In one poll, president Obama is up a point over Mitt Romney. In another poll, the lead is 6-points.
In one poll, likely voters are sampled. In another poll, all adults are sampled. So, how do we know which polls are accurate?
Toledo-based pollster Fritz Wenzel works for candidates all across the country, most famously, presidential candidate Ron Paul.
Wenzel says the political conventions are over and the candidate debates are 3- weeks away so polls are helping fill the void of presidential news.
Voters have noticed the dozens of polls recently released, including Suzanne Browning, who told 13abc reporter Bill Hormann, "I do kind of sometimes like to hear what other people are thinking."
But what people are thinking may not be accurately reflected in the polls.
Wenzel says "You get vastly different results depending on what sample you choose to use.'
He says two big factors affect poll results. One factor involves the sample population.
Wenzel says polls can sample a variety of publics. "Do you have a registered voter sample, a likely voter sample or just all adults sample," he wonders. Wenzel insists using a likely voters sample is more accurate because likely voters are more engaged in the political process, have reasoned opinions and will likely vote.
But including the other samples in the poll can skew results. For instance, and ABC News/Washington Post poll sampled likely voters and found the president had a 1-point lead at 49-48.
Yet among registered voters, Obama had a 6-point lead, 50-44.
To Wenzel, this means one sample is better than another. "Here, close to an election, everyone should be using a likely voter sample."
Another problem with polls: sampling more Democrats than Republicans.
Voters sometimes sense when a poll doesn't seem right. Sherri Jacobs tells 13abc, "I just don't believe in them. I think (the pollsters) call people they know who are going to go one way or the other."
Wenzel says that polling companies do sometimes skew the poll by not accurately sampling the public. "Democrats are more likely , our surveys show, to answer a political poll than are Republicans," Wenzel says. "They're easier to reach and they're more willing to talk."
For instance, a recent CNN poll uses *likely voters* and found the president leading Mitt Romney by 6-points, 52-46.
But the poll *over-sampled* Democrats and vastly *under-estimated* Independents.
A more accurate sample would have had *Romney up* 8-points, 53-45.
Wenzel says voters need to investigate the polls they're monitoring. "Always check to see that the Democrats to Republican to Independent are properly adjusted so that you have a poll that has the right proportion."
Despite the various ways polls can be manipulated, Wenzel believes today's polls *are more accurate*. But still, voters are left wondering what to believe.
When asked if opinion polls matter to her, Melissa Slaske told us, "No. My opinion matters to me."
Fritz Wenzel says voters need to do their homework. He says likely voters make for more reliable polls and you need to know whether the Democrat- Republican- Independent sample is roughly 35-percent each.
He also says polls can be skewed by gender, age and race.
And by the way, that Margin of Error? Two or three points probably means an accurate poll. Five points or higher should raise an eyebrow.
All content © Copyright 2000 - 2013 WorldNow and WTVG. All Rights Reserved.
