[Ole Okie]

Now I’m not defending anything here, I’m just stating that there is nothing inherently wrong with a sample size of 160.

Under ideal circumstances, perhaps.

But this was 161 respondents in 8 states. And the WaPo projected individual battleground state results based on an average of 20 respondents per state.

Bah and humbug!! The published "poll results" are meaningless.

But they served their purpose. Now a majority of Americans believe the election is irretrievably lost for Romney.

[RoseofTexas]

NOT SO FAST grasshopper! 60% of folk polled have said the liberal mediaWHORES are extremely in the tank for Obama...BY them virtual quoting polls day in and day out knowing those polls are slanted to benefit their evil master and depress our side in order to carry the BLACK TURD over the finish line ONLY TICKS OFF the BASE and I assure you we will be CRAWLING THROUGH BROKEN GLASS, WALK ACROSS RED HOT COALS to get to the polls on Nov. 6!! AND TO TOP IT OFF we will be taking family folk who have NOT voted in past elections WHO are itching to get to the polls. THE AMERICAN FOLK who love this country AREN’T blinded by the evil demonic communist agenda driven MEDIAwhores ONE bit!!

[USFRIENDINVICTORIA]

You're 100% correct, with no margin of error.

If the 161 residents were drawn from a single state, then the 8% margin of error would apply. (Although, that wouldn't mean there weren't many other things wrong with the poll.)

However, given the electoral college system, a sample cannot have the same margin of error, if drawn from 8 states. If the sample size in each state was 20, that yields a margin of error of 22%!

You can check the margin of error for any sample size, using this handy calculator:

http://americanresearchgroup.com/moe.html