Election Time Is Poll Time

Background: Originally published September 15, 2004. This column is admittedly outdated for the times, at least in regards to local polls. Television stations used to do "dial-a-vote" surveys, in which they'd ask viewers to call a certain telephone number to vote yes or no to a particular question. Such local surveys have nearly vanished, as social media outlets such as Facebook and Twitter have replaced the means in which television stations gather information. Instead of surveying viewers about a particular question and broadcasting the results, stations now solicit information about any topic and sometimes share the results with viewers (weather photos, cute pet pictures, feedback to stories, etc.)

     "The only poll that matters is the one held on election day."
     That statement is often quoted by political candidates when asked how they feel regarding a poll that's just been released. The closer to election day, the more the public is bombarded with polls.
     Professor Jim Lemert taught "Polls and Public Opinion," a senior-level course at the University of Oregon School of Journalism. The course title may not elicit much enthusiasm, and consequently, neither did the subject matter among some students. But I nonetheless considered it a course well worth taking.
     Scientific polls were the only "real" polls that we were concerned with in class. Scientific polls are supposed to be an actual barometer of what the public is thinking at a given point in time.
     This is in contrast to non-scientific polls, which can be stacked or tainted by special interest groups. Newspapers that solicit reader input over the internet or television stations that conduct "dial-a-vote" surveys are examples of non-scientific polls.
     Another variation of the non-scientific poll is the "M.O.S./P.O.S." (man on street/person on street) surveys conducted by broadcast stations. People on the street may be asked a question at random, but there's no guarantee the sampling isn't tainted. For instance, if the POS poll is conducted on the sidewalk next to the Democrat or the Republican convention, obviously the sampling of respondents won't be representative of the American public at large.
     Meanwhile, scientific polls use a variety of criteria not only in determining WHO is asked the question, but also in HOW the question is asked.
     The least expensive way to conduct a scientific poll is over the telephone. In this method, a random selection of phone numbers is compiled with a specific pattern. (example: 555-1111, 555-1114, 555-1117, etc.) One funny anecdote shared by Professor Lemert in class, was about an irate person who wanted to know how the pollster obtained his unlisted number. The pollster tried to explain the numbers were selected at random, which meant unlisted subscribers had the same chance of being contacted as listed numbers. But the respondent still couldn't understand why his unlisted number didn't "protect" him from pollsters.
     As you can see, polls are not always popular subjects of conversation. But they're something we'll always have around. As we, the public are inundated with multiple sources of information during election season, it's important to remember how some information may be more "accurate" than others.
     Scientific polls are pollster-driven, because the pollster contacts the respondent. They're more accurate because the sampling of people is more random. Non-scientific polls are respondent-driven, because people choose to participate in the poll. They're not as accurate because a person or special interest group can stack the poll with their viewpoint.