Looking at the details [1], it seems to be a small (N=180, then N=152) correlational study. How can we tell whether the effect is causal?
One way would be to run a randomized controlled trial, where you show some people slightly happier or sadder Facebook streams and you see how they react. This experiment has, in fact been run, [2] and what they found was that seeing happy status updates seemed to make people happier, not sadder. [3]
[3] Though what they actually measured was what people posted, not their happiness. Perhaps seeing happy things leads you to be sad but post happy things to fit in.
One way would be to run a randomized controlled trial, where you show some people slightly happier or sadder Facebook streams and you see how they react. This experiment has, in fact been run, [2] and what they found was that seeing happy status updates seemed to make people happier, not sadder. [3]
[1] http://guilfordjournals.com/doi/abs/10.1521/jscp.2014.33.8.7...
[2] http://www.pnas.org/content/111/24/8788.full
[3] Though what they actually measured was what people posted, not their happiness. Perhaps seeing happy things leads you to be sad but post happy things to fit in.