I would add two things. The small samplings are not worrisome as there is statistical analysis to determine the accuracy of such small sample sizes. That is what gives rise to the plus or minus number after each poll.
In addition, many of the pundits take an average of several of the more trusted polls, thus significantly enhancing their statistical relevance.
Much ballyhoo is made of "Polls". The pundits love them. Personally, I think they are based on what the pollsters want to hear. Some of the 'pollees' outright lied about their answers when interviewed later. During the 50 odd years of my working life, I have never been 'polled' about anything. For that matter, no one I have ever talked to during that time has been 'polled' either. Who the hell are these people that get asked the questions? Anyone here ever been a participant in a poll? Do the same 1,500 people routinely get asked all the time? Is it a man-on-the-street thing, do the pollsters go to the farm towns of the mid-west, or only the major metropolitan areas? Done by phone or in person?
I don't know how I get selected.
So, it's likely you haven't been polled yet.
They have to chose the people somehow at random. The procedure has been to generate lists randomly from telephone directories. That has probably become obsolete due to the number of people who use cellphones and have disconnected their landlines.
Cellphones aren't listed in directories, and that makes the polls biased now, though both parties think they have the lead among cellphone users.
Even the people taking the polls want to know truthfully how good their candidate is doing, but there might be some unconscious bias in polls.
As for liars, it is presumed that they are evenly distributed across the political sides, so that they cancel each other out.
I would add two things. The small samplings are not worrisome as there is statistical analysis to determine the accuracy of such small sample sizes. That is what gives rise to the plus or minus number after each poll.
In addition, many of the pundits take an average of several of the more trusted polls, thus significantly enhancing their statistical relevance.
between the polling instruments.
Polling is an interesting exercise.... and I oft wonder if it provides the self-fulfilling prophecy - that is - lets say candidate A - by an artificial inaccurate or biased polling result, gets ahead by a significant # of points... over candidate B... Does Candidate B then loose because his/her supporters feel the actual voting exercise to be futile? and therefore fails to vote...
I wonder if this is also why third party candidates do so poorly... (Well, that and they are mostly nut jobs...)
Very interesting point. I have a number of family members living on the west coast. As exit polls are broadcast shortly after the east coast voting is completed, voters out west, depending on the exit polling data, have shown a propensity to not vote if they feel their vote will have no affect on the outcome of the election. What they fail to consider, however, is that exit polling is routinely flawed. Hence, "If it is not even close, why should I bother voting"? Exit polling should be banned.