Free cheers?

Share this on social media:

Discovery research shows that around 60 per cent of article downloads come from free resources, writes Tracy Gardner.

In March Simon Inger and I released the results of our large scale research project: How Readers Discover Content in Scholarly Publications. The research was the culmination of 18 months’ work and the report forms the latest in a series of research projects spanning 10 years.

Research for this latest project was carried out via an online survey and attracted 40,000 responses from people all over the world, working and studying in all sectors, subject areas and from a variety of roles – from students to professors, clinicians to researchers. We have 11 supporters to thank as between them these publishers and intermediaries sent out well over a million emails to invite people to participate in our research.

As a result of all of this work we now know an awful lot about content discovery! The report was published under a CC-BY NC licence and can be downloaded from our website at

The very top level headlines of the research included:

  • Abstracting and indexing databases (A&Is) are in decline, but remain the most important starting point for search when looked at across all sectors and roles in aggregate:
  • Academic researchers in high income countries rate library discovery as highly as A&Is, and rate academic search engines as the most important discovery resource when searching;
  • Academic search engines are now more important than general search engines in the academic sector in high income countries;
  • Online book discovery varies significantly by sector;
  • Publisher web sites have become more popular as a search resource;
  • Google Scholar is used more than Google in the academia, but less than Google in all other sectors;
  • Access to scholarly content by mobile phone accounts for only about 10 per cent of the use; and
  • More than half of all journal content delivery appears to be from free incarnations of articles. PubMedCentral is popular in the medical sector and social media sites appear to be a significant source of free articles in lower income countries.  

For the purposes of this article, we would like to explore the last point listed – the fact that more than half of all journal content delivery appears to be from 'free resources'

To give some context, the question we asked was: 'What proportion of the journal articles you read do you access from each of the following resources?'

The options offered were: a) The publisher website, journal website, full-text aggregation or journal collections; b) A free subject repository; c) A university’s institutional repository; d) Researchgate, Mendeley, or other scientific social sharing site; e) A copy emailed by the author or colleague. Respondents were then asked to select the percentage of content accessed via each of these options. We didn’t use the terms 'free' or 'paid', we have just used those terms here for illustration purposes.

We found that, regardless of sector, subject area, or income bracket of the country they resided in, respondents told us that on average around 60 per cent of content was being accessed from one of the free sites. In lower income countries this figure was slightly higher, and in higher income countries it was slightly lower but nevertheless the results are surprising.

In the medical sector, we surmised that more than 25 per cent of content is being accessed from subject repositories (the PubMedCentral (PMC) effect presumably) and in the academic sector around 20 per cent from institutional repositories. Around 10 per cent of content is from a social media site and between 10 and 20 per cent from the author or colleague.

When asking the question 'why this is so high?', we only need look at the key discovery resources people will use to find content – the starting points are key to this.

In the medical sector the most important discovery resource appears to be PubMed, if the full text of an article is hosted in PMC, PubMed will provide a very obvious link to the 'free version', so this result is fairly easily explained in this sector.

In the academic sector, Google Scholar is an extremely important starting point. Google Scholar indexes many subject and institutional repositories (as well as, it appears, ResearchGate) and puts a link to the full text when it finds it within its search results. Whilst library technology can facilitate an additional link which takes the user via their library link server and thus onto “paid” content, this function relies on the library setting up their link resolver details with Google Scholar, and their patrons updating their Google Scholar preferences with their institutions details. 

News announced earlier this month which will see 1science and EBSCO Information Services collaborate to provide academic libraries access to a vast number of openly accessible scholarly articles when performing a search in EBSCO Discovery Service will only add to the percentage of people finding and accessing free versions of content. (

Either way, a link to a free version of the full text can often be found in the Google Scholar search results page as can be seen below. The Find it @ Oxford link is how what a user from Oxford University would see if they had set up their preferences to take them via their library.  

This was a new question for us in 2015 as many of our supporters were keen to get data on this key question. Of course, some may question these results as, quite rightly, they will say that many people don’t ever really know where they are getting their content from – they just click on a link and hope it appears. Even even if they don’t know where the content is coming from, there is of course an argument to say that it doesn’t really matter what they ultimately do – if they believe they are getting their content from a resource which is effectively free, they may well not be happy to have to pay for that same content in the future.