Political Polls Undergoing Massive Changes
BY BUDDY NEVINS
Political polls are undergoing their biggest change in decades.
Stymied by millions who are unplugging their landline telephones, pollsters are scrambling to find new ways to survey voters.
Some are using the Internet.
Others are using more cell phone calls.
Pollsters claim both are accurate.
In the next few years, there will be a massive transformation in the way political polls are conducted — the greatest since pollsters switched to telephones from going door-to-door decades ago.
Wireless technology is the trigger.
I sat in on a class last week on polling at the University of Florida, part of the graduate program in political campaigning. In the class of 12 students, not one had a landline.
No surprise. Neither of my sons – aged 25 and 32 and holding professional jobs – have landlines.
According to the Centers For Disease Control, roughly 40 percent of the adults in the United States lack landlines. That figure is growing by millions each year, especially among the younger voters.
Over 60 percent of those under 29 live in homes without any landline, compared with only 10 percent of those over 65.
This leads to an obvious problem for political pollsters.
Using Cell Phones
Some, like the New York Times, fold cell phone number into the diminishing number of landline numbers when they do polls. This has its own set of problems.
Cell phone numbers can be harder and more expensive to match with registered voters. Some voters carry disposable cell phones that are much harder to survey.
And many users resent being called on their mobile device because they can be busy outside of their homes. They refuse to answer blind cell calls.
Internet Polls
Another relatively new method is Internet surveys.
Saint Leo University, near Zephyrhills north of Tampa, is making a name for itself with Internet polls.
Using large online panels, the university randomly selects a cross section of all demographic groups. Surveys are then sent to those selected.
Jim Kitchens, a long time Florida pollster who now conducts Internet surveys, uses more than 20 million email addresses to build the panels. He combines the e-mail addresses with names, addresses, census data, and other public documents. He then randomly picks a smaller group in a geographic area to poll.
One advantage is that those responding can answer the surveys when and where they want.
Kitchens said that a large number of the surveys are answered early in the morning before work, during the lunch hour and late at night before bed.
Both Kitchens and Saint Leo “incentivize” those answering the poll. Saint Leo deposits 50 cents into the iTunes or Amazon accounts of participants.
Are the Internet polls accurate?
Kitchens and Saint Leo believe they are. Others are not so sure.
One pollster, who asked that his name not be used, believed Internet surveys would have more participants who have time on their hands. He also thought that less underprivileged would be surveyed because they have little access to the Internet.
Candidates and consultants can decide for themselves. Before throwing away money on an inaccurate poll, they need to ask how the survey is being done and be comfortable with the methodology.
My take on all this: Kitchens and Saint Leo are on the cutting edge of tomorrow. Internet polling is the future
That’s fine as far as I’m concerned. I’m sick of being interrupted during dinner by calls from pollsters.
April 8th, 2014 at 11:12 am
It doesn’t matter what the polls say. They said Mitt was going to win–so much for that. What does matter is that the GOP will never win a general unless they change their views on women, gays, abortion, contraception, etc. The only reason they will win local–is because of redistricting. And those lines should be re-examined. With glasses…..
April 8th, 2014 at 12:13 pm
Buddy,
This was a great post. Thank you.
As Jim Kane and other honest pollsters will tell you, though, the much bigger “dirty little secret” of polling besides reaching a relatively large sample size cheaply and accurately is the nature of generating a random sample. Randomly selecting a sample is absolutely essential to being able to generalize from a poll to the population being studied (which is obviously the whole point of polling). If you don’t have a random sample of 1,000 people out of a population of whatever, all your poll tells you is about those 1,000 people.
Randomness has been getting harder and harder year over year. And frankly, I don’t believe ANYBODY has figured out the best way to generate a random sample yet without being cost-prohibitive.
I know I haven’t.
Kevin.
April 8th, 2014 at 3:49 pm
These pollsters who think just dialing cell phone numbers will work apparently haven’t heard about the zero-cost Google Voice internet telephone service. As people ditch their pricey calling plans and use WiFi & tablets instead, cell phones will either have to sharply cut their cost structure or die.
April 8th, 2014 at 4:42 pm
There is truth in the old cliche that the only poll that counts is on election day.
April 9th, 2014 at 4:49 pm
Isn’t Kitchens a Democratic pollster who finds the Democrats are ahead in every race? Has he ever predictedaccurately?
April 9th, 2014 at 6:02 pm
Jim has been Judy sterns in house pollster for years, looking at how few races she has won since she lost Israel in 2008 not all that impressive.
Jim was also Judy’s puppet in the bs law suit against Israel for a supposed unpaid bill Judy set up to make sure scott would never run again. Lol that didn’t work out well ether.
Looks like since 08 karma has been a bitch.
April 10th, 2014 at 9:35 am
Wasn’t Roger Stone the mastermind of the “Draft Trump”? Wasn’t he the one telling Trump to go all out on the birther movement? Wasn’t he the one who attempted to rationalize it by telling the press that Trump was raising a very legitimate question, that Trump was making no allegations but simply asking questions, that republican voters agreed with Trump, that the birther movement was “a brilliant political strategy”, that the media immediately attempts to marginalize anyone raising questions about the presidents birth place,and that Trumps poll numbers would rise and that he would do well in republican primaries?
Same Roger Stone.. right?