Q.E.D. (quod erat demonstrandum) Is statistical demonstration scientific?

seminar from group Science, Reason and Faith.

Jesús López Fidalgo. Pamplona, 17 January 2017.

News

Jesús López Fidalgo is Full Professor of Statistics at the ICS; he has made extended stays of research and teaching at the Universities of Manchester, Glasgow, UCLA and California at Riverside; Associate Professor of the Universidad Nacional de Colombia in Medellín; vice-president of the Sociedad de Estadística e research Operativa and publisher-head of the journal Test. He is also an elected member of high school International Statistics. He has been manager of the National Mathematics Plan in the Ministry of Science and Innovation (2009-2011) and director of the School of Industrial Engineering of the UCLM (2008-2016). He has published more than 100 articles in scientific journals. Author of the book of knowledge dissemination "el azar no existe". He has directed thirteen doctoral thesis and coordinates a node of the network national biostatistics.

summary:

The scientific research , among many other things, is deductive and inductive. Deduction is used in all disciplines. Its demonstrations, irrefutable, usually end with a q.e.d. On the other hand, in the so-called experimental sciences, induction is more and more essential. Modern statistics has come to give mathematical consistency to induction by means of the laws of chance. This seminar is intended to show how it does so, using the natural processes of the human mind to draw conclusions from simple data . It will address the topic of possible manipulation at all stages of the research, from planning the collection of data to making decisions from the interpretation of the results. Some of today's hot topics such as causal inference, Bayesian procedures, ethical-statistical issues, the treatment of massive data , the statistical detection of fraudulent data or the difficulty of making predictions when human freedom comes into play will be addressed with simple examples.