Almost every statistics instructor emphasizes to
their classes that a strong association does not
necessarily imply causation. How researchers could
draw causal interpretations from the results of
their studies remains a challenging yet interesting
question. First, the question concerning how
potential causal factors can be identified prior to
modeling is even more important than making causal
interpretations after the analysis. This book aims
to evaluate the claim that automated data mining
constitutes a paradigm shift in causal discovery.
The data mining approach overlooks how...