Evaluating causality in quasi-experiments tends to be extremely difficult and so far we have only one such study that concerns social media directly.