The problem with the statistic you bring up is that it lacks context like what percentage of the general population eats dark chocolate. If only 20% of the population eats dark chocolate then it's a very interesting statistic.
When I look at the article I don't see any crimes against statistics, and I'd be curious what statistic you think is contextless or misrepresentative.
I find your complaint rather baffling: statistics is how we figure out what is going on in large populations when many variables are in play. How else would we advance medical science?
Yes, as the recent "reproducibility crisis" has shown us (http://blogs.discovermagazine.com/neuroskeptic/2015/11/10/re...), there is severe danger in playing with statistics, even mildly. But that does not mean we should stop using statistics altogether: it means we should come up with better protocols and procedures to prevent the biases we discover.
