Lies, Damn Lies, and Bad Statistics?
- PDF / 157,528 Bytes
- 2 Pages / 595.276 x 790.866 pts Page_size
- 88 Downloads / 175 Views
BRIEF COMMUNICATION
Lies, Damn Lies, and Bad Statistics? Italo Braghetto 1
&
Manuel Figueroa 1
# Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract A wide spectrum of research such as experimental, randomized trials, cohort or epidemiological studies, technical or control case reports, systematic reviews, and meta-analyses has resulted in a huge amount of publications. These studies and publications may be subject to errors due to poor application of statistical tests, which can lead to misinformation, misinterpretation, and erroneous conclusions, sometimes even considered as lies. In this article, some ideas about this issue are discussed in order to adopt new directions in the future and thus avoid lies and bad statistics. Keywords Publications . Statistic
A wide spectrum of research such as experimental, randomized trials, cohort or epidemiological studies, technical reports, or control/case reports, systematic reviews, and metaanalyses results in a huge amount of publications that require statistical support in order to be validated. Different requirements such as materials, number of patients included, methodology, and of course, statistical analyses are needed in order to achieve the desired results. Thus, knowledge, expertise, capability, and availability to perform these statistical analyses are needed. However, these are not always readily available and hence, we must ask for help. Surgeons have ideas, experience, knowledge, casuistry, and frequently, new questions. These questions lead to new hypotheses, which must then be proved through different methods of investigation. Currently, there is an enormous quantity of publications and a hypertrophic production of papers due to the “publish or perish” imperative. This occurs because different and controversial points of view arise that are not always deemed acceptable. Some are considered lies and even worse, as “damn lies.” Generally speaking, surgeons are motivated through personal interest to answer questions, share results, and experiences; however, confusion or mistakes do occasionally appear. Although they may be due to poor methodology or poor statistical application, I believe that
* Italo Braghetto [email protected] 1
Department of Surgery, Hospital “Dr José J. Aguirre”, Faculty of Medicine, University of Chile, Santos Dumont 999, Santiago, Chile
surgeons are reliable and responsible and that these do not occur for ill-intended reasons. So, what are the possible causes? What are the underlying problems? Lack of expertise? Lack of knowledge about statistical analyses? We must trust our colleagues and believe that there are no intended lies, just honest mistakes. Why do these mistakes occur? There are probably many causes such as poor hypotheses, conceptual flaws, poor experimental design, wrong methodology, wrong performance, subject variability, heterogeneity, incorrect statistical planning, or simply a lack of expertise and knowledge. These mistakes could conduce to a lack of reproducibility of the results t
Data Loading...
