The link to the actual study appears to be broken. It's interesting that they've found the largest effect by far in people who drink the equivalent of 3 bottles

The link to the actual study appears to be broken. It's interesting that they've found the largest effect by far in people who drink the equivalent of 3 bottles of weak beer per week. It would be good to know how many people in the study fell in this category. If a small percentage then it could be a statistical fallacy due to small samples having larger variances.