Think of it this way... in my previous example 300 people voted for the winning selection. I will spare you the math, but if hundreds of thousands of people play SWTOR, 300 is a fraction of a percent. We can't sway entire decisions based on just that poll. However, people look to polls to justify decisions and unfortunately things just aren't that simple.
Actually the issue is not sample size. 300 people is actually pretty close to the necessary sample size for a 500K population with a 5% margin of error. http://www.surveysystem.com/sscalc.htm
for a handy dandy calculator. The issue is representative population. When a proper survey is done relevant demographic information is used to map the sample to the population being meaured. Taking a ridiculous extreme, surveying 300 people at random asking Which character class in SWTOR is currently OP, with a true random sample would likely yield a 99% response of "what the **** is SWTOR", because a majority off the world population doesn't play SWTOR. By definition any poll in the forums is not hitting the general demographic of SWTOR as a majority of players don't read the forums. (I of course base this on the anecdotal evidence of the players that I know who play SWTOR and the frequency which they read the forums. which is not a statistically valid mechanism
) The forum readers will tend to be the more "hard core" users.
Now if you want to qualify and say that x% of the SWTOR players who visit the forums believe .... then a 300 person sample size of self selected participants would be sufficient.