[gizmo686]

A large enough sample size can find statistical significance in anything.

Also, this is a question will well understood theoretical backing. If you consider the entire law as an experiment, it is probably best to view this as a preliminary report, not a final result. Similarly to how no one thinks twice when CERN gets excited about a bump in their data, only to find out later that it turned out to be a statistical fluke.

[cimmanom]

But there’s also an assumption here that 3 months is long enough to have a measurable impact, rather than that quarter’s results being largely dictated by decisions made earlier.