For polls not backed by statistics it's fine for the media to report results. If the poll is random, and anyone can answer as many times as they like, then it's not something that should be reported in the media. There's really nothing we can learn from uncontrolled polls if they're not somewhat regulated.
If the news media can cover essentially any topic, why should covering the results of online debate polls be any different? I think that the American population that actually votes can decipher the metrics of polls and have a clear understand of which are valid and which are not, so there's really not an issue with it.
Online debate polls are often stuffed by people from all over the globe, when in reality people from outside of the US have no real bearing on the importance of the debate. They are not conducted in a very scientific manner and do not give people a good cross section of what the results really are.
No, the news media should not report the results because many of these polls could be taken by people who did not even watch the debate. Plus, many of the people who voted could vote many times in a row. Therefore, the polls are not an accurate representation of the outcome.