ABSTRACT
Anybody working professionally with data and statistics knows (or should know) that one should never deduce the existence of a cause and effect relationship between two variables just because they are statistically associated. Indeed, most professionals will almost without provocation provide the catch-phrase “Association is not causation.” Whilst this phrase is undoubtedly true it is a bit like writing “May contain traces of nuts” on all food packages; it protects the statisticians (or food producer), but also lessens the value of the conducted analysis (or food label). In this chapter we will explore when and how association can indeed be interpreted as causation in particular in the context of survival analysis.
