So the assumption is that in this broad cross section, a contruction worker represents the interests of ALL contruction workers, a soldier a soldier, a Doctor a Doctor, a housewife a housewife? That is why the data in flawed from the start in such a low number of people questioned, because it makes assumptions based on the person(s) taking the poll. In order for this poll to be accurate it has to contain a much larger number of people to have a true reflection as to the feelings of most Americans. Let me cite you an example, I can during an election cycle poll 895 people and come out with an opinion that will tell me mickey mouse will be the next president of the United States if I question the right people and use a low enough number of people to represent the intentions of ALL the voters. In short this poll is flawed based on the following, the poll sample represents the intentions of exactly .00000385ths of Americans and even if you used a factor of 20 or 19 the data is still flawed because the poll sample number is too low. This poll while interesting is meaningless, because it represents the interests of less than 1% of the people that need, want, or have healthcare.
Most national polls have a sample of around 1,000 people that are accurate within 3% of the sample and a confidence interval of 95%. In English, that means given any specific poll at any given time, we expect it to be accurate within 3% 19 times out of 20.
Let's look at the last Presidential election. The final results for the popular vote was Obama 53%, McCain 46%.
RealClearPolitics - Election 2008 - General Election: McCain vs. Obama
So, the confidence interval tells us we would expect Obama to receive 50% to 56% of the vote 19 times out of 20 and McCain to receive between 49% and 43% 19 times out of 20.
What actually happened in the election? Here are the polls on the last days of the election.
RealClearPolitics - Election 2008 - Latest Polls
There were 15 polls on the last weekend of the election. Every poll had Obama between 50% and 55%. Every poll had McCain between 48% and 42%. So the polls were pretty accurate even though the typical poll had about 1,000 respondents.
If you averaged out the polls, Obama was at 52.1% but received 52.9% while McCain was at 44.5% and received 45.6%. We would expect the average to be closer given that the 15 polls account for a population sample of around 20,000 (out of 300 million).
So, yeah, 1,000 people is usually a pretty fair assessment of the national mood. Political parties don't pay pollsters millions of dollars for nothing. There are many reasons to be skeptical about this poll, but generally, the methodological construction is probably not one of them.