The Atlantic reports:
Robert Nemiroff and Teresa Wilson, two researchers at Michigan Technological University, thought they might. In a study released online last week, the two scoured Twitter, Facebook, Google+ and a few other websites to find “prescient information”—that is, tweets and statuses about current events posted before the events became current. The only way someone could write such a post, they reasoned, is if they were visiting… from the future. [...]
The study, alas, turned up no time travelers. But that doesn’t quite mean anything. The authors admit that the study might have failed for many reasons: Time travelers might not have the ability to physically adjust the past; they might not have posted about the events the authors were looking for; they might have posted about the events but not turned up in a search. Time travelers might have also read the study or this news story about it, and been sure to making avoid any careless mistakes.
You can also read the full paper, which hasn’t been peer reviewed yet, on Arxiv
For more on the difficulties in using social media for academic research see my recent Wired piece on the lack of peer review of social data.