In my previous post, I briefly mentioned Ethan Perlstein’s two posts on publishing in the era of open science (part one and part two). I just wanted to quickly highlight something that I haven’t paid much attention to, but perhaps should have, regarding the importance of post publication. In his post, Perlstein mentions the ratio of HTML to PDF downloads, and how he used this as proxy for the quality of readership (i.e. expert, academic readers versus non-expert readers):

Some of you may be miffed by my appar­ent con­fla­tion of site traf­fic and read­er­ship. With­out more sophis­ti­cated ana­lyt­ics, I con­fess that it’s dif­fi­cult to gauge the back­ground of read­ers (sci­en­tists vs. non-scientists), or how much of the paper they’re actu­ally read­ing (abstract vs. full text). How­ever, the ratio of HTML views to PDF down­loads (HTML/PDF) may be infor­ma­tive here. After the ini­tial surge, HTML/PDF was 1 in 20 and remained there until the sec­ond surge, after which it fell to 1 in 30. If we assume that PDF down­loads are a proxy for “expert” read­er­ship, then the sec­ond surge diluted qual­ity read­er­ship. Con­versely, the third surge lifted the ratio to 1 in 15, with as many as 20% of read­ers, many of whom were pre­sum­ably aca­d­e­mics, on Day 29 choos­ing to down­load a PDF ver­sion of the paper.

It certainly rings true for me that I only download the PDFs of papers I intend on referencing and making use of on more than one occasion. Still, my intentions sometimes verge on the side of completely unrealistic and I’ll download a large number of PDFs that don’t survive past a cursory glance. I guess, for me at least, there is a stronger driving motivation behind my downloading of PDFs: the cluttered design of HTML pages for many journals simply drives me insane. Rather than any in-built preference for reading in PDF, the need to get away from the HTML page is more than enough motivation for me to download a small document. Still, with better online reading and annotating tools being just around the corner, I wonder how long this variable will hold as a viable proxy?

In short, I certainly think more research needs to be done into readership, with one aim being to develop metrics that accurately capture expert and non-expert audiences.