Observing open access

Share this on social media:

As a quantum theorist by training, I was taught early in my career that observation is not a passive act and that by measuring a physical system you cause it to change.  

It turns out that it is not just quantum systems that are sensitive to being measured – human systems also change when you measure them. By measuring a quantum system, the change that you make is profound, forcing the system into a definite quantum state that is determined by the nature of the measurement that you make. Measurements of human systems precipitate less instantaneous change than in the quantum case, but are no less profound in how they change the system under observation.

Measurement can be a powerful tool, since it can help bring about change. Equally, we must be careful about what we measure. Arguably the whole Responsible Metrics movement was born from the realisation that measurements are not necessarily used positively in the research system. A preoccupation with reductionism, ranking and overly simplistic approaches by all sorts of actors across the system can lead to unintended consequences. Here, we look at one case where measurements can be used as a positive force to help to change behaviours.

In a recent report from Digital Science, based on data from Dimensions, we argued that open access has progressed due to ongoing waves of innovation that force the field forward, rather than through a quiet measured attrition. In countries where there is less will for progress, and hence fewer drivers for change, we see less rapid development. We have seen waves of initiatives that are not linked to a specific geography and which benefited the development of open access globally. Typically, these are the creation of new platforms or new technologies, for example: the launch of arXiv in August 1991; the foundation of RePEc in 1997; the establishment of PLOS in 2000; and, the creation of PeerJ in 2013, to name just a few. At the other end of the scale, we have seen many institutional initiatives with investments in digital repositories, and the creation of institutional mandates being the most obvious.

However, it is not until more recently that we have seen the emergence of truly international efforts to coordinate OA. The emergence of Plan S, through cOAlition S, is the newest wave of innovation that seeks to break through the impasse that has developed in some countries. In our report, we observed the US had faltered in its progress toward increasing OA, levelling off at around 42 per cent of overall publication output through OA channels in both 2012 and 2016.  

Of course, it will be several years until Plan S is implemented and we still don’t know how it will finally be realised. However, with recent progress in Germany on Projekt DEAL and the new guidelines from REF 2021, it is clear that not only are both the UK and Germany travelling toward an open future, but also that, since those two countries are among the most collaborative in the world after the US, there are many countries who benefit from the stance taken by those with a progressive agenda.

The UK’s overall percentage of OA content has grown rapidly, outpacing both Germany and the US in recent years. As shown in Figure 1, the UK’s approach to policy around OA has paid dividends. This is not to say that this hasn’t taken significant sustained investment and resource – but innovation seldom comes for free. The spearheads of innovation that have pushed the UK forward include the Finch Report, delivered in 2012, through to the announcement that REF 2021 outputs would need to be made available through OA channels in 2014.

As you might expect, those universities with more research are more exposed to policy innovations around OA. In 2000, 43.3 per cent of the UK’s OA output was contributed by the Russell Group universities.  By 2016, this number had grown to 58.6 per cent. Even within these top universities, a natural stratification emerges with UCL leading the pack with 13.8 per cent of the Russell Group’s output (translating to around 8 per cent of the UK’s total OA output); followed by Oxford, Imperial and Cambridge producing between 9 per cent and 11 per cent of the Russell Group’s output. There is then a further cluster of Manchester, King’s, Edinburgh and Bristol each accounting for 6 to 7 per cent of the Russell Group’s output. 

Finally, there is a less stable grouping that includes Southampton, Leeds, Sheffield and Nottingham, all contributing around 5 per cent in 2016 but which also may be considered to extend to Glasgow and Birmingham. This latter grouping is less stable than the upper groupings, which see less change with time. All these institutions are publishing more than 60 per cent and, in many cases, close to 70 per cent of their output through OA channels.

Measurement of open access through government mandate is clearly driving behaviour in the UK’s research institutions in a way that is not so evident either on the continent or in the US. The natural competitiveness in the Russell Group is mixing with the impending REF 2021 exercise to be the newest driver of OA in the UK. Institutions, seeking to ensure that a maximal number of amount of outputs for the return, are strongly incentivised to increase their proportion of OA outputs.  

We began this piece by observing that measurement drives behaviour, and that there can be unintended consequences. In this case, the obvious effect is that the research-intensive institutions that are members of the Russell Group are more susceptible to this type of measurement than most of the UK’s research institutions, since such large portions of their funding depend on Research England’s block funding. However, what we see is an increase in the overall percentage of the UK’s research output being centred around Russell Group institutions – at 58.6 per cent in 2016, up more than 15 per cent since the turn of the millennium. This indicates that the ability to pay for open access at scale appears to be centralising in Russell Group institutions, with others being left behind the curve.  

In 2021, this will inevitably have an impact on the choices that smaller institutions can make regarding their REF returns, the results of those returns, and the potential funding balance going forward. Research England is clearly not insensitive to these challenges, as it has included a number of options for institutions to argue for exceptions and include a percentage of non-OA outputs. However, the direction is clear: open access will form part of the REF for the first time and it has been thoroughly embedded in the most recent guidance on submission, panel criteria and working methods published by Research England. This is a strong signal to the community and a strong ‘measurement’ that pushes the sector toward open access.  

Achieving change forces us to embrace new initiatives that renew our progress. Nonetheless, measurement remains a tool to be used with care. It will be interesting to return to the data in years to come to quantify the large-scale effects of these interventions and initiatives.

Daniel Hook is CEO at Digital Science