#IDIBELLseminars: How to lie with statistics to a scientist

Beyond qualitative research, all modern science is quantitative. Depending on study design, reporting results is guided by STROBE, CONSORT and other EQUATOR guidelines, aimed to enhancing the quality and transparency of health research, available elsewhere https://www.equator-network.org/reporting-guidelines/.

However, we need to understand that both mathematics and statistics can be tweaked, as they both can be presented and interpreted in many different ways. It can be shocking to many that the objectivity of presenting data in tables can be masked with figures, more attractive to the eye, but misinterpreted by the brains. By posing figures with varying axes, scales, thresholds, and distributions, literally anyone with a vested interest can exaggerate or mask data, and deceive readers.

The classical book How to lie with statistics, written by Darrell Huff (a journalist, not a statistician) in 1954, can be revisited together with the more modern Thinking, Fast and Slow by psychologist Daniel Kahneman, who was awarded the 2002 Nobel Prize in Economics Sciences, and passed away last March. The later ‘s main thesis is a differentiation between two modes of thought: «System 1» is fast, instinctive and emotional; «System 2» is slower, deliberative, and more logical. When reading papers, both systems cannot be taken for granted, otherwise we misuse results.

Revisiting some published examples may serve as food for thought to scientists, from PhD candidates to experienced line managers, especially on the best/most correct way to present data objectively. It is often said: “it takes one thief to catch another thief”. Well, by knowing how to lie with statistics, particularly in papers in predatory journals from windmill editorials, many  with soft or even absent peer review, we can better present our own data in the right way.

Hosted by Cristina Martínez – Tobacco control research group

Dr. Joan B. Soriano works at the Dept of Respiratory Medicine of Hospital Universitario de la Princesa, and he is Honorary Professor of Medicine at Universidad Autónoma de Madrid, in Madrid, and Associate Professor of Medicine at Universitat de les Illes Balears, in Mallorca. Joan belongs to the CIBERES team led by Prof. Esteve Fernández at ICO.
He has 500+ publications in PubMed and 10+ book chapters in the fields of clinical epidemiology and treatment of respiratory and tobacco-related disease, and a SCOPUS Hirsh index of 102 with 90,100+ individual citations. In May 2011 he received the Josep Trueta Award for scientific and medical achievements, in 2014 he was appointed Foundational Fellow of the ERS and Fellow of Chest, and in 2022 received the NeumoMadrid Award for his Scientific and Research Career.
Just like many other, since February 2020 Joan suffered a “covidization” of his research. Up to July 2021, he served as Senior Consultant of the COVID-19 Clinical Management Team, at the headquarters of the World Health Organization, based in Geneva, Switzerland, helping with the updates of COVID-19 WHO’s Clinical Guidelines and defining Long COVID. Currently he is a Consultant for chronic respiratory disease for WHO-Europe in Copenhagen, Denmark.
Beyond a life-long interest in asthma, COPD and tobacco control research, Joan is active in COVID-19-related research, leading several research studies and clinical trials on Long Covid.