Polling has lost the confidence of people across the country. More people believe in the local weather forecast than in political polls, especially in polls done by and for the media.
People in actual campaigns know the difference. When campaign professionals run polls, they want to know the truth about where their campaigns are in actuality, and how to craft their game plan to win. They demand their pollsters structure polls properly and get accurate results, and they know how to do it.
Then there are spin polls designed to propagate a public narrative. The pollsters know the results they want to see and they know the results will be made public. These polls depress the turnout of one side, depress fundraising of a certain candidate, make it hard to recruit volunteers – literally most of what a campaign does can be fueled or diminished by these spin polls.
These aren’t “push polls”. Push polls are polls where no one sees the results or even cares what the results are. Push polls are formalized rumor-mongering. “Would you be more or less likely to vote for Republican X if you knew bribed his business partner who then committed suicide.”
Spins polls are different because they are published and made to look legitimate. We can look at spin polls done for nefarious purposes and usually see how they stack the decks to get the results they want. In years past, it was done by slanting the questions or stacking one political party with an advantage.
Those days are gone. Those ways are too easy to debunk. There are specific things pollsters do today that are harder to detect, and most will never notice anything wrong until Election Day when the poll is off by fifteen points.
The 2020 election season was teeming with bad polls. In Maine, polls showed Sen. Susan Collins losing by five. She won by nine. Polls for weeks showed South Carolina’s Lindsey Graham even, and he won by over ten points.
Larry Sabato’s idiotic UVA “Crystal Ball” said Biden would win 321 electoral votes. He’s backtracking big time now.
These aren’t mistakes. You don’t get so much so wrong so often without doing so on purpose.
So how do they do it?
Including voters you want
With the data that is available, it’s easy to only contact voters who actually vote. If someone isn’t in the voter file, they don’t get polled. Simple right?
But that’s not what many pollsters do. They know the propensity of non-voters to vote for certain candidates, so if they want to pad the numbers for one side, they will contact voters who registered, but who haven’t been near a voting booth in years, or ever. Then they “ask” how likely they are to vote in the upcoming election.
People always tell pollsters they are going to vote. Pollsters know this. They can call someone who hasn’t voted in five years, and know that all they have to do is ask if they plan on voting, and they’ll say yes.
Just like when Nielson television ratings stopped using written viewer diaries where people recorded what they said they watched. When they’d ask people what they watched on TV, it was amazing how many said PBS. When they started measuring what people actually watched through meters, they sure weren’t watching PBS like they said they were.
Contact non-voters and ask them if they’re going to vote, and they will almost always say yes. You’re on your way to getting that call quota.
Any poll that says they contacted “registered voters” or “adults” isn’t worth publishing or reading.
Excluding voters you don’t want
This is how a pollster judges if they are polling “likely voters.” Want to exclude people who only vote in Presidential years? Then filter only respondents that have voted in two of the last four elections. Want to exclude people who only voted for Obama in 2008 and never showed up again until 2012? Use the “2 out of 4” rule. Want to screen out Trump voters who only show up for Trump? “2 out of 4” rules will do it.
Pollsters know that this tactic screens out many voters who will be voting in the Presidential election. Why would they do that? Because they don’t want them included. They want a slanted sample pool.
Another exclusionary tactic is using long surveys. Most people will not answer a 20-minute survey. They just won’t. Most people will hang up after 5 minutes or so because they just don’t have time. It’s dinner time or the kids need attention.
So who does sit and talk to pollsters for 20 minutes? Highly educated liberals. Want to debunk a poll in 5 seconds flat? Look at the percentage of college graduates in a poll. Roughly a third of American adults have a college degree according to most sources. I’ve seen polls regularly with 55, 65, even 70% of those surveyed with a college degree. It’s dreamland.
But they do know who it helps in the results.
One random poll I saw had Biden leading Trump 64-30 among college educated respondents. Trump won non-college graduates 54-34. So how do you make sure you poll lots of Biden voters? Make your surveys so long that highly educated voters will be thrilled to spend a half hour talking politics with you, and make a lot of working-class voters say “I have to go”.
It’s bias without showing bias. They can set up a poll to get the results they want and still claim innocence.
Finding the right Party split
Pollsters have gotten too smart to be debunked by the usual “Republican vs Democrat” split in a poll. If they show +13 Democrat in a poll demographic, they know they’ll be skewered. So they include the kind of Republicans they want: lots of highly educated Republicans, knowing they’ll check the box for Party ID but only support people like John Kasich or Joe Biden. Never a Donald Trump.
Bingo! A defensible poll that still veers off to Liberalville.
Another way to signal a biased poll is the income question. You’ll never know how rich your area is until you see the income percentages in a liberal poll. This goes hand in hand with the college bias, but when you bulk up the college-educated into your survey, you’ll get a pretty wealthy sample group.
The problem is that the people in line on Election Day don’t usually match the six figure incomes who made up the polling sample.
And Some People Lie
Pollsters know people lie, or at least fudge. Amazing how many people who vote for the same party year after year tell a pollster they are independent. Ask them if they are conservative or liberal and you won’t hear a peep. Everyone’s a moderate. Contact a voter who voted for Gore, Kerry, Obama, and both Clintons and they’ll tell the pollster they are moderate and independent.
It’s not because people want to be dishonest. It’s because they don’t want to be criticized. Some of the most liberal voters in the world will say they’re independent. People who vote Republican for decades will tell the stranger on the phone “they vote for the person, not the Party”.
Every elected official will tell you that after a campaign win, everyone they meet says they voted for them. Amazing how 51% becomes 100% in person!
Why does all this matter? Because in this age of data, you can have the voting history of people in polls in a snap of a finger. If someone votes in Democratic primaries year after year, a pollster could know without asking. Most pollsters won’t check.
Why? Because then their samples would actually be accurate.
Pollsters want an outcome, and they can only get there if lifelong Democrats say they are Independents or if conservative Republicans say they’re moderates.
And if all else fails, they can apply weighting to the categories they like to get the results they want. These weighted models are never publicized and never scrutinized, but they are used to make sure their sample is “representative”. If the model is wrong, the poll is wrong.
But they’ll never tell you. They’ll just say they need to do better next time.
Meanwhile, how many voters stay home, don’t contribute or decide whom to support when the pollsters tell them it will make no difference? Polls say the campaigns are over long before Election Day.
It’s media-fueled voter suppression.
And it’s legal.
Brian Kirwin is a political consultant based in Virginia Beach, Virginia with over 60 campaign victories.