Listen now | This week it’s the UK General Election, and lots of other countries either have elections coming soon or have recently voted. Lots of pollsters and political scientists have been attempting to predict the outcomes - but how successful will they be? In this
I was late to this episode. I studied political science I thought it was really good. I would draw a more distinct line between political science and political consultants / pollsters. I'm being pretty nitpicky. You guys did draw some good distinctions at the end. My professors in grad school emphasized that political science was not about prediction, but explanation. Academia does seem different these days, so I don't know if they draw the same distinction today. Political science is a weird amalgam of psychology, sociology, history, and philosophy. Probably the reason I chose to study.
Election polling is usually more journalistic. Most of the big ones are trying to gauge public opinion as accurately as possible, but There are a lot of polls with an agenda. I remember when a congressional candidate in my area was touting a poll that found that she was within the margin of error. It was plus or minus 8 points or something like that. Turns out it was a poll of 300 people and really done just as a tool to convince people that they weren't throwing their money away by donating to her doomed campaign.
Another intersting episode. As to the "accuracy" of the polls in the 2024 UK general election, there is a case to argue that they understimated the size of the Tory vote / overestimated Labor. See this by DecTech https://www.linkedin.com/feed/update/urn:li:activity:7216105448641282052/ (on LinkedIn) for an overview. Another episode - or a short - to follow up on this, perhaps?
You briefly mention that polling fraud would be unlikely... there was a case of polling fraud in 2010, although perhaps not exactly like the other fraud issues you mention. In 2009, the US political website dailykos.com hired polling firm Research 2000 (R2K) to conduct a wide range of weekly polls for them. By June 2010, Nate Silver's site fivethirtyeight.com---incidentally, Nate Silver got his start by posting on dailykos.com---ranked the R2K polls as some of the least reliable, and a trio of DailyKos readers (a political consultant, a very well respected physicist, and a wildlife researcher) became suspicious. They dove into the reported polling numbers and found all sorts of numerical artifacts that strongly suggested that the polling data were made up. Some other statisticians looked at their report, agreed, Dailykos sued, there was legal blustering, and if Wikipedia is to believed, Dailykos eventually won the suit by default.
Here's the post on Dailykos laying most of it out:
Another winner! Have you ever thought about doing an episode on the claim that ‘diversity means better decisions’? I hear this all the time, and it gets repeated in business articles but are there any actual studies?
Loved your episode. I can’t tell you how long I spent with transparent paper (I am pretty old) sliding it over a distribution to help me get it into my head the idea of the difference between my sample and the underlying data.
I also love your poking fun…but I do have to squeak up for Mother Theresa. In an odd turn of events for an atheist I worked in her hospital for the dying destitute in Calcutta for 6 months in the later 80s. It was both a very sad and very inspiring place. We feed, cleaned and cared for people who had no one else. We tended to wounds as best we could and we used our modest pharmacy to treat infections and ease pain (the was always a trained physician who gave orders). Sometimes we found a donor who would pay to have an otherwise healthy person treated at hospital for compound fracture. For sure she had some ideas that I didn’t like but she did provide care for people who no one else would. I was also very impressed by all the young people from the Catholic Worker movement who came to lend a hand and were so much more able to cope with the poverty.
The last thing about her was when I was leaving she had a little talk with me. She thanked me for my help but gently reminded me that there were plenty of people in Canada that needed care and comfort and I didn’t need to travel around the world to help.
I *knew* Stuart taking potshots at Mother Theresa would get us comments. (And my instinct was that when loads of people say someone's a saint, there's probably some reason for it, even if Christopher Hitchens didn't like her.)
I was late to this episode. I studied political science I thought it was really good. I would draw a more distinct line between political science and political consultants / pollsters. I'm being pretty nitpicky. You guys did draw some good distinctions at the end. My professors in grad school emphasized that political science was not about prediction, but explanation. Academia does seem different these days, so I don't know if they draw the same distinction today. Political science is a weird amalgam of psychology, sociology, history, and philosophy. Probably the reason I chose to study.
Election polling is usually more journalistic. Most of the big ones are trying to gauge public opinion as accurately as possible, but There are a lot of polls with an agenda. I remember when a congressional candidate in my area was touting a poll that found that she was within the margin of error. It was plus or minus 8 points or something like that. Turns out it was a poll of 300 people and really done just as a tool to convince people that they weren't throwing their money away by donating to her doomed campaign.
Another intersting episode. As to the "accuracy" of the polls in the 2024 UK general election, there is a case to argue that they understimated the size of the Tory vote / overestimated Labor. See this by DecTech https://www.linkedin.com/feed/update/urn:li:activity:7216105448641282052/ (on LinkedIn) for an overview. Another episode - or a short - to follow up on this, perhaps?
You briefly mention that polling fraud would be unlikely... there was a case of polling fraud in 2010, although perhaps not exactly like the other fraud issues you mention. In 2009, the US political website dailykos.com hired polling firm Research 2000 (R2K) to conduct a wide range of weekly polls for them. By June 2010, Nate Silver's site fivethirtyeight.com---incidentally, Nate Silver got his start by posting on dailykos.com---ranked the R2K polls as some of the least reliable, and a trio of DailyKos readers (a political consultant, a very well respected physicist, and a wildlife researcher) became suspicious. They dove into the reported polling numbers and found all sorts of numerical artifacts that strongly suggested that the polling data were made up. Some other statisticians looked at their report, agreed, Dailykos sued, there was legal blustering, and if Wikipedia is to believed, Dailykos eventually won the suit by default.
Here's the post on Dailykos laying most of it out:
https://www.dailykos.com/stories/2010/6/29/880179/-
Another winner! Have you ever thought about doing an episode on the claim that ‘diversity means better decisions’? I hear this all the time, and it gets repeated in business articles but are there any actual studies?
Is this the fMRI study on political ideology and threat that Stuart said didn't replicate with a larger sample? https://doi.org/10.1016/j.cub.2014.09.050
The 2014 study doesn't look underpowered, and I can't find the failed replication.
Loved your episode. I can’t tell you how long I spent with transparent paper (I am pretty old) sliding it over a distribution to help me get it into my head the idea of the difference between my sample and the underlying data.
I also love your poking fun…but I do have to squeak up for Mother Theresa. In an odd turn of events for an atheist I worked in her hospital for the dying destitute in Calcutta for 6 months in the later 80s. It was both a very sad and very inspiring place. We feed, cleaned and cared for people who had no one else. We tended to wounds as best we could and we used our modest pharmacy to treat infections and ease pain (the was always a trained physician who gave orders). Sometimes we found a donor who would pay to have an otherwise healthy person treated at hospital for compound fracture. For sure she had some ideas that I didn’t like but she did provide care for people who no one else would. I was also very impressed by all the young people from the Catholic Worker movement who came to lend a hand and were so much more able to cope with the poverty.
The last thing about her was when I was leaving she had a little talk with me. She thanked me for my help but gently reminded me that there were plenty of people in Canada that needed care and comfort and I didn’t need to travel around the world to help.
Anyway great show
I *knew* Stuart taking potshots at Mother Theresa would get us comments. (And my instinct was that when loads of people say someone's a saint, there's probably some reason for it, even if Christopher Hitchens didn't like her.)