Alexandra Freeman is Executive Director of the Winton Centre for Risk and Evidence Communication at the University of Cambridge

Science has a problem.

It’s one caused by conservativism in a field that many people might assume was constantly pushing the boundaries and exploiting the latest technology. And to get to the heart of the problem, we need to consider what science really is.

Part of the solution?

Put aside the stereotypical images of laboratories and pipettes beloved of picture editors – at its heart science is an approach to problem solving.

Much of human – and other animals’ – problem solving is a matter of trial and error. We try, we watch, we learn, all by practical experience. What we might call true science, though, is to take a step beyond that. It is to try to understand the underlying generalisable principles that govern why things happen the way they do – and hence to make predictions about what might happen under conditions that have never yet been experienced. These predictions allow human imagination and creativity to stretch more readily beyond personal experience, to invent things based on the foreseen not simply the seen.

For example, for centuries people experimented with substances that could make explosions, mixing up different chemicals and heating them in sealed containers to see how to make bigger bangs. All of that was trial and error. The science came in when the theory of chemistry and the physical gas laws came in to allow predictions about what sort of chemical compositions would create maximum gas expansion. It is understanding the underlying theory that allows the most finely controlled explosions, such as those that create exactly the right amount of gas, at exactly the right time, inside a car airbag.

As part of that move away from purely personal experience and towards a more theoretical underpinning, though, science demands a more collective approach. It requires the sharing of knowledge, ideas and experience. Because to understand the principles that govern the world around us we need to look for patterns – to see past the chaotic ripples of the everyday to search for underlying stability beneath the surface.

Stable genius

What never changes? Whilst some of the laws of physics are relatively unobscured to our perception and allow a single person to take measurements themselves and understand the principles (as many of the early scientists of the 17th and 18th century did), working to understand the complexities of the climate, ecology, the human mind or the behaviour of society requires more skills, experience, data and imagination than a single person can possess. It requires a collective effort.

This is where science’s current problem lies. The scientific establishment developed in centuries where communication was either done by direct speech or the printed word sent from hand to hand. Experience and results could be written down in a few pages of numbers and letters. The specialisms required to carry out each part of the scientific process – from developing ideas to designing experiments, carrying them out, analysing and interpreting the findings – could feasibly all be gained by one person, with perhaps a little assistance. The system of communication and sharing of scientific knowledge that grew up within that establishment was that of ‘journals’: magazines which people could subscribe to for a fee and in which authors would write about their thoughts and experiments in the form of ‘papers’.

Now there are over 33,000 different journals in the English language alone, publishing around 3 million scientific papers each year – a number that is growing by about 5% a year.

Paperless science

These “papers” are no longer serving science well. It’s not just that it is virtually impossible to find and read all the relevant work within this vast, disconnected mass of poorly-searchable (and expensive) material. Nor is it just that experimental data is rarely easily contained within a printed form and so if often only summarised, preventing others from fully using it or checking the conclusions of the authors. The problem is much deeper.

This is because the very purpose of publishing “papers” has changed. Now, instead of being mainly a means of communication of fact, papers are the main measure of success for a scientific researcher – determining their future career prospects. This, naturally, demands a different kind of communication style, and a different kind of content.

Who wants to publish ideas that didn’t turn out to be supported by data? Perhaps it’s not surprising that a suspiciously high 85% of papers claim to show data supporting their authors’ hypothesis. But we need to know the unadulterated truth – to learn from the “error” bit of trial and error.

Who wants to publish long explanations of exactly what they did to obtain data? Not only is it tedious to write, and uses up article space, but maybe it will encourage rivals. Yet without it, how can others interpret the results, to know how they may or may not be generalisable? Who wants to share anything except a beautiful, polished account of their idea and how they proved that they were right all along? After all, that is the kind of story we all like to sell about ourselves on our public-facing CVs to potential employers or patrons.

A good story

This pressure to create a good story even influences what and how research is done in the first place: if the aim is to create a good “paper”, a good story, then the harder, more specific, slower (or even more thorough) repetitive checking is just not worth a researcher’s time or a funder’s investment.

This is a far cry from the original premise of science.

In order to make progress in scientific research in the complex areas we now tackle, researchers need to work together more and more. And in the twenty-first century, digital-first world we have the tools to make that increasingly easy. We can bring information together, search it, collate it, link it in multi-dimensional ways, translate it automatically from one language to another, and all at minimal cost. By doing this we can allow specialists to collaborate with other specialists across time and space. And, most importantly, we can incentivise good science rather than simply selling a good story. Since the main driving force behind the problems of the current system is that the metrics of success have become divorced from measures of good science, our new twenty-first century system of scientific communication must realign them in order to drive ambitious researchers towards good and useful practices.

The kraken awakes

This might all sound a pipe-dream, but in fact such a system is almost laughably easy to construct. It is far from the hardest problem that scientific and engineering minds have had to crack, and in fact over the past year a mostly volunteer force has put together the technical basis of it, an entirely free system called Octopus, which should be ready for release early in 2020.

The problem, then, is not a technical one but purely a social one. The scientific establishment has become deeply conservative, because so many people’s money and career prospects ride on every aspect of it. Individual researchers feel they cannot change the way they publish because they are caught in a system that will penalise anyone who doesn’t play the game.

I designed Octopus to allow researchers to use it and publish in it without jeopardising their ability to publish in the traditional journal system in parallel, and to make it worth their while to do so. But to truly allow the switch to happen will take a political change, from institutions, funders and science-based industry. Social change can happen quickly, and there is already a perception that there are problems with the current publishing system. What I don’t think is realised is how deep those problems are, and how much they are holding back scientific effort.

With the world facing unprecedented challenges in terms of preserving ecological and species diversity whilst maintaining a healthy and happy human population, the need for more efficient, more collaborative and more meritocratic scientific research has never been greater. We can only achieve that if we change the way we share scientific knowledge. To my mind, there is no single thing that we can do in the world that would have such a huge impact – and yet it is far from the top of any agenda.

2 responses to “The journal publication system betrays the purpose of science

  1. I wholeheartedly agree with everything in this article. It’s absurd to imagine that the best way of sharing and communicating science in the 19th century remains the best way today.

  2. It seems that the development of Octopus has slowed down because of the coronavirus pandemic. This idea is very promising; I hope the slowdown from the pandemic won’t kill off the momentum for this project.

Leave a Reply