I think the problem with polls is the reliance on telephones. I'm 70 years old and haven't had a landline in 9 years, and never answered it before that once caller ID came about. I set my smartphone to silence unknown callers. No one wants spam calls, but if I knew it was a political pollster I'd be glad to talk to them. Leaving a voicemail won't work either because I wouldn't believe it. I think there needs to be an update to the present in the way polls are conducted, although I don't know what that is.
Not the way I wanted to start a Sunday morning, but this is probably one of the most important things I'll read all day, or all week.
For Oregonians, this should really turn up the power on our work on the Governor's race, which started (from Cook) as likely D, eroded to lean D this summer, and now is toss up. Adding Dan's caution to the mix, a MAGA Governor is now more likely than not. And Tina Kotek's campaign isn't helping - she would make a fantastic Governor, but her campaign is mediocre at best.
Meanwhile, Betsy Johson (independent) is an anti-climate, pro-gun insanity, pro-crony capitalism, tough-talking paper tiger running an excellent, highly persuasive campaign. She is deftly exploiting MAGA imagery and language, painting Portland as a smoldering hell-scape that is spreading feces and needles throughout the state. I am stunned at how many Dems I've talked to who are persuaded by her ads. All of that opens the door wider than it's ever been for Christine Drazan (Country Club MAGA) to waltz right in.
Democrats and particularly other pro-choice voters have to hear who and what Betsy Johnson is. (If you need a soundbite: Before the pandemic, Clatsop County - Betsy's district - had the highest rate of homelessness in the state - by a mile.) We've got far more work to do than we could have imagined, particularly in our own house.
Polls can be as addictive as Twitter and I'm sure you remember what you said about Twitter several years ago? As best as I can remember, and correct me if I'm wrong, "Someone told me that Twitter was something I would like, and I tried it and before I knew it, I was living homeless outside of a supermarket with all of my clothing and belongings in the shopping cart". I loved that...and I do notice that you have backed off a notch, which is a good thing. Now, about the polls: I tell myself never ever to believe any of them, although, like you, I am addicted to them. The bits about the Latino voters in the Cohn latest, is very interesting...but again, I just write the postcards for the candidates and contribute 5 bucks each to about 20 a month because if I do more, I will be living outside that supermarket and all of MY belongings in the cart.
I never realized how sharp the split was between media and campaign polls. That’s great info. Thanks for that.
Are campaign polls generally more accurate than media polls? To your knowledge did campaign polls have the same problems over the past six years?
I realize it has been over 50 years since I had a statistics course, but my first job was at Fair, Isaac (the FICO inventors). The chief statistician way back then hated the trends of minimal, shrinking interview bases and over-mathematical manipulation to render a flawed sample supposedly random and representative of the electorate. There are probably parameters, biases and choices in all of that which would remove this from the realm of science. Thus error-prone.
So many people charged with fixing this stuff have a vested interest in continuing to publish the results. Are any polling orgs handing these problems off to Academia with grant money?
I confess I never did political polling. Building credit models are incredibly easier. And far more accurate.
Hey Dan, so if polling is simply broken -- if you can't gather good data from the population -- then how do you envision politics working without polling? I'm sure that the "entertainment polling industry" will still exist, because they just can't help themselves. Is it simply a matter of the "back to basics" model where politicians and staff really work to talk to constituents, via focus groups or other methods?
For those of us who are data geeks, the Pew Research Center has a great article on polling methods and why they have switched from phones to online surveys using panels of demographically representative voters. They even provide these individuals with cell phones and data plans if they don't have internet access. https://www.pewresearch.org/our-methods/u-s-surveys/
Yet polls tell us little about what voter turnout will be--maybe they give us a general idea and that's it. Having Trump on the ballot in 2020 turned folks out on both sides--having abortion on the ballot is likely to turn out women, especially college-educated ones. If I were a Republican running in a swing suburban district, this would make my knees shake.
Do I really think that Tim Ryan can beat JD Vance in OH? Only if the Trump vote is low and the suburban and young new voters (especially women) turn out in droves.
I breathed a sign of relief on Thursday when Fetterman told MSNBC that he’s running like he’s 5 points behind even though recent polls had him up by 9.
Everything I’ve seen so far on abortion has it being a priority for Dem voters and makes them more likely to vote. I haven’t seen a poll that reflects that sentiment among Republican Women or Independents. If Rs continue to rally around inflation and scare tactics, and they have more money and are better organized than us, I’m worried we might be screwed.
What do you think about Dem priorities always chasing the issue flavor of the week (for lack of a better word)? There was a Marist poll last week that showed healthcare priorities were dwindling, gun violence prevention was in the single digits and climate didn’t event make the list. Meanwhile, GOP voters are consistent with economy and immigration.
Polling in the modern age of Trumpism is garbage, in part because of Trumpism. Those infected by it see pollsters as the enemy as much as they do the legitimate media. Yet pollsters still aren’t acting like they know this.
The truth that few want to publicly admit is that one must lie to Trumpers convincingly to get them to think you’re in their in group. Only then will they tell you what they really think.
Yes, it’s lying to the mentally ill & those who are intellectually & ethically challenged to act in that manner. But if pollsters want the truth from that group of significantly defective beings, that’s what they need to do.
As for how Democratic & Lefty individuals should treat the polls until the pollsters fix themselves? Use polls as very rough guides, and ALWAYS assume your candidate or issue is at least ten winnable points behind, even if the polls say you’re up by twenty.
Pollsters, like the media, are not our friends. That’s a hard lesson Dems & Leftys still haven’t fully learned.
Seldom mentioned is the effect COVID is going to have on the election.
Not long ago, thousands were dying per day. Today the average is 400 per day. If the CDC is correct, the majority of those dying are elderly and not vaccinated. That demographic largely votes for gop candidates. Pollsters can't poll dead voters.
Our recent elections have come down to a relatively small margin of votes in key states. If that many are dying on a daily basis shouldn't that raise the question that there aren't as many gop voters as in the past due to COVID?
The basic problem with polls is that the problem is asymetric, as pointed out in some of the comments. I would go further- the MAGA universe loathes the mainstream media and are likely to intentionally lie to pollsters, who they assume are an arm of the media, if they talk to them at all, in order to own the libs. So dems should assume they get 4-5% less than the polls indicate....at least in presidential elections
Do you think one of the reasons polls are off (or maybe just flip when actual ballots are needed) is because Dems get complacent and start thinking there are enough votes without theirs? Is it the difference between older (more likely Republican) voters and young Democrats (who get busy, forgetful or haven’t been told enough times that every vote counts)? And that independents are even less motivated to vote if they think it’s decided?
Great perspective, thanks Dan. We have to assume all these races will be closer than we'd like. The work stays the same.
I think the problem with polls is the reliance on telephones. I'm 70 years old and haven't had a landline in 9 years, and never answered it before that once caller ID came about. I set my smartphone to silence unknown callers. No one wants spam calls, but if I knew it was a political pollster I'd be glad to talk to them. Leaving a voicemail won't work either because I wouldn't believe it. I think there needs to be an update to the present in the way polls are conducted, although I don't know what that is.
Not the way I wanted to start a Sunday morning, but this is probably one of the most important things I'll read all day, or all week.
For Oregonians, this should really turn up the power on our work on the Governor's race, which started (from Cook) as likely D, eroded to lean D this summer, and now is toss up. Adding Dan's caution to the mix, a MAGA Governor is now more likely than not. And Tina Kotek's campaign isn't helping - she would make a fantastic Governor, but her campaign is mediocre at best.
Meanwhile, Betsy Johson (independent) is an anti-climate, pro-gun insanity, pro-crony capitalism, tough-talking paper tiger running an excellent, highly persuasive campaign. She is deftly exploiting MAGA imagery and language, painting Portland as a smoldering hell-scape that is spreading feces and needles throughout the state. I am stunned at how many Dems I've talked to who are persuaded by her ads. All of that opens the door wider than it's ever been for Christine Drazan (Country Club MAGA) to waltz right in.
Democrats and particularly other pro-choice voters have to hear who and what Betsy Johnson is. (If you need a soundbite: Before the pandemic, Clatsop County - Betsy's district - had the highest rate of homelessness in the state - by a mile.) We've got far more work to do than we could have imagined, particularly in our own house.
Polls can be as addictive as Twitter and I'm sure you remember what you said about Twitter several years ago? As best as I can remember, and correct me if I'm wrong, "Someone told me that Twitter was something I would like, and I tried it and before I knew it, I was living homeless outside of a supermarket with all of my clothing and belongings in the shopping cart". I loved that...and I do notice that you have backed off a notch, which is a good thing. Now, about the polls: I tell myself never ever to believe any of them, although, like you, I am addicted to them. The bits about the Latino voters in the Cohn latest, is very interesting...but again, I just write the postcards for the candidates and contribute 5 bucks each to about 20 a month because if I do more, I will be living outside that supermarket and all of MY belongings in the cart.
Thanks, Dan. I was text banking in AZ yesterday and told a voter just that...we can't trust the polls we need to get out and vote!!
I never realized how sharp the split was between media and campaign polls. That’s great info. Thanks for that.
Are campaign polls generally more accurate than media polls? To your knowledge did campaign polls have the same problems over the past six years?
I realize it has been over 50 years since I had a statistics course, but my first job was at Fair, Isaac (the FICO inventors). The chief statistician way back then hated the trends of minimal, shrinking interview bases and over-mathematical manipulation to render a flawed sample supposedly random and representative of the electorate. There are probably parameters, biases and choices in all of that which would remove this from the realm of science. Thus error-prone.
So many people charged with fixing this stuff have a vested interest in continuing to publish the results. Are any polling orgs handing these problems off to Academia with grant money?
I confess I never did political polling. Building credit models are incredibly easier. And far more accurate.
Hey Dan, so if polling is simply broken -- if you can't gather good data from the population -- then how do you envision politics working without polling? I'm sure that the "entertainment polling industry" will still exist, because they just can't help themselves. Is it simply a matter of the "back to basics" model where politicians and staff really work to talk to constituents, via focus groups or other methods?
For those of us who are data geeks, the Pew Research Center has a great article on polling methods and why they have switched from phones to online surveys using panels of demographically representative voters. They even provide these individuals with cell phones and data plans if they don't have internet access. https://www.pewresearch.org/our-methods/u-s-surveys/
Yet polls tell us little about what voter turnout will be--maybe they give us a general idea and that's it. Having Trump on the ballot in 2020 turned folks out on both sides--having abortion on the ballot is likely to turn out women, especially college-educated ones. If I were a Republican running in a swing suburban district, this would make my knees shake.
Do I really think that Tim Ryan can beat JD Vance in OH? Only if the Trump vote is low and the suburban and young new voters (especially women) turn out in droves.
I breathed a sign of relief on Thursday when Fetterman told MSNBC that he’s running like he’s 5 points behind even though recent polls had him up by 9.
Everything I’ve seen so far on abortion has it being a priority for Dem voters and makes them more likely to vote. I haven’t seen a poll that reflects that sentiment among Republican Women or Independents. If Rs continue to rally around inflation and scare tactics, and they have more money and are better organized than us, I’m worried we might be screwed.
What do you think about Dem priorities always chasing the issue flavor of the week (for lack of a better word)? There was a Marist poll last week that showed healthcare priorities were dwindling, gun violence prevention was in the single digits and climate didn’t event make the list. Meanwhile, GOP voters are consistent with economy and immigration.
Polling in the modern age of Trumpism is garbage, in part because of Trumpism. Those infected by it see pollsters as the enemy as much as they do the legitimate media. Yet pollsters still aren’t acting like they know this.
The truth that few want to publicly admit is that one must lie to Trumpers convincingly to get them to think you’re in their in group. Only then will they tell you what they really think.
Yes, it’s lying to the mentally ill & those who are intellectually & ethically challenged to act in that manner. But if pollsters want the truth from that group of significantly defective beings, that’s what they need to do.
As for how Democratic & Lefty individuals should treat the polls until the pollsters fix themselves? Use polls as very rough guides, and ALWAYS assume your candidate or issue is at least ten winnable points behind, even if the polls say you’re up by twenty.
Pollsters, like the media, are not our friends. That’s a hard lesson Dems & Leftys still haven’t fully learned.
Seldom mentioned is the effect COVID is going to have on the election.
Not long ago, thousands were dying per day. Today the average is 400 per day. If the CDC is correct, the majority of those dying are elderly and not vaccinated. That demographic largely votes for gop candidates. Pollsters can't poll dead voters.
Our recent elections have come down to a relatively small margin of votes in key states. If that many are dying on a daily basis shouldn't that raise the question that there aren't as many gop voters as in the past due to COVID?
The basic problem with polls is that the problem is asymetric, as pointed out in some of the comments. I would go further- the MAGA universe loathes the mainstream media and are likely to intentionally lie to pollsters, who they assume are an arm of the media, if they talk to them at all, in order to own the libs. So dems should assume they get 4-5% less than the polls indicate....at least in presidential elections
Do you think one of the reasons polls are off (or maybe just flip when actual ballots are needed) is because Dems get complacent and start thinking there are enough votes without theirs? Is it the difference between older (more likely Republican) voters and young Democrats (who get busy, forgetful or haven’t been told enough times that every vote counts)? And that independents are even less motivated to vote if they think it’s decided?
Thank you. Useful. Turn out. That is where it is all at. Turn out. Really--thank you.