Breathe.
Breathe.
The Economist election map is red and makes you blue. Polls needs a tobacco warning. “Our 3% margin of error elevates heart rates, leads to anxiety, and has other harmful side effects.” Breathe.
You see this and your pulse quickens. Why? The poll that’s stressing you out isn’t science. It’s abstracted impressionist art. Stats hitting the proverbial fan. Except instead of Jasper Johns’ Map we get an electoral college map. When did a survey of 2800 people stretched across 2800 miles have so much power?
The Economist explains its own poll’s total uselessness like this, “In fact, our model finds that there is a nearly 50/50 chance that one of the two candidates gets at least 306 electoral college votes.”
A poll has power because we don’t know what goes into it. Unlike, say, a Big Mac. Other than E. coli, we don’t fear a Big Mac because we know what goes into one. Two all beef [sic] patties, lettuce, cheese, pickles, onions, and a sesame seed bun. If you sang the song in your head, you know something is missing. Special sauce. Every poll has a secret special sauce. Once you know a poll’s secret, it’s not scary at all.
Pollsters know 56% of voters with party affiliations are registered democrats. And, 44% are republicans. So polls should be 56/44 democrat. Nope. More than half of registered voters – 105 million of them – don’t party with a party. Pollsters need to know what to do with the unparty people. A poll from Gallup says 28% of people feel like democrats and 31% feel republicany. But… that leaves 41% of voters who feel like they’re stuck in neutral.
No matter how many rabbit holes pollsters go down they end up with a large group of people they can’t pigeonhole. That’s where the guessing comes in.
“We expect this many dems, this many repubs, this many indies, this many women, this many men, this many blacks, this many whites, this many Latinos, this many black men, this many suburban white soccer moms, and on, and on, and on it goes.”
Eventually, pollsters decide they need one person in their sample to represent the set of fat, balding men over the age of 50 who self-identify as “mutt.” They have a list of thousands of people who fit that description AND who still have a land line phone AND AND who will still answer their land line. Ok, maybe not thousands. But dozens. Then, the pollsters call those folks until one answers. “Who are you voting for?” Done. Then, they do that for each demographic in their survey. “We need three white soccer moms in the northeast. Calls!”
Concocting a poll is a lot like kitchen condiment alchemy. Everyone knows that Mom’s Thousand Island dressing has this much mayo and that much ketchup. The secret part is guessing the right amount of relish. Pollsters don’t relish this election because a lot of people are sampling a different party. Which makes their guesses worse than useless.
To sort of quote Will McAvoy from The Newsroom,
“If polls are so fucking smart why are they wrong so goddam always?”