Scott Rasmussen's bias is so obvious that reading his poll questions alone makes you realize it. His bias stems from the fact that he is a conservative ideologue.
On the other side of the coin, we have Research 2000. I think they are the Left's Rasmussen. I have not had the occasion to examine whether Research 2000 push-polls people. What I have seen is bias in a different form. Research 2000 does not do job approval during President Obama's tenure. Instead, it publishes "favorability" ratings; that is, how "favorable" or "unfavorable" Obama is in the eyes of the people.
What's the difference? You might ask. Here's how Obama's average favorability numbers look like:
Compare to Obama's job approval average:
These statistics are from Pollster.com. As you can see, favorability runs 6% higher than job approval, and "unfavorable" runs 8% more negative than "disapproval."
My argument would be neutralized if it were proved that Research 2000 used the same method with regard to George W. Bush or Bill Clinton. Internet searches don't reach as far back as the Clinton years, and I do not even know whether Research 2000 even exist back then. So let's stick to the Bush years. Research 2000 seems to have conducted very few polls on Bush. The one I found is referenced in a Daily Kos diary in July, 2008 (written by Markos). Excerpt:
KOS: For one, Bush is at 22 percent approval rating. No one likes him or fears him. Well, except Congressional Democrats, of course, which is probably why they're at 13 percent approval. Meanwhile, Latinos flock to Obama 65-24. Too bad Latinos won't vote for the black guy, huh? By the time this is all over, Obama will have won them at least 70-30. Mark my words.
As you can see, Daily Kos/Research 2000 trusted a "job approval" system when Bush was President. His numbers would have been crappy anyway, but the larger point is that independent pollsters without conflict of interests are much more reliable, in my view.
If I were conducting polls for the Daily Kos, I would want to make the website happy; and the bigger Obama's approval, favorability, or whatever it is looks, the happier the Daily Kos is. I'm not saying Research 2000 does this. Just stating what I would do in order to satisfy my client.
*note: Research 2000 had Coakley and Brown (Massachussets Senate) tied shortly before the elections, at a time that several other polls had Brown ahead. Brown won by 5%.