The fact is, more Democrats receive handouts. That also means that they are more likely to be available to the pollers who are trying to contact them.
It’s kind of like the unemployment numbers. They are calculated from questionaires sent out to a very small sample of people (0.02% of the population). So, who’s more likely to respond? The long term unemployed who are suffering from depression because they have been out of work so long that they are no longer receiving benefits or people in a better mood?