Also: Facebook is a very incomplete and often inaccurate view of our lives. I assume that most regular users create their Facebook timeline as an idealized, curated mental facsimile of their lives, filtering out real inadequacies, exaggerating stories that might make them look more wealthy, busy, free, etc.

Perhaps they're accounting for this but I'm not really a fan of doing social science based on a data set that's more fantasy than scientific measurement.

I agree i hope at least they use more data sources for the study. Personally, i'm too lazy to update my information like my current job and so on, so my data would be totally wrong.

It depends on what questions you're trying to answer.

My intuition tells me that predicting socioeconomic status with Facebook's data would probably be reasonably successful, and you could look at feature importance from that. If they're trying to test every theory that Munoz came up with in 4 years of brainstorming or w/e though, then I think they'll likely find it's an insufficiently complete dataset.