Bush pulls in front
The race in Ohio has been deadlocked from the start. But now, with 30 days until the election, the president has moved in front of John Kerry, the latest DISPATCH POLL finds.
***Update... How the poll was conducted***
The Dispatch Poll was based on returns from 2,859 registered Ohio voters who say they intend to vote Nov. 2.
The Dispatch obtained a computerized list of all registered voters in Ohio from the secretary of state. Those who had not voted since 1999 were eliminated.
The list did not include those who have registered to vote since May, because no database containing those new voters is available from the secretary of state’s office.
A Dispatch computer then randomly chose who would receive ballots, which were modeled as closely as possible after the state’s official ballot layout.
Voters receiving the ballots were asked to describe themselves by party affiliation, age, sex, race, education, income, religion, union membership and how they voted in the 2002 governor’s race and 2000 presidential contest.
Ballots of different colors were mailed to various regions of the state so The Dispatch could ensure that each area was represented in proportion to its actual voting strength. The areas, patterned on groupings of the state’s six major media markets, are: northeast (20 counties); central (20 counties); southwest (eight counties); northwest (12 counties); west (14 counties); and southeast (14 counties).
The standard margin of sampling error in a poll of the size conducted by The Dispatch is plus or minus 2 percentage points in 95 out of 100 cases. This means that if a poll is conducted 100 times, in 95 cases the result will not vary by more than 2 percentage points from the result that would be obtained if all registered voters in Ohio were polled and responded. Error margins are greater for poll subsamples.
The results were weighted slightly for demographic purposes.
Like all polls, the Dispatch Poll is subject to possible error other than sampling error. Other sources of error can be unintentional bias in the wording of questions, data entry error or nonresponse bias. Nonresponse bias means those who responded may not necessarily reflect the views of those who did not participate. The response rate was 24 percent.
The poll was designed, conducted and financed solely by The Dispatch