Fostering Productive Statistical Skepticism