Between these two firms, you see a representative sample of the general polling averages as you would see on RCP. For the most part, the polling averages underestimate the part that will win most of these races. Generally the margins were larger, though they did (overall) pick most of the correct winners in each race.
My point is, that PPP is not a firm that overestimates Democrats as an overall rule. Actually, between these two cycles, in just the Senate, PPP overestimated Republican candidates more on average. PPP, while a little more to the left than Rasmussen, is still to the right of the end result.
As a last note, both of these firms did better in the Presidential vote. While both overestimated Romney, they did by lesser margins than they overestimated the GOP in Senate races.
Lesson: PPP polls should not be just thrown out as "partisan trash."