Are political polls biased and why?

Great question! I think it's tough to avoid bias in the type, number, and phrasing of the questions (if you even try). 2nd, it's very difficult to target a truly random representative sample, because I think busier people who value their time more, so will be less likely to complete the poll, when contacted.

The problem by and large with political polls is the way the information is presented. Take the current presidential polls. Statistically, if I take a large enough random sample and present the choices in nonbiased manner (non-leading question phrasing, intonating my preference for the poll taker, counter balancing the answer choices, etc.) I should be able to accurately reflect the mood of the general population with the data that I have collected. However, it should come as no surprise that the major political polls are showing Trump and Clinton both well into the 40%'s. If I present you the question "are you more likely to vote for Donald Trump or Hillary Clinton" then you are forced into an A or B answer psychologically. If I ask you "in this election who do you plan on voting for?" Then you may say Gary Johnson, Jill Stein, or Mickey Mouse.

Not to sound like a conspiracy theorist (because it isn't a conspiracy that the "two parties" do not want to lose power) but if Gary Johnson is polling above 20% in some polls conducted at state levels and Stein polling at or near 8% in polls conducted at state levels...why do the polls that determine the debate stage have Johnson polling >9% and Stein polling at >5%?

A poll done properly can provide some insight. However, polls can very easily be manipulated and I believe the polls usually show the result that the people paying for them want it to show.