If the population you wish to represent is the people of Earth, certainly not... And numbers isn't the issue... Consider that there are all sorts of reasons that people on DDO might not represent a fair sample: access to computers, access to internet, literacy, etc.
But even if we're only interested in people on DDO... Well, then a very popular debate that's on the front page for a while might well accrue enough votes to be representative of DDO's average views... But what use would that be? You don't even have the stats of where in the World the DDO membership comes from, so it's hard to see what use that would be.
I don't believe the sample size the site is allowing is big enough to be considered the general opinion of people. For example on most of these debates I have gone through there is only 3 or 4 voters so saying that 100% of voters agree with the question would mathematically be correct yet I think that they should give ratios instead of the percentage. If 1 person were to vote "yes" on a particular question and someone used that for an essay or for proof to something that would be based on a 1 person survey. I believe for this to be accurate it should be a random sample of at least 10 to be considered the slightest bit accurate.