) precisely to cut down on this sort of bias, but on reflection ten years is still much too long. Should probably be limited to, say, 2011-2013.
October 10, 2013 at 12:48 pm
Great idea! I’d suggest figuring out if the LinkedIn API might be used to work out people’s career stages over the years — otherwise you’d need to sift through tons of personal webpages to get this data.
The next problem would be matching LinkedIn users with their publications reliably — however some users do indicate their publications as far as I know and we might even use a combination of users’ names and their listed institutions to increase accuracy.
I’d be happy to contribute if anyone wants to tackle this collaboratively.
Not sure if useful, but if any interest exists to work on this online, tracking ideas in a hackpad might be good:
https://hackpad.com/Analysis-of-Open-Access-Involvment-by-Career-Stage-PzAK6r4kefd
October 10, 2013 at 3:10 pm
Have a look at http://project-soap.eu/documents/. In 2001, Project SOAP (Study of Open Access Publishing) surveyed nearly 40,000 active researchers, and have statistics on how many OA articles they have published, broken down by career stage, discipline, country etc. The full dataset is available. It would be good to update this research certainly.
October 10, 2013 at 3:13 pm
Thanks, Deborah, that’s useful. I assume you mean in 2011?
October 10, 2013 at 3:15 pm
I did mean 2011. Sorry about that!
October 10, 2013 at 4:07 pm
It would be very interesting to do a bibliometric environmental scan from a OA group and paywalled group to see a snapshot of researchers, instead of following their careers specifically.
October 10, 2013 at 9:55 pm
I think ORCID may offer a good opportunity to do the kind of coupling suggested above between publications and career stages via LinkedIn, both by recording authors’ institutional affiliations for their whole career and by eventually linking ORCID iDs to other web identities such LinkedIn’s. Checking the level of detail of institutional affiliations provided by the ORCID affiliation module about to be released as a pilot could indeed be helpful for this purpose.
(Thanks for the Hackpad entry, @georgwalther, I included this there too).
October 11, 2013 at 2:43 pm
Great contributions so far in the hackpad!
We should probably find a way to separate chatting about the project and bouncing ideas and the working on the project itself.
Anyone have any good suggestions for group chats?
hipchat? Twitter?
October 11, 2013 at 2:47 pm
I’d use good ol’ IRC for group chat because it can be used with any (web, mobile or desktop) client without the need to register to any social network.
October 11, 2013 at 2:50 pm
Cool .. down with that. Can you set something up so that the discussion is also logged for future reference?
October 11, 2013 at 3:07 pm
There are logging bots for IRC available, but I don’t have any experience with them. So it is possible to set up automatic logging, but with my current knowledge, I can’t
October 14, 2013 at 6:03 am
You might want to check out concepts like the Polymath project. People collaborating on serious mathematical problems online, and so far it has been through blogs (e.g. the most recent one has been collated here: http://michaelnielsen.org/polymath1/index.php?title=Bounded_gaps_between_primes#Polymath_threads)