On October 26th, 2000 the Gallup Poll showed George W. Bush leading Al Gore by 13 points among what they identified as 851 likely voters. That was less than two weeks before election day, when Gore ended up besting Bush in the popular vote by some 540,000. The aggregated data from 2000 is available at PollingReport.com here:
http://www.pollingreport.com/wh2gen1.htm
More cool history and analysis in extended text...
That Polling Report historical data makes for interesting reading. Overall, most of the major polls showed Bush leading throughout the fall by anywhere from 2-13 points. The general consensus among the media was that Gore was a terrible and unlikeable candidate who would probably lose. Moreover, Republicans were seen, in the wake of the Clinton scandals, as more motivated and energized. That was the conventional wisdom, and the polls only fueled that perception of the race.
After the 2000 election many pundits, most notably the Wall Street Journal's Daniel Heninger, concluded that the only way Gore closed the gap was that the Democrats had mounted an unprecedented "turnout miracle" as he put it. Karl Rove, grasping for an explanation to offer his angry boss, claimed that some "four million" evangelicals had failed to show up on election day. This seems rather dubious. Evangelicals tend to be organized, well-informed, and active in their communities. The idea that so many of them just didn't bother to vote is not credible. More likely the overall estimates on the number of evangelicals has been slightly inflated. As is only natural, most groups, including national gay and Jewish organizations, tend to favor and promote only the most optimistic estimates of their size. Evangelicals would be no different. Rove compared such statistics to the VNS exit poll data to make his claim.
In the four years since the 2000 election both parties have focused on turnout and GOTV efforts. Fearing that the campaign finance reform measures on party soft money would handicap the Democratic Party's traditional turnout operation, progressives established the 527 group America Coming Together (actforvictory.org). However, while ACT has been established and fielded an impressive organization this year, the DNC has found unexpected success in small-donor hard money contributions that it had previously never been good at. What this means is that not only will ACT be in the field, but the DNC will also be able to fully-fund its traditional GOTV operation, something it did not expect to be able to do when McCain-Feingold passed.
Lastly, a note on national telephone polls in the age of caller ID and cheap cellular. Two critically important Democratic leaning groups are now unavailable to pollsters: younger voters and the economically distressed. The rapidly increasing number of voters under 40 who have dropped land lines in favor of cellular-only communication are unreachable by pollsters. Lower income groups have likewise found it more convenient and sometimes also cheaper to have only a cell phone than to maintain a land-line. In addition, any voter who mistrusts unknown callers fearing a bill collector or a hospital collection agency is unavailable to these national poll samples. Other groups unavailable to evening telephone pollsters include those working two jobs, those who work night shifts, those who go to school at night, those who work late, those who are socially active at night, and those who live abroad.
So when conservatives refer to 2000 as the Democrats "turnout miracle" they really ought to call it the Democrats "turnout reality."