The author of this article acknowledges the possibility that the polls could just be wrong, but gets tripped up when the Republican polls seem to be more accurate than the Democratic polls -- and then concludes that the only explanation must be fraud.

I think what's really going on is that older voters and younger voters behave differently, affecting the accuracy of the polls for one set of voters more than the other.