[amu]

Now it is 3-3, it seems like the poll has not helped us one bit! However, one may say that if the result of a vote is 50-50 that may mean that one result is more likely to have occured than the other. Are there anyone with good statistics skills that have occured a phenomenon like this?

If we were in court, we have to assume she is innocent. In this case the reasonable thing is to assume that the fart did not occur. So according to the vote, there is not overwhelming "evidence" that the fart did occur. So if we use this principle we can not reject the hypothesis that she did not fart.