Intentionally lying or not, I simply have no confidence in them, whether they're for or against my guy. The pollsters rely on landline telephones and the available respondents are less and less representative of the full-spectrum electorate. The error is not all on the side of the Dems, there are large Dem demographics that don't have landlines who go unsampled as well. I think the pollsters assume the omitted/undersampled on each side cancel out but the rate of change in communications tech and usage likely invalidates that assumption since any statistical basis for it would be drawn from the 2000 census, 8 years ago.