Last week I had the joy of speaking to Ben Walker, founder of Britain Elects, Labour councillor, and overall the person who knows by far the most about the intricacies and eccentricities of voting patterns in the UK. We talked about how he built his election forecasting model in 2024, which came closer to predicting the general election outcome than any of the professional pollsters, and about what polls of voting intentions do and don’t tell us.
You can listen to the full episode on Spotify or Apple Podcasts. Don’t forget to subscribe to get new episodes as soon as they’re released.
You can follow Britain Elects on X: https://x.com/BritainElects
and on Bluesky: https://bsky.app/profile/britainelects.com
TRANSCRIPT:
GEORGINA:
My guest this week is Ben Walker, senior data journalist at The New Statesman and co-founder of Britain Elects, a website that aggregates polls of people’s voting intentions and which dabbles very successfully in election forecasting. At the 2024 general election, Ben’s model was the one that came closest to predicting the end result, an incredible achievement considering that it proved more accurate than all the professional polling companies.
I’m sure that he’s already been likened to a British version of Nate Silver. I can see him cringing, but if not, I’ll be the first to make that comparison. Ben’s also a reporter and commentator on data and politics and has made many appearances on the radio, on BBC News and on podcasts. In 2023, he crossed over from the sidelines into politics itself, being elected as a borough councillor in Cheshire, representing Chester City Centre. Ben, it’s great to have you on the show.
BEN:
Thank you for that introduction.
GEORGINA:
Well, I’m glad that you appreciated the Nate Silver comparison. We maybe will get onto that later on. Briefly, to start off with, could you tell me about what you do? What does being a data journalist with a side gig in election forecasting involve?
BEN:
Not much sleep, not much sleep. And I don’t perhaps know your audience who’s going to be listening to this, who’s going be watching this. But I would say this to you, if anyone aspires to do a job in journalism, when you sign the contract as I did in 2020, I read you’ll be working from 9am till 6pm. And then I realised your job doesn’t really have to be done within then and your job often won’t be done within then. It’ll be done all the time.
You’ll always be logged on. You’ll always be aware of what’s happening throughout the world and you’ll be covering it. Embargos get lifted at midnight. Polls come out at 10. It’s very fluid. don’t, if I may sort of, if I might be selling it poorly here, you don’t often switch off, which might be my own experience, but I wouldn’t give it up. I wouldn’t give it up because to tell you the truth, the reason, and I might be answering further questions here, the reason I’m into this, the reason I really like doing this is because I think the fundamental wrong with so many of us in the West is that we don’t understand our neighbors. We don’t understand why our neighbors think the way they think. Too many of us make assumptions. And I think through polling, through an understanding of data, you can better understand your neighbors. You can better understand your country. I’ll give you one example. It was June the 24th, 2016, the morning of the sun was rising up. And I was just checking my Facebook feed as the country awoke to the fact that Britain had voted leave.
And all my Facebook friends were realizing who on earth voted Leave? I had no idea anyone would ever vote Leave and they were in Yorkshire. They were in Leave voting Yorkshire. It was the message to me really, that so many of us who may be so logged on, maybe even into politics, don’t have a clue what this country thinks, which is, I suppose, why I do this.
to get on to lots of the things that you talked about there. But before we get into that, I want to backtrack a little bit and ask you about how you got into this in the first place. Britain Elect was a project that you started in 2013 with another colleague, Lily Jane Summers, who sadly is no longer with us. Can you tell me about that Britain Elect project, how that started and why and how it’s evolved since then?
Yeah, so I set it up when I was 17. I was a high school student and I, tell you the truth, was a bit of a recluse as a kid. And the friends I made were the ones online. And one of them was a girl called Lily and she was a member of the Labour Party and shock horror, she was also trans. And I, having no clue about that kind of world, we just got to know each other and we just talked very honestly and it just…
bloom from there. She was running a little blog where she was just writing about politics. This was in the era in which everyone was blogging. And I sort of got interested in this, started trying to round my hand at writing something. And it was at that stage when you’re just talking to anybody of any party. Tory members, Tory councillors, some of whom I can still call friends even though I am a Labour man, Liberals, kippers, members from UKIP back then. And you just realise everyone, it’s kind of like a sports commentary.
everyone’s interest is in their own party and how well it does and it made me realise every Thursday night some of these activists, hardcore activists, are tuning in to Twitter or wherever to find out how their party is doing, not in the opinion polls but in by-elections. So for the uninitiated, every Thursday almost without fail even up to the final day, even up to Christmas Eve sometimes, a random corner of the country is going to the polls to elect a councillor.
So right, almost every Thursday we’ve… Almost every You know, all the time. The councillor died, resigned, got kicked out for non-attendance, got arrested for doing something indecent. It always happens. There are about 15,000 councillors in the country, I think. And I noticed that and I thought, why doesn’t someone put that in one place? Get the results together, put it in one place and just measure how well the parties are doing. Who’s up, who’s down? It’s the horse racing.
the raw local horse racing, it’s real votes and it’s political as well. People like seeing it and we cover it to this day. But that’s how it started. Built up a following for quite a bit. And then by about 2015, I was working in London, initially just to design graphics for campaigns. And then I worked, know, analyzing poll data as a consultant and it went from there. But then I realized I don’t have any qualifications to my name.
So when I was working in London in 15, 16, I realized I should really get a degree. And then I did. So I paused, went to Chester, made my life there and Britain Elects just carried on growing. And by 2019, was still doing the very same thing, occasionally writing, occasionally commissioning polling as well. I’ve got to that extent, I started to raise some money. The New Statesman noticed the spread of Britain Elects, the value in its following and said, why don’t you do something for us for the 2019 general election?
And then after that they said, well, you know what, you bring a following, you’re pretty good, why don’t you stay on as a data journalist? And I did and I have for five years now.
GEORGINA:
So it started really with the interest in local election figures really. But there may be some people listening who, well, I mean, all of us really, I think, are pretty familiar with the idea of polling, election polling. You’re asking people about their voting intentions, how they’d vote if there was a general election. But I’m not sure that all of us completely understand the ins and outs of how that actually works. Where do these polls come from?
What exactly are these numbers that we’re looking at? Could you just run me through how polling works in a fairly basic way?
BEN:
Yeah, absolutely. To do an opinion poll, you have to ask a representative sample of the public. Ask anyone a few questions and you can wait it. So here’s an example. Let’s say if for instance, I went on the high street of my nearest town or city and I was able somehow to get an answer for something from a thousand people. Importantly, I’d also need to log who they are, age, gender, background, where they live, income, whatever. Because if you want an opinion poll to mean something,
A random poll of the high street of a city tells you nothing. This is what absolutely irritates me about, you know, news studios. They’re obsessed with Vox Pops, which is just going on the high street and asking, what do you think of the budget, Mrs. Muggins? That’s not representative. So an opinion poll is getting a thousand people, but making sure it’s representative. Try and aim for it to be 50 % female. If it’s 60 % female, you just weight it down a little bit. But if it’s a 90 % female, I’ll consider that poll.
very, very worthless because you’d be weighting up that 10 % of men to 50 and that would be absolutely ridiculous. That’s a non-starter. That’s how you make a poll representative, but you’ve got to make sure the sample is neutral. There’s a lot of polling companies out there who have over the years built up a pretty big audience. So you’ve probably all, many of your listeners may have heard of YouGov and when you go on there, you’re often asked, sign up to our membership, answer questions every week and you might earn a few quid.
after 800,000 questions. Okay? It’s a bit like that. This is how YouGov gets rapid polling, because they’ve got hundreds of thousands, if not millions of people willing to answer their survey questions. The problem with polling is that who are the people filling in the polls? They are people wanting to give views. So you risk undersampling people who have socially unsavory views. There’s a risk there. So pollsters, how do they respond? They either go to extra lengths to seek them out.
They make their market research to reach out to them much more appealing to them, a bit more anti-establishment, a bit more, you know, put two fingers up to the politicians, tell us what you really think. Adverts can be a bit like that. There is a risk always with a poll being unrepresentative and there were some misses in America. The reason Trump was under-sampled, the reason Trump over-performed the polls in 2016 was that in so many areas, our assumption
of who the people would vote was based on their enthusiasm to turn up to the pollster, to the pollster survey. And the problem is the people who, you you had a lot of Latinos who just wouldn’t answer the pollster, who wouldn’t answer the polls, call to arms, but were going to vote, but they were embarrassed to say so. We’re having a lot of that in more anti-establishment, in a more anti-establishment period in time, which I would say is now, you’re having a lot of people who are not exactly proud of admitting to their opinions but nevertheless have them and nevertheless will turn out. Polster’s job is to wait against that and then it becomes science and assumptions.
When did election polling become a thing? And I suppose I’m asking in the UK and also more broadly, is this something that’s fairly recent? I think we can all kind of assume it’s not, but is it maybe a bit more widely used and a bit more, say, widely reported on nowadays than it used to be?
Well, yeah. So when did polling start? It’s been going for that. It was certainly around in the whole of the 20th century. Polling, I think, really took off 1920s and 30s. Behind me, I’ve got two big books called the Gallup Books of British Polling, and it goes back to pre-war Britain. And it’s fascinating. But again, assumptions about who is going to turn out. Thank goodness for them, almost everyone turned out in elections like those. Now we have differential turnout and it becomes a little more complex. And by the way, even though despite that, Indian polling back then when more people turned out was less accurate than it is now.
GEORGINA:
So you’re quite embedded in the politics world. You’re obviously now a councillor in Cheshire and you’ve been involved in Labour Party work, or at least on the periphery of it for a while. Can you give me any indication of how polling data is actually used by politicians and by the people who advise them?
BEN:
Politicians use it, campaigns use it all the time to squeeze. We are a first past the post democracy. And yet, despite that, more often than not, people tend to rally to one or two parties. Look at the opinion polls now. Even though the polls are fractured, even though reforms on 29, Labour’s on 20, Tories are on 20, maybe 21 now, who knows? In seat by seat, you’ll find those shares not replicated, you’ll find the squeeze. You’ll find the Labour share still high, you’ll find the reform share still high.
and the other parties squeezed to high heaven. Politicians will use what polling data to have to say it so close. Every campaign uses it. In the general election, the new Labour landslide of 97, you had around about a few dozen extra Lib Dem MPs because of tactical voting. We’re going to see more and more of that on a scale we don’t think expect anymore because as a country, we are not loyal to our parties as we used to. We are now a nation of swingers. We are politically swingy. We are less loyal, less willing to stick to one side despite all the disappointment. We’re willing to swing, we’re willing to switch and safe labour seats, councillors in safe labour seats and indeed councillors in safe Tory seats. I don’t think they’ve woken up to that yet.
GEORGINA:
Do you get the impression that they’re looking at polling data a lot of the time anyway, even say now, when a general election is pretty far off probably?
BEN:
Yeah, I am the master of our nation’s own demise. The success of Briton Alex has seen more and more people look at the polls when they absolutely shouldn’t be, think. Thanks to me, sorry. Well, not thanks to me. Well, to an extent actually, Briton Alex has triggered other accounts to set up and succeed in their own right. it is a very prevalent.
GEORGINA:
I mean, people are absolutely obsessed with polls in a way. The rise of reform over the last year or so has obviously been a real news story. But there have been periods in that time where these snapshots that have been taken have been treated as front page news. Britain Electors is a poll aggregator. Well, I think you said you commissioned some polls of your own, but for the most part, you’re aggregating. Can you just explain what exactly that means and how it works.
BEN:
So initially at the start, when I set it up with Lily, it was just putting it in one place. That’s what aggregation is at the end of the day. And putting it in one place is fine and dandy if the data is good. But people pick and choose. You have one poll with reform ahead by 12 points, massive reach. That’s the narrative definer. The senior journalists at the big broadcasting houses, they’re always quoting that one and they’re ignoring another equally valuable poll, which puts the reform leader only two or three points.
You know, the most popular poll wins there, which an aggregation service is meant to deny because some pollsters have better reach than others. That, you know, we’re meant to stop that. That was at the start. Now what I do is I build my own models to tell you what the polls overall are truly saying. I’ll give you one example. We have one pollster prolifically putting Labour behind the Greens right now. And every other poll puts Labour ahead of the Greens by between two points to 8 points to 11 points, okay? I think by and large you can say therefore, based on the total number of polls, that Labour are ahead of the Greens. Not by much, but they are ahead. That is what I do. I get the data together and I churn it through, I weight it, judge it sometimes by historical accuracy, even though a pollster sometimes changes its methodology without telling you. You make assumptions as the statistician in the room about which has more value.
But to just go back on what you said earlier, now is not, I think, the time to be doing this. I’ve been teasing out the Britain Predicts model at this time for certain by-elections. Does okay. Does okay. Got the reform-sharing Caerphilly bang on, but the level of tactical voting for Plaid at the expense of Labour was not expected. We are in more uncertain territory, and the reason we’re in more uncertain territory is this. After the general election, there was a big drop in enthusiasm for Labour.
And what you saw was voters who normally vote. So if you had a median Britain, semi-detached, mortgage paying, middle-aged, married, a lot of the enthusiasm from those types of people dropped like a stone. And so you had a field, think of the voters like a pitch. You had a much sparser pitch taken up by the loudest voices. And these were reformed voters and these were to an extent smaller party voters. What you’re seeing now since then, is more people are coming back onto the pitch to say they’re certain to vote, but they are not regular voters. We’re seeing unreliable areas, apathetic areas apparently, turn out now, which means polling assumptions go back to the drawing board, I think.
GEORGINA:
I did want to go back to your Britain Predicts model, which hopefully you’re very proud of and happy to talk about, because like I said, this was the most successful model at predicting the 2024 general election outcome. You projected every seat for the Greens and almost every seat for reform. The BBC had reform on course to win another seat, which actually it didn’t in the end, and your model had predicted that. So you proved the BBC and other respected pollsters wrong. How did you do that? What did you get right that others got wrong?
BEN:
It’s strength. It’s where your party is strong. It made assumptions about party strength based on the local elections. It’s my view. Everyone likes to say in council elections people vote differently. It’s about pot holes. It’s about bins. True, true, true. Yes. But it’s not a significant deviation. It’s not a massive deviation. And local strength, ward level strength is a very important indicator from my work at least.
into gauging whether a party will win this or that or won’t. And that helps. That obsession, I suppose, with the horse race nature of council by-elections helped my model become better than most. I’ve been spending the past 10, 15 years building up an archive of telling you where in Cheltenham the Lib Dems are strongest. I’m giving a hypothetical example here. If any of you know of electoral calculus, Martin Baxter, he built a model long ago called Strong Transition Model.
It’s an assumption basically that if you’re a party and if you have a hundred of your voters in the room, 20 of them might be swing, 50 of them might be swing and by swing I mean prone to moving elsewhere and the other 50 or the 80 will stay with you till death do us part. Those assumptions, I copied that but I based that ratio off local election data of the past 20, 30 years which showed you if know Lib Dems Historic here but
many of them can really turn away as quick as they came for another party and if they do turn away quick that ups the value of their weakness so you have a seat where the Lib Dems their strong voters aren’t plentiful their weak voters are plentiful meaning in an election in which they fall by a little bit nationally they fall by a big amount in that seat that’s just an example
GEORGINA:
So a lot of people are not clued in or tapped into what’s happening in politics. That suggests there must be a lot of people who are in the category of don’t know when it comes to voting intentions. I just wondered whether you could explain what pollsters do with people who are in the don’t know camp.
BEN:
Again, it’s all about assumptions. It’s the same thing I do. I’m making assumptions here in my model. So I take the opinion polls. That poll defines 20 % of Britons are still undecided. You can just exclude them and recalculate 100 % for the, you know, those who’ve declared certainty. That’s fine. That is the sort of, we don’t know, we’re not going to try it. We’ll assume they’ll split fairly. It’s not going to be like that.
Well, can be and it can’t be, you just don’t know and this is up to you, the pollster, you, the statistician, you, the political scientist to start to make assumptions.
GEORGINA:
I wanted to look at a couple of recent polls. So I just looked at the couple of ones that you’ve posted on Britain Elects feed on X this week. And there’s a couple of different polls you’ve got here, four days apart from one another. So this is asking people if there was a general election tomorrow, who would you vote for, something along those lines. Reforms ahead in both of them, but in one of them they’re on 32%, the other they’re on 24%.
And you’ve also got some differences in terms of the support for Labour, Conservative and the Greens mainly. What do you make of these kind of findings? Do you read much into either of them? And I guess have you got any advice for how to interpret results that on the face of it appear quite different?
BEN:
Yeah, I’ll take you back in history for a wee while. Oh god, this is so geeky. On Sunday, I was watching the election results program of the 1964 general election covered by the BBC. And what I noticed was they covered the polling prior to the election and there were two pollsters and it was quite a tight election for those who don’t remember in which Labour just emerged ahead and the Tories weren’t far behind. And you had one pollster which consistently put Labour ahead.
And then by election day, narrowed. And then you had one pollster which consistently put the Tories ahead, but by election day, narrowed to the Labour lead. So there was convergence. There was a convergence that no methodology should have anticipated unless the methodology was changed, unless the polls were, you know, fluffed up, I suppose. Variety is the most important thing you can have in political opinion polling, because it means the pollster, the statistician,
is not looking at their opponents, looking at their competitors and putting a finger on the weighing scales and changing it according. We can’t have herding. Herding is the worst possible thing for political data. So having here a reform leader four or a reform leader 14, the Labour share on 15 or the Labour share on 19, the green share on 14 or 17, variety, absolutely brilliant, absolutely important. Where is it coming from? Why is it happening? It’s a good question and there’s a few assumptions. I said it earlier.
Who are the people answering opinion polls? It’s people who are determined to tell you what they think. Where are we getting our opinion polls from? Historically, was always a face-to-face interview, there was always a fella knocking on your door, hello, do mind if I come in, sit down, ask you a few questions? That’s old school. We don’t, I don’t know any pollster that still does it anymore, which is very unfortunate actually, because I think it was a decent way of measuring. We also used to do it by phone conversation. Some pollsters, although very few do it, but the cheapest and easiest way,
is an online sample. Send a survey, survey monkey, you know, and then wait accordingly, spam a load of emails and see what you get. So you’re basing it off often samples of people who are very online. Who were the people online answering surveys? They’re not representative. It’s your job to make it more representative. These are, I regret to say, because I know we are looking for clarity here, these are unknowns. Two years ago, or was it three years ago now, reform were polling 8 to 10 % in the opinion polls and the heady days of Labour on 40%. And they were never, never following it up in a decent performance in by-elections or council elections. So reform appalling 8%, 10%, which was pretty, you UKIP got that and did well enough, but they were never matching the UKIP performance about long ago. They weren’t doing particularly well. So my assumption, my working assumption from them is you’re having some pollsters with very online respondents. And then it got to the stage where we had the general election and most parties performed in line with the polls with the exception of Labour and the Tories and ended up with a form starting to match with what it got. And again, to sort of fall back, we don’t really know what proportion of our respondents are incredibly online. All we know is this, the pollster is constantly changing its methodology. A pollster is constantly trying to readjust.
GEORGINA:
And just briefly, you mentioned herding. Could you explain what that is?
BEN:
Yeah. So just go back to 66, you have a pollster historically pro-Labour and a pollster historically pro-Tory and up until the week before poll, that one’s three points Labour lead and that one’s three points Tory lead. And then for some reason they converge. What are they doing there? They are looking over their shoulder at what’s he doing? What’s he doing? Why is he doing that? Maybe we should come closer to him. And the other guy’s looking over their shoulder and going, basically they don’t want to be the odd one out.
damage their reputation, so they come together. We saw that to a degree in 2015. So many pollsters couldn’t believe their eyes when their initial samples were showing a better showing for the conservatives and they thought, you know, we’re the odd ones out here. It was the same in 2017. A lot of pollsters weren’t catching up with the Labour surge under Jeremy Corbyn, except Survation. Survation was the one that went out on a limb and said, you know what, stuff it, we’ll just publish it. And they did and they got the closest photo share. They didn’t hurt.
GEORGINA:
But most of them do this kind of like a bit of hedging at the last minute because they don’t want to be the one that gets it completely wrong.
BEN:
It’s happening in America as well. It’s really problematic. have to tell you this, in Britain, I would say this because I’m British, but I think we have one of the best polling industries going compared to the rest of the world. In America, they will publish anything and they sometimes will publish stuff that a campaign has commissioned them to do and they will ask questions prior to the voting intention question that they don’t publish. In Britain, and the one thing I do is Britain, Britain Elects.
I only publish polls from people who are signed up to something called the British Polling Council. Now there’s not much transparency there, but it means British Polling Council firms have to publish their data tables and the full questions. Okay, that’s great. That allows transparency. That means I’m not going to give any value to a pollster that is clearly giving very silly poll questions prior to the headline voting intention survey. In America, there isn’t as much regulation there. 538 was often trying to clean through, but they couldn’t clean through all. So they ended up saying this is a Republican poll, this is a Democrat poll, and they still included it.
GEORGINA:
Yeah, I’ve seen some analysis by our friend Nate Silver at FiveThirtyEight or wherever he is now showing that basically a lot of the polls that were coming out before the most recent US presidential election were so close. It must have been because they were hurting. It would have been such a coincidence otherwise, you know, one in several million or what have you, that they would have results that were that similar to one another, which kind of undermines the result.
BEN:
No one wants to be the odd one out. I was sort of guilty of this. was desperate to find some strain of change that might tell us something. There was something about white college educated voters that were moving away from Donald Trump, even though Biden won them quite, didn’t win them, but saw a swing towards him. And it was just low samples. It was very poor sampling. It did happen to a tiny extent. Problem is everyone else swung the other way, you know.
just moving away from polling slightly. In America, of course, they do early voting, which was very popular during COVID, something Trump tried to denounce very quickly, didn’t this time around. whatever, even the most optimistic of models, and I tried this, even the most optimistic models for Harris of early voting still pointed to Trump sweeping that state. The data was there. The data was there to point to a clear Trump win in so many of these states.
But so many of us, myself included, I regret to say, saw that and thought, surely not. Why don’t we just stick to the what we know, the opinion polls? Why don’t we stick to that? Try and crunch that, try and model through that. I think someone like myself, so used to covering it from that angle, we need to just take a step back, take a beat and just realize, look at it from a different perspective. Always be innovative, always consider alternative methods. And I did, I saw it, but I didn’t believe it.
GEORGINA:
And I suppose there’ve been so many method failures now, some quite notorious ones, that you’re always perhaps a bit on the lookout for one of those as well. So when you get these results that look a bit anomalous, that’s another instinct as well, right, at this point to question, well, is that just something that’s gone wrong with the methodology? And maybe to sort of overcompensate for that possibility by toning it down a bit.
BEN:
Yeah, one example is multi regression MRP polls, you’ll hear them. They’re the gold standard apparently, they’re gold dust. you send to a newspaper, we’ve got an MRP poll for you, they’ll drop everything and say, right, we want this on our front page because MRP, everyone knows it’s like 50 to 100,000 people. Everyone knows it can give you seat by seat breakdowns. But the truth is it’s just as unreliable as a poll, because it’s the same assumptions and basically it’s one poll. It’s the assumption that
You’ll get a decent sample of women, men, 18 to 24 year olds. But if it’s from one body, one group, they have a certain skew, you know, it’s better to stick to all the polls and look at models. I’m a shameless self-promoter. Look at models like mine over MRP. Sorry.
What do the next few years look like for you at Britain Elect? Obviously, we should be quite far off a general election right now. What are you going to be doing in the meantime?
I know. I’m always wanting to do a bit more with Brit and Alex. think 2025 was a pretty bad year for Brit and Alex. My day job got in the way. I was couch surfing for four months. I was moving into a new house. just, all of this kind of just fell by the wayside a little bit. Now though, trying to get in more, because I’m realizing there’s a lot of fake data analysis out there. Not to sound like Donald Trump, but there’s just a total miss, a wilful misinterpretation of data by some.
popular Twitter accounts and others that just needs calling out. I am going to do more, I suppose, but this year, in Britain at least, we’ve got the Welsh Senate elections, which to tell you the truth, politically for Britain could be the success or not of Keir Starmer’s premiership because I think the minute Labour loses Wales, if indeed that happens, and that’s what the polls say, I think so many in the Labour rank and file will say, right, that’s it, that’s it. So covering that is absolutely imperative who succeeds, absolutely imperative. It’s going to be business as usual, but we have so many more elections, so many more pivotal elections in France in which the far right and the former Jordan Bardella and others are leading and not just be leading by a little bit, leading by quite a bit. Always remember this, even in Emmanuel Macron’s worst throws of his first term, never did Marine Le Pen top Macron in the head to heads.
Now we have Bardella at the far right, the Rassemblement Nationale, formerly National Front, polling ahead in the two-party trial heats. These are abnormal times and we got to be accurate and good and cut through that noise as much as we can.
GEORGINA:
Great, lots of things to look out for this year then. And if listeners would like to see what Ben is up to this year, he’ll be publishing data analysis from his journalism work on the New Statesman website and in the publication. And I’d encourage listeners to also look out for Britain Elect. There’s a website and as I mentioned, they post things regularly on X. If people are still willing to go on.
And on Blue Sky and there you go, it’s on all the main social medias if people would like to go and look at the latest results. Ben, thank you so much for talking to me today. Thanks for listening to this week’s episode of Talk Data to Me. I’m your host Georgina Sturge and I’ll be back next week with a brand new episode.
Originally published on Substack: https://georginasturge.substack.com/p/the-art-of-election-polling-with