Scientists aren't very scientific

Science has granted us many things, across the world it can be attributed to our increase in wealth, living standards, longevity, and technological prowess. Science accomplishes this by systematically trying to understand the universe, with testable theories and predictions. Science is certainly an accomplishment to be proud of, but in many way scientists are not being scientific. In fact, many of their actions to further their careers may be impeding science.

In the rather bleakly titled The Death of Science by Andrew Holster, the institutions of science are bemoaned, cursing them as sources of greed, ambition, competition and bureaucracy. Whilst Holster seemingly nit-picks everything to do with modern science, and I disagree with him forcefully on many things, he may have some points. With increasing numbers of students looking to pursue the scientific career, but without comparative increases in the number of academic positions, and progressively scarce funding, science has indeed become very competitive.

Competition is not an evil unto itself, it should surely ensure that the best and the brightest are the ones that get the jobs and the money. In fact, by allowing competition in science we’re making sure that science is criticised – I’ll be much more likely to criticise one of my rivals than one my friends. These lines of thinking are certainly the ones that are employed, but as Fang and Casadevall argue, they may be misguided. Competition has made the marker of a good scientist not good science necessarily, but prestige. What does this mean? It means that scientists may not be inclined to do the best research, but the research that will get them the most notice. Through prestige scientists can progress in their careers and get job security.

In a competitive environment, scientists will vie to become the first to publish their bit of research and before they do they will keep it a secret from their peers. Moreover, scientists may actively block those who they see as threats to their particular niche, and there have been associations found between increased competition for funds and misconduct. Indeed, Eric Poehlman, a disgraced scientist serving 12 years in prison for falsifying data, claimed that competition, a desire to pay people’s salaries, and defining self-worth by grant income were driving forces for his crimes.

Scientific misconduct is fortunately rare, but secrecy and political manoeuvring are sadly not. These lead to anti-collaborative efforts and can erode the quality of science conducted. Secrecy can lead to a duplication of effort with many labs working independently toward the same goal, rather than together. Some of the greatest feats in science can be linked to collaborative efforts, rather than competition; consider the wonderful collaborative efforts surrounding the discovery of the Higgs-Boson. The value attached to prestige, also means that many papers that are produced are flawed, and some are just plain wrong.

In a paper, dripping with sarcasm, Higgison and Munafò describe a model which shows the best ‘strategy’ for a scientist. They find that to advance in their careers scientists should focus on ‘low-hanging fruit’: easy, exploratory, small studies with rubbish statistical power that are bound to show ‘statistically significant’ differences. They should also not bother trying to replicate studies, as these are less highly prized. This is despite replication being one of the backbones of science. Worryingly the paper actually concludes that the model shows ~50% of all scientific research is just plain incorrect, which is close to the Open Science Foundation’s estimates of error in psychology.

If I could boil down the problem to one word, it would have to be ‘incentives’. Incentives in science are just plain wrong, whether if it’s for career advancement or for grant money, the focus on novelty and priority are forcing science to become a competitive enterprise and, as such, erodes the quality of science. How can we then move forward? We need to change how we do science, incentivising collaborative efforts and ‘repeat’ studies. Higgison and Munafò describe how just a small reduction in the weight given to novel findings could increase the value of research. Focusing on common goals rather than individual ones could help forge collaborative efforts, scientists should be recognised not just for their publication record, but for their generosity and coordination of research efforts. I also think that more focus should be put on methodology and analysis of results before scientists conduct studies, perhaps creating a register for studies that can be open for critique and making participation contingent on grant income. Similar to what is being tried for clinical trials.

I’ve probably sounded quite derisive of science, but that’s not my intention. My reason for writing this is I love science and I want to help it. Science is designed for such criticism as well, out of all human activities there is nothing as critical of itself and as (hopefully) self-correcting. The problems I’ve outlined are being recognized, indeed there’s now a ‘Parasite’ award for replicating papers, and it is my hope that they will be addressed. Science can only be as perfect as those who practice it, and humans will never be perfect, but we can at least try. I’ll leave you with this quote from Brian Nosek:

Science is wrong about everything, but you can trust it more than anything.