I know, I know, the polls say otherwise. The polls are crap and have been for years. Just how many times do they have to get the easy things like elections wrong before you stop treating them as gospel?
Yes, elections are easy. If you sample enough and correctly, more often than not you should correctly predict election outcomes. Once in a while it’ll go sideways but 99% of the time the polls will be close enough as to make no difference. But the last few elections haven’t been ‘close enough’ or even correct. That’s because the polls are messed up. Simple really.
Part of the problem is that people are changing their voting behavior – we call that a ‘volatile electorate’. Just means they don’t vote as predictably as before. It’s not a big thing – happens every decade or so – but it can mess up your polling if you aren’t aware of it and don’t compensate. The easy way to compensate is also the expensive one – use bigger samples. A LOT bigger.
In the days of cell phones, that is REALLY expensive. See. most people do what you do – they hang up on unknown numbers. If they do answer they usually tell the nice pollster no, they don’t want to participate, however, they often aren’t that polite about it. People don’t like being hassled and polls feel like hassles.
To compensate, you call more people. Eventually enough answer to do the survey analysis but you’ve also ticked off about 10 to 1 more. Self defeating, really.
Finally, you get a desperate enough client that you send real people to talk to people. The cheapest of these are the exit surveys but predicting an election by exit survey these days is just stupid. The real tallies will be in as fast as you can tally the exit poll. Why bother? Oh, we still do them to find out why people voted as they did but they aren’t used to predict any longer.
The only real solution is to knock on doors. This is a thankless job where you get doors slammed in your face a lot. Needless to say, pollsters have to pay people more to get them to actually finish the job. The days of cheap labor are done. But this yields the most reliable results.
There are other ways but people stopped hanging out in malls so they aren’t as easy. Basically, find a place people hang out and set up a booth. Best if you offer free stuff. Best of all if you pay for filling out the survey. Well, at least you get more people to answer – a lot of their answers will be complete crap as they just want the payoff and don’t care how they answer.
Polling is fun!
It actually is but it comes with some challenges. That’s where the trouble comes in. Most clients don’t want to pay for large sample sizes, let alone door to door surveys. They want cheap and quick.
Enter the micropoll. These are the crappy little things you see EVERYWHERE. They use the lowest rational sample size and weight the HECK out of the results in hope of getting decent results. Guess what does not work if the electorate is even a little volatile? You got it – these things become complete garbage in short order.
Shy of a course in Statistics and one in Survey Research want proof? Go search Pew, Gallup and Zogby for Presidential Approval. If you even find one it’ll have more footnotes than graphs. Pew, Gallup and Zogby are the top polling agencies in the country – their reputations are their bread and butter. They can’t be getting things horribly wrong on a regular basis so they stop surveying on things where they can’t trust the results. They all dropped out of predicting the 2020 Presidential race for exactly that reason – they couldn’t afford the kind of polling that would have yielded reliable results.
Now, ABC, CBS, NBC, Fox, MSNBC and CNN along with a host of others either do or sponsor polls themselves weekly or biweekly. You’ll see Presidential Approval all over their sites. Why? The stupid things can’t be calibrated until the election – and a few weeks before that they all switch to more reliable polling methods and categories.
If you can’t calibrate Presidential Approval, how the heck do you calibrate something like ‘polarization’? You don’t. You just pretend that your stupid micropoll actually measured something. Remember, the sample is tiny – too tiny to analyze as raw data. That’s where weighting comes in – and amazingly, they find what they expected to find!
Sadly, they aren’t really even cheating. They are basing their weights on the last election. With a very stable electorate, this can work – although polarization isn’t measured by an election – but with a volatile electorate it’s pure fantasy.
Of course the results are internally consistent and consistent with other polls weighted the same way – that doesn’t mean much. We expect that to happen with weighted polls regardless of how representative they actually are. The only way to know if they are correct is to calibrate them to much larger polls – polls we no longer do.
Then how do I know polarization is a myth? Good question! Glad you asked!
Because there are LOTS of other indicators besides polls. Polls are like yardsticks – they are for measuring. That makes them tremendous tools especially for campaigns when they are reliable. But if your fuel gauge is broken, can you still tell if you are running out of gas? Sure – you use other indicators and some common sense. If you haven’t filled up in a week, you head straight for the gas station. If the engine starts knocking, you pull over, grab the gas can and start walking.
Fuel gauges are better but not the only game in town.
What other indicators do we have?
Turn out is a HUGE one. People willing to turn up for a political rally are also going to vote for that guy 99% of the time. Ditto political yard signs. But there’s a bigger one.
Donations. People vote the same with their wallets as they do in the voting booth. We aren’t interested in dollar amounts – big donors tend to hedge their bets and they can only vote once anyway – we’re looking at the number of small donors. If they bother to donate, they will very likely check the box on Election Day.
Trump rakes in small donors. Biden can’t get people to show up for rallies. Draw your own conclusions.
But what about polarization, not just elections?
You are getting so good at this! Exactly the right question! Those indicators are more muted on issues than on elections – people seldom stick up yard signs on issues and donations can be harder to sort out why they are exactly donating. They still tell us quite a bit so don’t discard them but we do have other ways.
Political activity. This is stuff like protests, rallies, spontaneous expressions, voter drives and a host of other stuff. Let’s look at protests for the moment.
We saw rioting in 2019 after the George Floyd incident. But where was all this rioting? The bulk was in Democrat strongholds – not exactly the places you expect. A few weren’t, that is true, but most and most of the well reported ones were in California and Portland, Oregon. Why does this matter? Two reasons: it bring the motivations of the organizers into question – they are picking very safe places where they are unlikely to face arrest, not the places where the problem supposedly existed; and there was little to no appreciable spread.
That latter is critical for our purposes. If the US is so incredibly polarized, these riots should have spread far beyond what was being organized but even the organized attempts failed to gain traction outside Democrat strongholds. Okay, there was a really funny attempt at Stone Mountain, Georgia but where are the protests in Birmingham or Dallas? If they occurred at all they were minor and peaceful.
Contrast that to the Civil Rights Movement and the Freedom Riders. Doesn’t really hold a candle to it, does it?
Unless Portland, Oregon was the center of police violence – and the stats don’t begin to support that – the protest weren’t happening where they should have been and they showed no sign of effective spread. After Kenosha, they showed no spread at all.
That’s far more likely to be activists and rabble rousers than a real organic political movement.
Far more organic were the apparently spontaneous outbursts in sports arenas around the country in 2021 – remember ‘Let’s Go, Brandon’? Only that wasn’t what they were chanting originally….
But no counter cheers. Despite a few years of silly athletes kneeling for the American Anthem, there weren’t crowds willing to chant pro-Biden or even pro-Democrat slogans. Democrats do go to football games. right? Okay, what about soccer? No soccer moms thought to cheer their favorite president while their kids lost 20/0?
Taking a much more thorough look than I am describing here, I do see political trends rightward – the political pendulum was about due for a shift – but I don’t see this polarization outside the bases. They’re supposed to be polarized. But the usually complacent middle doesn’t seem to be.
The Left calls for boycotts almost daily and they scare the heck out of retailers but almost never does anything come of it. Bud Light sends a specialty can to the wrong TikTok celeb and all heck breaks loose. Again, points to a rightward shift but not to a polarized electorate. If we were really polarized, both sides should be able to create major, effective boycotts but right now. only one side can.
There are few things that will ignite the middle, much to the chagrin of the rest of us. Ramming politics into their faces is one way; messing with kids at all is the other.
Bud Light went for both with Target right behind. Idiots.
But what about abortion?
What about it? It moves the base more than the middle. The Court’s decision might have provided fodder for the Left in 2020 but where, exactly, were the protests as states began outlawing abortion? The two sides dutifully lined up on the capitol steps and waved posters at each other but in most cases, the whole protest with both sides really did fit on the capitol steps and had room to spare.
You’re right, this is the issue that should have mobilized the Left but it very quickly fizzled. Not a great indicator of polarization. Heck, it’s a better indicator of losing support.
Rallies, when they occur, tend to be larger on one side than the other – that’s an indicator of a political shift, not political polarization.
To be fair, as a shift begins, it can look like polarization in the polling. But if it really is polarizing, you should see strong indications of alliance on both ends – we just don’t see that right now. The only positive indicator I can see for polarization is in the state legislatures but that’s a weak indicator at best. We have so few truly purple states – another indicator that argues against polarization – that we should expect legislatures to represent their state’s majorities and not a shifting between majorities which is what we’d see in a highly polarized environment.
Unless somehow all these highly polarized people manage to never be in the same state?
Despite all those lovely red and blue maps, that’s not how reality works. There are large enough enclaves in a lot of states that we should see some fun fireworks if those populations were really polarized. Instead, they all go to the game and either chant ‘Let’s Go Brandon’ or they hit the concession stand. Usually, they just cheer their team and complain about the weather.
We’re all united on the weather. We all hate it, love it or wish it were different.
That ain’t polarized. That’s just America.
