When I look into the 2016 presidential candidates, the comments below every news post seem to be biased (either apparently or potentially). I try not to think these data should be discredited because of the obvious political stance, but what can we do with them in any research? What do they reflect, for example?
In this particular online news report, when candidate number 3 is "mocked," a large number of users who support him flooded in and left comments. In this case their political is made exceptionally clear thanks to profile picture generator. Of all the 74 comments, 40 of them (excluding 8 by the same users) are by users whose profile pictures say loudly 'number 3.'
**And now the election is over (01/16); it turned out that candidate 3 got 12.84% of all votes. It seems that the comments, the threads, and posts on the internet neither correspond with the polls nor reflect the actual voting. Even though it may seem self-evident that the overwhelmingly supportive comments do not predict anything, I can't help wonder, then, what can we gain from them. (18 de enero)
http://news.ltn.com.tw/news/politics/breakingnews/1553460