Obviously that was wrong. The suspicion that a good many of these polls were crap and biased was right.
Tell me how a Pew poll has Obama up 15 two weeks in a row and then, for their last poll, they say “just kidding, his lead is really only 6, not 15”. Now they brag how accurate they were. Rasmussen didn’t play those kind of games.
I have learned over time to believe in the AGGREGATE of ALL polls. Put ALL polls in the hat, regardless whether they look good or bad for our side. Then tally them all up and divide by the number of polls measured. More often than not you get your final result, basically eliminating margin of error.
While Rasmussen was right on with their popular vote count, their individual state polls weren’t as good (i.e. Pennsylvania.) But, look at the aggregate tally from some of the aggregation sites like RCP or pollster.com, and you get an incredibly close picture to what ended up happening. Take pollster.com actual vs. predicted as an example:
http://www.pollster.com/blogs/actual_vs_predicted_electoral.php
Maybe this and 2006 can serve as a lesson for many here. Come 2010 take ALL polls, good and bad (chances are some are too far in our corner, some are too far in theirs) and average them out to come up with the most likely result (barring any major external crisis or event.)
That attitude hurt us in some meaningful ways. For instance, Pennsylvania was never close, and anecdotals “proving” that Obama actually had no chance to win the state, despite what the polls were showing us, were maddening, and potentially damaging to our chances. McCain obviously started drinking the kool-aid about Pennsylvania, too. He sunk a ton of money (about 1/4th of his ad budget for most of October) and time into a state that looked FURTHER away from us than Michigan, but polls were to be shunned, not believed. Ignore them, the whole lot of them. Meanwhile Virginia, Florida, Ohio, North Carolina lacked those funds and were allowed to flip. Huge, huge blunder, based on blind reliance on the myth that ALL polls are inherently false or have a partisan agenda, that has been too prevalent on our side and had even found its way into the campaign. The end result: 55% to 44%.
Also, look at 2004. The one poll closest to the end result was the aggregate of them all, see for yourself:
http://www.realclearpolitics.com/bush_vs_kerry.html
Don’t shun ALL polls. Don’t believe whacked out polls too generous to their side, but also don’t JUST believe those closest to our side (chances are they are too generous to us.) Take them ALL and use the average.