Distinguished Speaker:
Dr. Frederi Viens
Full Professor, Department of Statistics
Member, Ken Kennedy Institute
Member, Rice Sustainability Institute
Rice University, Houston, Texas, USA
Title: Is it safe to use statistics to determine whether two environmental phenomena measured over time are related to each other?
Abstract: In 1926 Statistician G. Udny Yule showed that for two sequences of coin flips, their correlation has a tendency to be large, even when they are independent from each other. This phenomenon, which has since been recognized for many non-stationary time series common in observational studies, is in sharp contrast with the classical result for two sequences of independently sampled data points, by which the same correlation should be close to zero. Still, ignorance of this empirical fact, known today as Yule's nonsense correlation, has led practitioners to make dramatically ill-informed assertions about statistical associations. This improper use of methodology has occurred in recent times, e.g. for the critically important question of attribution in climate science. The mathematics behind this phenomenon comes from a classical theorem in probability and statistics. While technically challenging, the basic intuition behind this theorem is well within the grasp of the general public, as we hope to demonstrate in this presentation. We will also discuss the implications of a new mathematical discovery about these correlations, in practical testing for independence and for attribution in environmental time series, including applications to ecology, agronomy, and development economics in the US and in Africa.
Day & Time: Thursday, September 17, 2026, at 7:00pm
Location: Toldo, Room 102