ZenGiraffe
This is vulgar, but I love it
No it isn'tthat's why you poll more people, to iron out the variables!

An unrepresentative sample is unrepresentative regardless of size.
There is so much more to polling than just asking as many people as you can find how they will vote. There is checking how the characteristics of the people in your sample compare to the population as a whole (eg. On Gender or race). How likely are people to vote? What variables do you use to predict likely voting behaviour.
A bigger sample is better than a smaller sample all else being equal, yes, but you can't just point at the poll with the biggest sample size and say that it's the best.
The problem with polls herding is that it suggest the pollsters are doing something to make their poll more like the others, usually by tweaking with their weightings.