I think he's trying to say the stronger statement that not only does correlation != causation, but he also believes there are known confounding factors that weren't accounted for.