The debate launched a thousand polls
Ticking 'yes' for surveys. Asimov says 'no'. As does a 1960s gem on voter data
Welcome to This Week, Those Books, your rundown on books new and old that resonate with the week’s big news story.
🎧 Would you rather listen? The podcast drops at the w/e.
We’re back from our break with a subtly refreshed outlook and format, which is, hopefully, more pithy. To welcome back our entire community of more than 10,000 subscribers in 114 countries, this post is free to read.
– Rashmee
The Big Story:
The one clear result from that zinger of a Kamala Harris-Donald Trump TV debate on September 10: It set off yet another round of opinion polls. These will drive the conversation for the 56 days until America’s November 5 election.
But should political polling, a two-centuries-old American invention, really count for so much of the chatter about the US presidential election?
After all, as G Elliott Morris,1 editorial director of ABC News Data Analytics, has written, “polls suffer from the problem that most people expect too much of them”. Most people believe “that surveys are more like an assured statistical calculation than a rough estimate, such as a weather forecast”.
In fact, the only thing polls have really been able to say about the 2024 US presidential election thus far: It’s very close.2
This Week, Those Books:
A brilliant survey of the science and art of polling.
A short story that shrinks the sample size to one.
BONUS: The dangers of using bits and bytes to create political models.
And if you want another read on politics, this specially unlocked post from the archives will be freely available for one week:
The Backstory:
The opinion poll is defined as the systematic posing of a series of questions to a group of people with the answers adjusted to make them representative of a wider group.
The word ‘pollster’ was first used by TIME magazine in 1939.
Straw polls – simple tallies of support for political candidates – were the precursor to modern political opinion polls, and conducted from the 1820s at events such as Fourth of July celebrations.
The Columbus Dispatch newspaper started its own polling in the late 19th century.
In the early 20th century, the Literary Digest, a US magazine, started national surveys and correctly predicted five presidential elections (1916, 1920, 1924, 1928 and 1932).
In 1936, George Gallup predicted Franklin Roosevelt’s landslide win while the Literary Digest said the opposite.
Gallup changed polling by using a new method based on sampling. Polls were now considered high-quality not by the number of respondents but by how representative they were of the population.
From the US, polling first spread to the UK. Then, Denmark and Sweden (1939); Canada and Australia (1941); France and Germany post-WWII; Ireland (1961), Spain (1966) and further afield.
A 2012 study said that nearly 40 countries ban the publication of pre-election polls in one form or another.3
This Week’s Books:
Polling UnPacked: The History, Uses and Abuses of Political Opinion Polls
By: Mark Pack
Publisher: Reaktion Books
Year: 2022
British politician Mark Pack4 corrects the misreading by my tribe – journalists – of the polling industry. He argues that there is no crisis in polling. In fact, pollsters have a better track record of forecasting elections than the media understands or admits. Research on tens of thousands of polls for a half-century from the 1940s found that the “average polling error” ( the gap between the predicted vote share and the result) was just 2.1%. Still polls get a bad press because the media doesn’t understand they aren’t meant to call winners and losers but to calculate a range of possibilities. That’s what’s called the ‘margin of error’.
Pack explains the science of polling using small samples with a brilliant metaphor: “A blood test takes a tiny portion of your blood, yet those few drops are sufficient” to draw conclusions about one’s health. And polls are better than alternative measures of public opinion, such as crowd size at rallies, social media ‘likes’ and yard signs on the lawns of suburban America. At the outset, he quotes Sherlock Holmes: You can never foretell what any one man will do, but you can say with precision what an average number will be up to.
But he also admits some high-profile polling failures, not least the 2015 UK election, when the likelihood of a Labour Party victory and Conservative defeat was overestimated. Polls can fall short, Pack says, if biased samples are used and late swings in momentum are left unaccounted.
Choice quotes:
“…controversies over how much attention politicians should pay to pollsters. There is a tension between the roles of politicians as leaders, telling the public what they think is right, and as listeners, letting the public determine what happens”.
“…during the nineteenth century, the idea of sampling began to take hold. That is, rather than having to count everything, counting a representative sample would still enable you to reach the truth – more quickly, cheaply and easily. A pioneer was Statistics Norway, the country’s first official statistics organization, created as an independent body in 1876”.
Franchise
By: Isaac Asimov
First published by magazine If: Worlds of Science Fiction
Year: 1955
This short story, republished in the 1957 collection Earth is Room Enough, imagines a future in which computer modelling has become so refined that American elections involve only one person voting! In 2008, the “Voter of the Year” is store clerk Norman Muller of Bloomington, Indiana.
Choice quotes:
"Listen, I was around when they set up Multivac. It would end partisan politics, they said. No more voters' money wasted on campaigns. No more grinning nobodies high-pressured and advertising-campaigned into Congress or the White House. So what happens. More campaigning than ever, only now they do it blind…I say, wipe out all that nonsense. Back to the good old-"
– Matthew Hortenweiler, Norman’s father-in-law to the family before the 2008 voter is announced
“Any American can be brought to Multivac to have the bent of his mind surveyed. From that the bent of all other minds in the country can be estimated. Some Americans are better for the purpose than others at some given time, depending upon the happenings of that year. Multivac picked you as most representative this year. Not the smartest, or the strongest, or the luckiest, but just the most representative”.
– Secret Service agent Phil Handley to Norman
“Multivac already has most of the information it needs to decide all the elections, national, state and local. It needs only to check certain imponderable attitudes of mind and it will use you for that. We can't predict what questions it will ask, but they may not make much sense to you, or even to us. It may ask you how you feel about garbage disposal in your town; whether you favor central incinerators…"
– John Paulson, ‘Senior Computer’ to Norman before the 2008 voter ‘speaks’ to the machine
Bonus Pick: The 480
By: Eugene Burdick
Publisher: McGraw Hill
Year: 1964
A bestselling novel by the glamorous political scientist and TV pundit of the JFK era, it criticises the use of polling and computers to the point of manipulating voters. The title refers to the 480 categories (such as rural, eastern, white collar) into which the American electorate was divided by the 1960 Kennedy presidential campaign on the advice of Simulmatics Corporation’s5 computers and data scientists.
I hope you find This Week, Those Books useful, thoughtful, and…a conversation starter. It’s a small operation here at TWTB, and support from readers like you helps keep this news literacy project going.
Email thisweekthosebooks@substack.com to say ‘hi’ or if you want a pdf of the Asimov story.
Connect with me on LinkedIn | Twitter | Bluesky | Facebook | Threads |YouTube
Quoted from G Elliott Morris’s 2022 book Strength in Numbers: How Polls Work and Why We Need Them
CNN data analyst Harry Enten says: “This is the closest presidential campaign 60+ years”. In the 15 presidential elections since 1964, a candidate has led by more than five points in the national polling average for at least three weeks but in 2024, that hasn't been the case for a single day
https://www.bbc.co.uk/news/uk-politics-35350419
Mark Pack is president of the Liberal Democrats. https://www.markpack.org.uk/
Harvard historian Jill Lepore’s 2020 book If Then: How the Simulmatics Corporation Invented the Future suggests the firm mined data, targeted voters, manipulated consumers and destabilised politics much before Facebook, Google, Amazon and Cambridge Analytica
The main thing polls do Today is to make most every race a neck-and-neck nail-biter. For any race where one candidate is a runaway leader, they cause weaklings to drop out quickly and those that remain to do _something_ to "up their game" to be more competitive. This will accomplish two things: The public and competitors will clamor for more polls (like taking a patient's blood pressure every half-hour, just to see if anything has changed). And the pollsters make A LOT more money. Meaning that the pollsters are motivated to interpret results to keep them within the margin of error.