Why open access isn't enough in itself | Higher Education Network | Guardian Professional

"I hit a career high when my research on university library data recently appeared in the Guardian education pages. However, my initial surge of excitement was tempered slightly when I saw the journalist hadn't provided references to my actual study and significantly when I realised the single finding she'd quoted was – to put it charitably – not very reliable. My feelings on browsing the comments section can't really be articulated in a family news outlet, but it was clear that many people had picked up an incomplete understanding of how and why the research was being done. But the worst part of the experience was this: I had blogged the whole project – background, rationale, false starts, findings, corrections – openly online. Aren't these misunderstandings exactly the kind of problem that openness is supposed to prevent? It's easy to see what happened. The project was a study using routinely-generated behavioural data to explore student library use and how this relates to degree results. Early results had suggested that low library use might be correlated with dropping out, but the measure was pretty crude: I stressed this in the blog post reporting the results. Later, we did some further analysis which provided a more nuanced, ambiguous but accurate interpretation of the data. Guess which one became the hook for the piece? The government's response to the Finch review, published about a year ago, revealed a subtle but important shift in rhetoric about open access. Public access to research was always in the background of earlier position statements by, for example, the Budapest Open Access Initiative, but became a central pillar of David Willetts' argument for greater openness. But my recent experience suggests that paywalls may not be the only barrier we need to overcome if the public is to benefit from access to research. Just because we make something open does not mean that people will read it, even when it is directly relevant to an issue they are trying to understand. Even if the Guardian journalist got the rogue statistic from one of her interviewees, the blog is easily discoverable for a bit of basic fact-checking. The same goes for the commenters who were railing against our study. We know that the relationships we're reporting probably aren't causal. We know there are ethical issues around using student activity data in this way. We know that, taken as standalone statistics, the findings might seem blindingly obvious. We've addressed all these issues in the project blog. But that is no use if nobody reads it. Being open about the messiness of the research process carries risks as well as benefits. We had a long discussion before blogging that early, flawed, but highly-quotable finding. I was worried that it might be picked up and used without the all-important context which explained its limitations. In the end, our desire to be open trumped this concern, but unfortunately I was proven right. And it's no good to argue that we made a rod for our own backs by sharing something we weren't completely certain about. Every published journal article is part of a bigger academic conversation; a new finding offered up for comment, replication and testing. Rising retraction rates, which tend to be higher in high-impact journals, show that even the peer-reviewed literature contains mistakes, and academic progress is based upon improving on the flaws in somebody else's work. Making scholarly content open to everyone, including people who aren't familiar with scholarly communications conventions is also risky. Academics write in a specific language to communicate with their peers: they use shorthands and references that assume familiarity with ideas, conventions and techniques. Medical charities require lay summaries of research proposals and findings for a reason: the general public can spend time and effort learning to understand academic-speak but not all of them want to. And of course, they will only do so if they know that they need to. Some of the Guardian commenters had a misplaced trust in their own understanding of statistics. Correlation is not causation, nor is it predestination. So, no, your experience as a high-achieving, non-library-using postgraduate doesn't disprove our two years of analysis on undergraduates. Sorry. None of this is to argue against open access. Anything that is published, open or otherwise, risks being picked up and misunderstood by a journalist. And as a researcher working outside a university I'm a big fan of being open, having crashed into paywalls more often than I'd like. If my recent research run-in shows anything, it's that throwing open the doors to the world's research literature is not enough. We need to tell people that the stuff is there, welcome them in, show them a map, and maybe warn them about some of the trickier corners where they might get lost or fall down. Just who does this, or how, I don't know. But shouldn't we be talking about it?"