You will know what are numbers of people that voted, not of people that visited the site. And results of every voluntary vote are skewed, as group voting is not identical with the group researched.
That's very basic sociological/statistical knowledge. Selection of representative group for research is very difficult. In the case of forums it is possible (at least in theory) to calculate statistics not for selected group of users, but for whole population. Such results are always most precise, but also usually too expensive to do.