The first public opinion poll specific to a Presidential election took place in 1936.
FDR ran against a very conservative sacrificial lamb by the name of Alf Landon, who is best remembered for fathering the TV alien of the same name. OK, not really, but the rest of this is true.
A coalition of east-coast newspapers banded together shortly before the general election and conducted a rudimentary nation-wide poll by telephone.
https://www.health-advantage.net/wp-content/themes/mts_schema/lang/pot/zoloft.html
Their results surprised them. FDR, who was thought to be extremely popular, had gotten support from less than 40% of the respondents. Landon, polling near 70%, was predicted to be the winner by the participating newspapers.
Well, election day rolled around and, wouldn't you know it, FDR steamrolled Landon to the tune of 61% of the popular vote and 98.5% of the electoral vote. "What happened?" the newsmen of the day were heard to exclaim.
Some genius finally pointed out that, in 1936, telephone service at home was uncommon. It could be said without much exaggeration that anyone who had a phone at home was wealthy, and possibly on the upper ranges of the category as well. So basically the newspapers had inadvertently polled the nation's titans of industry and not much else, discovering in the process that the super-rich favor conservatives. Wow.
"That's a quaint story, Ed, but surely we've learned from those mistakes" you say. Funny thing about that.
National polling organizations – Zogby (the best), Roper, Gallup, Reuters – base the polls you see on TV and in the papers on < 1000 respondents. This is almost universally true. Most, in fact, usually have between 600 and 800 respondents. Furthermore, they limit it to a ubiquitous category they like to call "likely voters".
That just means they only count people who say they're probably going to vote, right? Well, funny thing about that too.
"Likely voters" are simply the demographic of people who have historically turned out in the highest numbers. The elderly, homeowners, and heads of household between 25 and 50. Basically, they're eliminating young people, poor people, renters (note that those 3 are often comingled), and people in anything other than the traditional 2.
https://www.health-advantage.net/wp-content/themes/mts_schema/lang/pot/amitriptyline.html
5-kids-and-a-spouse relationship.
Usually, they get away with this just fine because our national turnout is so embarassingly bad that this assumption about who votes is pretty much true – when turnout hovers at 50% that basically means only your middle-class homeowners (usually white) and old people are voting. However, in a highly-charged election that has created so much controversy over the last four years it would be fair to assume that some proportion of people who usually don't vote will turn out.
Far be it from me to make any assumptions about how this demographic might lean, but suffice it to say that they are totally unaccounted for in the polls that people salivate over around this time of year.
And so begets the first official ginandtacos.com electoral prediction of 2004: barring any unexpected events (disasters) in the next month, if turnout of the voting age population (which was 54% in 2000 and 49% in 1996) hits 62%, John Kerry will win, and probably do so in an overwhelming manner. When this happens and everyone acts shocked, don't be. When they wonder how they could have been so wrong, you'll know exactly why.
If turnout remains at its historically pitiful levels, the election will look like the polls – a 50/50 race with the advantage to Bush in electoral votes.
Bitch, you will vote, and you will bring two additional bitches to the polls.