Scott Alexander comments on Be Skeptical of Correlational Studies