With just five days to go until election, poll aggregates are showing Mitt Romney slightly ahead in the presidential race, while several hot-topic propositions in California are locked in tight battles are well.
Real Clear Politics, which averages dozens of nationwide polls, is showing Romney up by 0.2 points as of Wednesday afternoon. Huffington Post, which is tracking 560 polls nationwide, is showing Romney up by 0.6 points.
Proposition 30, the widely-debated initiative that could potentially raise taxes and student tuition statewide depending on the outcome, is too close to call at this point, according to AroundTheCapitol.com, which averaged out five different polls, found 49 percent of those polled will vote yes, with 41.5 percent voting no and 8.6 percent being undecided on the issue.
Proposition 38, the competing initiative to Proposition 30, shows that 51 percent of those polled will vote no, 38.8 percent voting yes and 11 percent undecided.
Proposition 32, if passed, would prohibit the payroll deductions unions and corporations use for political contributions. Polls shows that 42.2 percent will vote yes and 45.8 percent will vote no.
Although these statistics may seem useless, Matthew Jarvis, Ph.D., associate professor of political science at Cal State Fullerton, said he believes them to be valuable.
“The way a poll works is you call a 1,000- ish people usually at random, and that random sample of people should on average look decently like the population,” said Jarvis. “It wouldn’t be identical, but it’s close enough. You should within about three points on a typical question on American politics.
“There is a 3-percent margin of sampling error. Any given poll should be within that 3-percent range,” he added.
Due to the random nature of polls, several things factor in that could throw off the results of any particular one.
Scott Spitzer, Ph.D., associate professor of political science, explained how the sampling of a population can be unreliable.
“There are groups in the population that are sometimes under-sampled; in other words we don’t get enough of them in the polls to accurately say with some statistical certainty what they’re feeling, so that can throw off a poll also,” said Spitzer.
Jarvis explained how to compensate for this—by polling companies weighing their statistics.
“They say, ‘I talked to more white people than I should’ve, or fewer Latino people than I should’ve. So I’m going to count Latinos as if they look like the population percentage.’ They rely heavily on the census which gives them a really good estimate of what the race, age, gender and education breakdowns of a population are and skew the sample in that direction,” he said.
Thus, the science behind the sampling and polling methods are completely up to the companies doing the survey, though both Jarvis and Spitzer agreed on the accuracy of these polls.
Jarvis, who kept statistics of polls taken by The Field Poll from 1996 to 2010, pointed out several trends among polls that account for the accuracy of polling in general.
“Generally speaking, if a proposition was down, it didn’t pass. Three of the 40 that were down passed, but the other 37 failed,” Jarvis said. “Just about everything that had more than a four point lead passed. Only two of the 47 propositions that were up by more than 4 points in the polls failed. One of them was up by 10 points and failed, and the other was up by 22 points and failed.”
According to Jarvis’ statistics, in the last 15 years, the most a proposition has come from behind (as reflected in the polls) to end up winning was from 17 points down. Jarvis said this was one of the times where polls were simply off.
“I think the Field Poll was off, but the likely voter model was off too,” he said. “One out of every 20 polls will be wrong. The sampling is just off—randomness happens.”
Kayla Coriaty, the chief governmental officer for Associated Students Inc., noted polls on Proposition 30 can be off.
“You don’t know who they’re polling, you don’t know what the populations are being polled, and although a lot of them are random it could be the luck of the draw. For example, the polls on Proposition 30, a lot of students might not be getting polled,” said Coriaty.
While new polls on the presidential candidates are regularly released, and will be up until election day, the polls for the propositions are less numerous and less frequent. Spitzer said this, combined with voters unfamiliar with the propositions, makes the polls on propositions leading up to election day slightly more subject to scrutiny.
“A lot of the most intense campaigning for propositions is starting now. This weekend is going to be a big deal,” Spitzer said. “Particularly with propositions, where people have not fully tuned in, people are going to start making up their minds and be more influenced by campaigning.”
This is no fault of the polls, however.
“A lot of these polls can be wrong because you actually changed your mind—it’s not the fault of the polls. People are not going to really change their mind on the presidential race last minute. The propositions on the other hand are really uncertain,” Jarvis said.
Polls can have positive and negative effects on viewers, depending on the numbers.
“I think it can be a motivating factor in making sure I get out and vote on the issues that I’m passionate about if they’re not doing as well in the polls,” Coriaty said. “I think there are a lot of people who say how something is polling 90 percent and wonder how their vote is not going to matter.”
To those voters, Spitzer said: “Vote based on your own conscience. Don’t vote based on what you think everyone else is going to do; vote based on what you think is the best decision you would like to see. I think it’s really important that people not look at the polls and get discouraged.”