Polls inaccurate, don’t represent voters

During the past few months, Americans have been inundated with the latest polls attempting to do what reasonable people realize is impossible: predict the future. A closer examination of such predictions reveals that the polls mean almost nothing for this presidential election.

Gallup, used by CNN and USA Today, is one of the most widely recognized and respected polling firms in the nation. It boasts a margin of error of plus or minus 2 percent, the lowest among the pollsters.

First, one must consider how respondents to polls are chosen. According to USA Today polling editor Jim Norman, Gallup sets quotas in each of the four regions of the United States, then calls random telephone numbers. Of the people polled, only the most likely voters’ responses are reported. These "likely voters" are chosen through a series of seven questions such as "Did you vote in the last election?" and "Do you know where your polling place is?" Based on respondents’ answers, pollsters rank them on a 0-7 scale and then report the top 50 percent as "likely voters."

There are several obvious problems with this system. To begin with, not every response is reported; logic dictates that the more responses, the more accurate the polled sample will be. When the increased intensity and polarity of the upcoming presidential election are taken into account, the "likely voter" in past elections may no longer apply.

Also, only the responses of those who answer random house phones are reported. Gallup does not poll cell phone users. Many students and members of the working class, who are far more likely to be Democrats than most other groups, use only cell phones. Hence, the Democratic vote may be drastically underrepresented in current polls, which have shown Bush in the lead in recent weeks.

Then there is the margin of error to consider. Plus or minus 2 percent may sound small, but when polls for this election have consistently shown one candidate leading the other by such a small percentage as 49-47, it’s really quite large. One candidate may actually get 51 percent of the vote while the other gets only 45 percent — a far larger difference than the polls would lead us to believe. Most polling firms have larger margins of error; plus or minus 4 percent is more common, so a "leading" candidate at 49-47 could even be down 45-51.

Online polls are even less accurate, as any one can vote as many times as they want. Politically charged sites show differences as drastic as Kerry leading Bush 70-30; most people realize this is a tight race and such polls are completely meaningless.

Consider the poll regarding the Democratic primary performed by the American Research Group, Inc. from January 2-4. This poll showed Howard Dean in the lead by far with 39 percent of the vote, while Kerry had a meager 14 percent. Clearly, something made this poll inaccurate, as Kerry had little difficulty capturing the Democratic nomination.

But in spite of the numerous fallacies of polls, many undecided Americans will turn to them as a guide and cast their vote for the same man they believe the majority of their fellow citizens will be voting for. Polls just might influence enough voters to make themselves accurate.

Chenin Simi is a first year public relations and Spanish major at CSULB.