Author: Mirte Golverdingen
Speaker: John P.A. Ioannidis
Department: Meta-Research Innovation Center at Stanford (METRICS), Department of Medicine, Stanford Prevention Research Center, Department of Health Research and Policy, Department of Statistics, Stanford University School of Humanities and Sciences, Stanford, California, United States of America
Subject: Making more scientific research true
Location: Erasmus University College
Date: 9 November 2015
In the week of 9 November the 102nd ‘Dies Natalis’ of the Erasmus University Rotterdam was celebrated. Because of this 102nd anniversary two persons got a honorary doctorate and one of those persons is professor John P.A. Ioannidis. He got the honorary doctorate because of the global orientation of his research and his broad and critical view on research in general. Because of this occasion professor Ioannidis gave an interactive lecture about two of his research articles.
In his research he tries to answer the question how do we communicate research and how can we do it the right way? He is not the only one who is busy with this research, a lot of different groups are investigating this question. They are looking at different topics that can be improved like: methods, reporting, reproducibility, evaluation and incentives. Each topic has its own difficulties and this concludes that it is very hard to make science better.
When scientist publish their articles 97% of them claims that they found something significant. This results in an abundance of information and we are not even sure if it is really significant. Although the percentage of significant results, due to the writers, slowly is decreasing it is still a lot. Do we need to change the scientific system so we can be sure that the published work is significant?
Articles can be labeled as insignificant when the study does a lot of work in an underpowered setting. This means that they have too less result to trust and therefore have a low statistical power. Moreover, a lot of articles have reproducibility issues. It is impossible to reproduce the same result. And when authors reanalyze their own articles they have to come up with a different result, otherwise it is not acceptable. This makes that re-analysis is used to make the original publication better.
How are we going to change this system? Ioannidis put forward a two new rewarding systems. You are more rewarded for doing peer review, for successfully translate publications and for contribute to education and training. You also lose rewarding for granting funding, see also table 1. Ioannidis thinks that this new rewarding system contribute to make more published research true.
However, are these changes in the system reliable? It is hard to say, but major stakeholders all have different looks at what is useable and what kind of paper you need. So some stakeholders will profit from this new system, while others prefer to stay at the old system. Yet the majority of research effort is still wasted and we know that changes in the system helps to make scientific research more successfully pursue its goals.
Tabel 1: An illustration of different exchange rates for various currencies and wealth items in research. (Source: Ioannidis, J. P. A. (2014). How to Make More Published Research True. PLoS Medicine, 11(10), e1001747. http://doi.org/10.1371/journal.pmed.1001747)
It is very important to stay focused on making scientific research better. More and more papers are published and we need to know if all this information is significant. To do this we can use different examples of rewarding systems. However, it still is very hard to change the way scientific research works.