Comment: A Fruitful Resolution to Simpson’s Paradox via Multiresolution Inference