Modeling dependence beyond correlation

As scientists, looking for relationships between variables is an essential part of our efforts to understand the world, identify the causes of illnesses, assess climate change, predict economic cycles or guard against catastrophic events such as tsunamis or financial crises. Statistical models are often used for this purpose. To many scientists and engineers, this is synonymous with correlation and regression because these techniques are typically the first, if not the only ones, to which they were exposed. However, thanks in part to the work of Canadian statisticians, we now know that there is much more to dependence than correlation, regression, and the omnipresent “bell-shaped curve” of basic statistics textbooks. As we will demonstrate in this chapter, those who are oblivious to new tools may miss important messages hidden in their data.