Wednesday, January 27, 2016

Academia.edu wants to commercialize its "recommendations"



Academia.edu has a system of "recommendations" for publications that I have never been fond of. Now they want to commercialize them by selling commendations. The way the recommendations work is that some scholars are invited to submit recommendations. I can't find the criteria listed on the website, but as I recall the only criterion was that one had published one or two papers ever. The recommenders are then supposed to recommend papers by clicking a button on the paper in question. This information become public, and the number of views on those papers increases. Individuals are given an "Authors rank" based on the number of recommendations their papers have received, adjusted for the rank of the recommenders. My author rank is 3.6, but I have no idea if that is high or low; the nature of the scale is not revealed.

I tried being a recommender for a while. I recommended some things, and then I'd get messages stating that views of those papers had increased dramatically after I had recommended them. Wow, I am an influential guy in Academia.edu! I'll put that on my CV. But without information on why one is recommending a paper, these recommendations don't carry much weight. And when the system got started I snooped around to see who was doing the recommending. Some recommenders are serious scholars whose views I take seriously (people like Gary Feinman and Linda Manzanilla). Others are low-quality scholars whose views I do NOT take seriously (I won't name names here. I manage to get enough people pissed off at me as it stands, I don't need make a bunch more people mad). So having one's papers recommended by someone like Gary Feinman (one of the top archaeologists, in my opinion) has the same weight as having them recommended by low-quality scholars. Not a very good system.  Plus, there is no way to give a negative recommendation. Some papers are terrible and deserve to be described as such, but that is not possible with this system. This is one more example of the facebookization of online scholarship. You can like something, but you can't dislike anything.

I just got the following email from someone on the Academia.edu staff:



 Hi Dr. Smith,
   My name is XXXXX, here at Academia. I noticed you had received a few recommendations on your papers. Would you be open to paying a small fee to submit any upcoming papers to our board of editors to be considered for recommendation? You'd only be charged if your paper was recommended. If it does get recommended then you'll see the natural boost in viewership and downloads that recommended papers get. Would love to hear your thoughts.




Here are my thoughts (this was my email reply) - I don’t have a very high opinion of your system of recommendation. As it stands, you have a bunch of low-quality scholars making recommendations, and I don’t consider the recommendations any kind of rigorous or useful measure of anything. Getting visibility through Academia.edu is useful, I guess, but it is not very high on my list of professional goals. The idea of paying a fee for recommendations sounds ludicrous. Who is on the “board of editors” to make these decisions? Non-professionals? Low-quality scholars? I have previously looked at Academia.edu as an alternative to the trend of increasing commercialization of scholarship. But now you want people to pay for some kind of recommendation? The recognition that matters to me is citations, not some social-media type of “liking” or fee-based recommendations. Please leave me out of it.

(END OF EMAIL)

 Academia.edu is strange. It will have some very positive scholarly practices, and then it will introduce a retrograde, anti-scholarly features like co-authors listing. I think the whole idea of recommendations, as currently implemented, is a pseudo-scholarly feature, and I don't trust it. One could design a better and more transparent system but that might be too complex. But this idea of selling recommendations is terrible. If this is implemented, I might consider leaving Academia.edu and posting my papers elsewhere.

2 comments:

Richard Price said...

Hi Michael,

Thanks for your thoughts.

You say "So having one's papers recommended by someone like Gary Feinman (one of the top archaeologists, in my opinion) has the same weight as having them recommended by low-quality scholars." That is not how the system works. A recommendation is weighted by the AuthorRank of the recommender. Someone's AuthorRank is based on the PaperRanks of their own papers. Assuming Gary Feinman's papers have been recommended more than others, his AuthorRank will be higher, and his recommendation will carry greater weight.

The background of the flare-up on Twitter is that if an idea comes up over lunch at the office, a team member will sometimes reach out to users to discuss it. In this case it was an embryonic idea that someone had, and not something on the company roadmap. We were not expecting the email would get shared on Twitter.

Most of the embryonic ideas we have in the office look bad on the surface, and end up actually being bad on closer scrutiny. But occasionally an idea makes it through the vetting phase. Typically the products we build (e.g. Sessions) started out in private communication with users, and evolved significantly.

The general background to Adnan's email is that we want to start the conversation around how to fund academic publishing when paywall revenues dry up (which I think they will over the coming years). The sciences are switching to an APC-funded model, but that model doesn’t straightforwardly work for non-grant funded people in the humanities. It seems to us that either someone figures out a super low-cost APC for humanities publishing ($50 or so) or you have the normal APC (around $1,500), and figure out a way for universities to cover the fee. When Adnan reached out to users, he was probing the first idea.

Richard Price
Founder, Academia.edu

Michael E. Smith said...

@Richard - Yes, I wondered if this was just an idea being floated. Even with your algorithm, I still don't see the score on a paper as a useful scholarly recommendation (too many non-transparent aspects). The question of journal funding over the long haul is a big question, too big to solve in blogs and minor changes to Academia.edu.

I do appreciate Academia.edu and the attempts to improve the service, and I hope that you continue to innovate and improve the system. Given that universities (like mine) are now starting to take down faculty websites, your service will be even more valiable going forward.

Mike Smith