> The data from extremistan will more likely follow a power law distribution.

That's exactly the kind of logical inference that the linked article warns against, unless you have some additional reasoning behind that assertion.