Down to the Wire They Come…To a Nail-Biting Finish

ROCHESTER, NY, NOVEMBER 7, 2000—This time it really is a horse race. The final Harris Interactive Election 2000 Internet poll shows Al Gore and George Bush in a dead heat at 47% to 47%, with Ralph Nader at 4%, and other candidates, including Pat Buchanan, at 2%. Gore holds an incredibly narrow Electoral College lead of 271 to 256 with Wisconsin and its 11 Electoral College votes a toss-up. In races for the Senate, this study’s best projection is anywhere from 0 to 3 seat gains for the Democrats. Three states are literally toss-ups: Delaware, Missouri and Michigan. Several other races remain exceedingly close. Neither party appears likely to make significant gains in races for control of governorships.

The final Harris Interactive’s election poll was conducted from October 30 through November 6. A total of 240,666 likely voters participated in the study. National results are based on results from 50 states. In addition, this release provides specific results for 38 states. This survey comprises the largest online election study in history and may well be the largest polling effort ever conducted by a private organization.

Highlights of the Presidential Race

The chart below shows how Harris Interactive expects the Electoral College to turn out on Tuesday based on the final results reported in Table 1. As can be seen from the chart, several of the races that Harris Interactive’s calling are extremely close (under 3% spread between the candidates). These states, and the candidate they marginally favor, are:

For Gore: Florida, Minnesota, New Hampshire

For Bush: Missouri, Nevada, Ohio, Tennessee, Washington

Obviously, how well the campaigns execute their turnout strategies as well as the vagaries of the weather or last-minute breaking news events can still impact the outcome on election day. Assuming nothing unforeseen happens and the campaigns execute equally well, however, Al Gore is slightly more likely to emerge the winner on Tuesday night. And he may well do it without winning the popular vote.

The Nader Factor

In a race this close, Ralph Nader can hardly fail to have at least a potential impact. In the following states, Gore trails Bush by a margin that is less than this study’s projected vote for Nader: Colorado, Nevada, New Mexico, Ohio, Tenessee and Washington. And in Wisconsin, Gore would lead Bush if it were not for the Nader vote. Clearly, one of the key factors to look for tonight is just what percentage of the vote Nader gets in these key states.

Race for Senate Control

Like the race for the Presidency, the race for control of the Senate is likely to make for extended light night television viewing. The following chart shows how we think the races are most likely to turn out. As can be seen, many races are extremely close and could easily go either way on election day. Three Missouri, Michigan and Delaware are literally toss-ups. Among the other close races are:

• Florida, where Bill Nelson (D) leads Bill McCollum (R) 50% to 46%
• Washington, where Maria Cantwell (D) trails Slade Gorton (R) by two points (47% to 49%)
• Virginia, where Chuck Robb (D) trails George Allen (R) by 47% to 52%
• New York, where Hillary Clinton (D) leads Rick Lazio (R) by 51% to 46%.
• New Jersey, where Jim Corzine (D) leads Bob Franks (R) by 49% to 45%.

Table 2, below, provides specific projections for each state in the Harris Interactive Election 2000 survey.

The Governorships

There are 11 races for Governor in this year’s election. It now appears that neither party will make significant gains. Democrats are poised to retain the governor’s mansions in Indiana and North Carolina despite spirited Republican challenges but appear likely to lose in Missouri. On the other hand, Democrats may pick up the governor’s seat in West Virginia. Overall, little or no change is anticipated in the partisan makeup of statehouse occupants. The following chart gives an overview of how we expect the races to turn out. Results for specific states in the Harris Interactive Election 2000 poll are shown in Table 3, below.

Where to See Results

Methodology

This Harris Interactive Election 2000 study was conducted between October 30 and November 6 with a total sample of over 240,666 likely voters from the Harris Interactive Inc. panel of Internet users. State sample sizes range from 585 (in Hawaii) to 25,017 (in California). Data were weighted by age, sex, education, income, race/ethnicity, urbanicity, and the propensity to be online, in order to generalize the results to the population of likely voters in each state.

In theory with a randomly chosen sample of this size and after weighting the data, one could say with 95 percent certainty that the results have a statistical precision of approximately one to five percentage points of what they would be if the entire likely voter population of each state had been polled. The sample used by Harris Interactive is not a random sample. While individuals have been randomly sampled from the database for this survey, they have previously chosen to take part in the Harris Interactive database.

There are several other possible sources of error in all polls or surveys that are probably more serious than theoretical calculations of sampling error. These potential sources of error include question wording and question order, non-response, and screening (e.g. for likely voters). It is difficult or impossible to quantify the errors that may result from these factors so the words "margin of error" should be avoided when reporting all survey data.

These statements conform to the principles of disclosure of the National Council on Public Polls.

Click here to view all 3 tables.