As of last Friday, Barack Obama was leading John McCain by 3 points in Gallup's Daily Poll, 48% to 45%. Even after a Republican convention the media deemed a "success," McCain seemed to experience no significant post-convention "bump" in his poll numbers as is often the case. However, as if by miracle, McCain has come out of the weekend with an 6 point swing in the poll numbers with him now leading Obama 48% to 45% in Gallup's latest daily poll. It would seem as if McCain is "surging," right?
Well, consider me a sceptic. I've never been a fan of poll numbers and they are generally as accurate as flipping a coin when it come to predicting elections. However, what bothers me the most is that they seem primarily to be "yellow journalism" tools for the media. If the media can get you to believe that the race is closer than most people think, it believes in can get more people to tune in to its coverage. Is it just a case of wanting more "eyes" for advertising or is it a blatant attempt by the media to influence the election by getting more people to view "slanted" coverage? Rather than attempt to answer that question, view this video by Noam Chomsky and decide for yourself:
While the polls would have you believe that the election between John McCain and Barack Obama is close, the electoral college polls tell a different story. If all of the states that voted Democrat for the last four elections vote Democrat again (a very strong likelihood), then Obama already has 248 electoral votes, just 22 short of the 270 needed to win the election. Obama also has strong leads in polls within several battleground states with significant numbers of electoral votes. The reality is that this election is very much Obama's to lose.
Why place more stock in the electoral college polls? Simple... their numbers are compiled in ways that more closely reflect how we actually vote for our President. What is to keep Gallup or any other polling company from skewing their national polling numbers simply by calling more people in the states most likely to vote for a particular candidate? Why not call more "red" states after the Republican National Convention or more "blue" states after the Democratic National Convention? It's very easy to manipulate national poll numbers in that fashion. Maybe it's done to increase drama, maybe it's done to actually influence the outcome of the election. But most national polling numbers are more than likely inaccurate... after all, Al Gore had more total votes nationally than George W. Bush but lost the election via the electoral college.
I'd place more stock in national poll numbers if they included the LOCATIONS represented in their sampling data. My guess is that those samples are skewed to particular locations to yield specific results. In any case, electoral college polls are a much better gauge of the ACTUAL state of the current Presidential election because the LOCATION represented in the sampling data is a known quantity and more accurately reflects how that state will impact the electoral college vote.