I can't remember where I saw the research either (it was recent though, maybe on Hacker News), but they did check casuation/temporal effects by following people over time and seeing that when one person became sad, people within a few social links of them were more likely to become sad, etc.