|
August 7, 2000
|
The Public Speaks, in Numbers
| Gallup-ing Along |
|
According to the Gallup Organization, one of the nation's best known polling groups, Texas Governor George W. Bush is on a roll in the race for president. A Gallup survey of likely voters on July 25 and 26 found that Bush leads his rival, Vice President Al Gore, by a 54% to 40% margin. Less than two weeks earlier, Bush led by only 48% to 46%. But not so fast! Political experts point to a Bush "bounce" leading up to last week's Republican convention in Philadelphia, Pennsylvania. A similar bounce may benefit candidate Gore later this month when the Democrats convene in Los Angeles, California. The lesson here is that a candidate's popularity in the polls may vary greatly from week to week. Below, you can get a longer view of how Bush and Gore have fared in Gallup polls:
Many organizations poll the voters, from news outlets to public interest groups to the candidates themselves. Their hope is to predict accurately how the entire population will vote, based on interviews with only a small percentage of that population. According to Gallup, which has been in the polling business for much of the last century, a poll of any kind is only as good as the sample selected to interview. The key, says Gallup, "is a fundamental principle called equal probability of selection, which states that if every member of a population has an equal probability of being selected in a sample, then the sample will be representative of the population." How accurately has Gallup predicted the outcome of recent elections? In 1996, the final Gallup poll predicted a 52% victory for President Bill Clinton. Clinton actually won 49.2% of the vote. In 1992, Gallup predicted 49% of the vote for Clinton, but he received just 43%. That was enough for victory in a three-candidate race.
|
|
While presidential candidates George W. Bush and Al Gore will be making plenty of speeches in the coming months, the American public will be talking back in numbers—the numbers of national polls. What can these polls tell us?
|
| Getting the Numbers Right |
|
The researchers conducting the poll need to include enough people in a sample to predict how the entire 170-million adult population in the United States will vote. Gallup and other major polling organizations use samples of about 1,000 people. The results they get will not absolutely predict how Americans will vote, but the so-called margin of error will be limited to + or - 3%. For instance, if 42% of voters in the poll say they will choose a particular candidate, there is a high probability that somewhere between 39% (42 - 3) and 45% (42 + 3) of all Americans will vote the same way.
If the sample size increases to 2,000, the margin of error decreases to + or - 2%. In the above example, the larger sample size would mean that between 46% and 50% of the voting public would be highly likely to choose Bush. Between 44% and 48% would be highly likely to choose Gore.
|
|
| Reading Other Polls |
|
The questions covered a range of topics including the following: |
|
Voters have also weighed in through other polls. Recently The New York Times and CBS News asked 953 adults to compare candidates Bush and Gore. This poll was different from the Gallup poll above in that the people responding did not have to choose one candidate over the other. They were asked the same questions about each candidate and could answer yes or no for each.
|
|
|||||||||||||||||||||||||||
|
|
|
|
|
|
|