Graduate Student Seminar - James McVittie
Title: Why Bother With Statistics?
Abstract:
A simple Google search defines the term statistics as “the practice or science of collecting and analyzing numerical data in large quantities, especially for the purpose of inferring proportions in a whole from those in a representative sample”. In general, this definition is correct, however, it does not seem to capture the subtleties hidden in modelling real world phenomena. In 1976, George Box stated ‘Since all models are wrong the scientist cannot obtain a “correct” one by excessive elaboration’ which simplifies to the common aphorism of “All models are wrong but some are useful”. Given the above definition of statistics and the statement of Box, how can the statistician bridge the gap between collected data and well-defined mathematical models which are potentially incorrect? In this talk, we discuss how statistics is introduced at an undergraduate university level and how seemingly basic concepts can be generalized using rigorous mathematics. Time permitting, some current research problems will be introduced.