The Telegraph has picked up an interesting story from the Los Angeles Times: “Polling putting Democratic Socialist in the lead was fake, admits firm”
Polling that projected a Democratic Socialist candidate had a commanding lead over her main rival was fake.
Progressive Francesca Hong was seen as a front-runner in the primary to be the Democratic nominee for Wisconsin governor, but went on to lose.
Median Strategies, which billed itself as a trustworthy, non-partisan polling company, projected that Ms Hong had a more than 20-point lead over moderate David Crowley.
However, on Aug 11, she narrowly lost to Mr Crowley, handing establishment Democrats a win following a wave of victories by Left-wing candidates.
Median Strategies has now admitted the polling data were faked and part of a “short-term social experiment”, following a report by The Los Angeles Times.
Note that, although in that primary Median Strategies falsely raised the hopes of the extreme Left, in another contest the company put out fake polling that exaggerated the lead of the establishment Democrat over the extreme leftist:
The company had also released incorrect data showing Karen Bass, the mayor of LA, had a double-digit lead over Nithya Raman, a city council member and Democratic Socialist, before the mayoral run-off in November.
I don’t know quite what to think of this. The Telegraph commenters, few of whom appear have read as far as the part about Bass vs Raman, think that this proves their long-held suspicion that all polls are lies put out by left wingers to artificially boost support for left wing ideas. This suspicion is not true. Polls are frequently wrong, they are sometimes skewed by the political biases of the people who set the questions and interpret the answers, but they are not generally made up. Reputable polling companies like YouGov and Ipsos scarcely make any profit from their political polls; their real business is market research. They do the voting intention polls to demonstrate their honesty, and to show that opinion polls can tease out meaningful information from the background noise. And I hate to say it, but millions upon millions of people truly support left wing ideas because it would be so nice if you really could make people rich and safe and good by decree.
Getting back to Median Strategies, I see that the US media fell for the hoax. For once, I do not blame them. It was a genuinely interesting experiment but it destroys trust. Perhaps that is a good thing in a world full of increasingly sophisticated lies, but, somehow, this obscure story about some fake polls in a foreign country disturbed me more than one would expect.
“The Telegraph commenters, few of whom appear have read as far as the part about Bass vs Raman, think that this proves their long-held suspicion that all polls are lies put out by left wingers to artificially boost support for left wing ideas. This suspicion is not true.”
And I’ve got a bridge for sale. Every single poll ever understates the right wing vote, and overstates the left wing vote. Funny that. There’s never been a poll that says the right’s vote is higher than it ends up being. One of two things must be true – either this is deliberate, or its a known error but they do nothing about it. When an ‘error’ happens in only one direction is by design, not happenstance.
Its a given that the sort of people who work in these companies will be Lefties, its all media work. Do you really think such people are going to tell the truth about the popularity of their personal political opponents? Would the type of people who work in polling in the UK be happy to promote the popularity of Reform or Restore? Of course they wouldn’t, they’d do everything they can to make sure the Right’s true poll numbers are suppressed.
Leftyism is ALL about control. Without control it withers because people choose something else. Controlling polling data is a key part of this. Oh look! Everyone agrees with my woke worldview and if you don’t then you’re a bad person.
Why, for example, did the government just veto publishing immigrant crime statistics? Well because the truth would fuck their intended narrative.
Jim,
I think a lot of people on the right (for various reasons) don’t “admit” it in public.
Jim writes, “Every single poll ever understates the right wing vote, and overstates the left wing vote. Funny that. There’s never been a poll that says the right’s vote is higher than it ends up being.”
There have been plenty. Most recently, those who followed the 2024 election will have noticed that YouGov’s flagship large-sample MRP model overestimated Reform support in several constituencies in 2024. For instance the Reform candidate in Barnsley North was predicted to get 34% of the vote but actually got 29%.
Going back a few years, nearly all polling for the 2017 general election
overestimated Conservative supportCorrection: it actually got the Conservative vote share about right but badly underestimated the Labour vote share.It is true that in UK polling there has been a consistent pattern of pollsters underestimating the number of “shy Tories”, although they do try to correct for that (and have over-corrected in some elections since 1992). They have a commercial incentive to get it right if they can. There are two opposite errors made about polling; one is thinking it is a magical crystal ball, the other is thinking it has no predictive value. Political movements that want to win need to have an accurate view of the world and properly conducted polling for all its faults is the best tool they have to do that. An educated guess is better than nothing.
There is also a consistent pattern of pollsters underestimating how popular the extreme left are. The Greens outperformed the polls in the 2024 election, and, as mentioned above, Jeremy Corbyn’s Labour outperformed the polls in the 2017 election.
While opinion polling has often underestimated support for right-wing parties, some recent examples have overstated support for Reform. At the Gorton and Denton and Makerfield by elections many polls predicted a better Reform performance than was achieved. Admittedly these were by elections so their applicability to a general election may be limited.
Jim. I agree with NickM that quite a number of people intending to vote right will lie to pollsters about their intention. My understanding is that some more responsible polling companies have taken steps to counteract the tendency of their polls to exaggerate support for the left but these have been less than entirely successful; the steps have reduced the extent of the exaggeration but not eliminated it. I have now read Natalie’s post and bow to her more recent information and knowledge.
On a different level, it is not at all clear to me that polls exaggerating support for the left help the left. They may have the result that less fervent left wing supporters assume a left victory and don’t bother to vote.
My view on the publication of fake polls is that it is both irresponsible and immoral because they may affect the outcome of the vote. It does not cure this to fake one poll in one direction and and another in the opposite direction.
My reference to Natalie’s post should have been to her comment.
In 2019, we had an Australian Federal election that the polls got wrong, predicting a Labor win. They were paying too much attention to the urban centers.
Anecdotally I have heard from a handful of normie centrist colleagues who didn’t vote in the EU membership referendum in 2016 but subsequently became rather radicalised remainers that they didn’t vote because they thought the polls and bookies odds (Remain was 1/10 odds on on the referendum day) showed Remain would win easily so didn’t consider it sufficiently important to vote.
Since polling is rooted in trust, I assume this polling company will now be going out of business?
I’d commend to you all this great scene from Yes Prime Minister that discusses this exactly. Note the point Humphrey makes about “reputable” companies. But when your reputation is gone you might as well be a dude in pajamas sitting in his mother’s basement posting his opinion on X.
If this had been short-term medical experiment – without informed consent – then the perpetrators would be prosecuted. I do not believe Median Strategies can be excused.
“in UK polling there has been a consistent pattern of pollsters underestimating the number of “shy Tories”, ”
What evidence is there for this ‘shy Tory’ effect? It is really just a way for the polling companies to escape their failures. ‘Its not our fault, the people keep lying to us!’. Strangely however only those on the Right are liars though, everyone else just tells the truth all the time.
The ‘shy Tory’ effect is just a cope, an excuse, a way of hiding the anti-right pro-left bias of the polling companies. If they aren’t capable of accurately assessing the support for Right wing parties then they should stop doing political polls. To produce polls that consistently underestimate one side of the political argument (whether by accident or design) is itself a decision to put your thumb on the scales.
@Jim
What evidence is there for this ‘shy Tory’ effect? It is really just a way for the polling companies to escape their failures.
I’m not sure that is entirely fair. If a polling company calls me I’m not answering their dumb questions, there is a systemic bias in these things. For example, if they conducted a self serving poll where the question was “Are you the sort of person likely to answer a polling question?” then they’d get pretty close to 100% result. But that doesn’t reflect reality.
The real question is do they use the real results of elections (or whatever is being measured) to correct for this “reluctance”. If they do it is a good thing, it they don’t then their data is garbage.
Ultimately we can determine how good a poll is by looking at their historical success with the same methodology. If they are consistently right we can trust them, if not then we cannot. It is extremely simple to test. However, polling companies are like weathermen and economists — they can be wrong most of the time and still, somehow, keep their jobs.
Brings out an interesting (to me) legal discussion.
Generally, if someone publishes a poll and screws it up, that’s life. Some people wail and gnash their teeth, others smile, and life goes on.
But when someone intentionally – admittedly – publishes a lie for fun and giggles, some interesting pathways might open up.
If a candidate or a party can show that it acted in reliance in some way on that poll (e.g., “we didn’t campaign in Chicago based on that poll”), and can show damage from acting thus, there may be some lawsuit potential.
Admittedly a stretch – but new legal theories of action are devised all the time.
@bobby b
and can show damage from acting thus, there may be some lawsuit potential.
Yeah, denying a democrat the opportunity to become a state governor and preventing her from millions of dollars of income from state corruption — I mean the potential damages could be huge!!
The real fraud occurred in the first round of the election.
The pro Fidel Castro Mayor Karen Bass is the worst Mayor in the history of Los Angeles – not even a articulate Marxist, Karen Bass just starred at the cameras when asked questions about the Los Angeles fire – on her return from her absurd trip to Africa, her general record is awful – and Karen Bass was declared to be in first place.
“Well Paul – perhaps there are just lots of Communists in Los Angeles who will vote for a fellow Communist, no-matter-how-bad-her-record-as-Mayor-is” – that does not explain the following….
An fairly obscure Communist (Marxist – the “Democratic” Socialists of America is a Marxist organization) from India was declared to have come second.
There may be a lot of Communists in Los Angeles – but first AND second?
Meanwhile the candidate who made all the running in the campaign, Spencer Pratt, whose home was destroyed in the Karen Bass fire (a fire started by a leftist activist – the media leave that out, and made vastly worse by the neglect by Karen Bass and her DEI hires) and whose internet films made all the running in the campaign – was declared to have come third.
So the non Communist, who had made all the running in the campaign (both on the internet – and in the debate) “came third” – and will not even be on the ballot in November.
Mass mail-in-ballots = fraud.
But the “mainstream media” (and the education system) look the other way.
Wake up – it does not matter which of the two Communist candidates “wins” in November – Los Angeles is dead either way.
Was the web page announcing the fakery real or fake? Nobody knows.
“every poll understates support for the right”. Agree and it makes me wonder: what if we could actually get an honest poll, or better, an honest election, with honest media coverage leading up to the election (that should be part and parcel of “honest election”)? I suspect you’d find that the US electorate is not 40-40-20 (right-left-swing) as we’ve been told, with the 20 deciding things, but rather its more like 60-30-10. 2:1 right:left and the “undecided” shrinking because they are getting more and better info.
But that is a parallel universe we will never be able to access!
If only AI and the internet could be harnessed for Good, no one would have any excuse for not having good information to base a vote on. And if pigs had wings…
The whole story is manufactured to get people interested in the “election” in Los Angeles.
There is no real election in Los Angeles – both candidates for Mayor are Marxist-scumbags.
The election ended when the first round was rigged – blatantly so, in order to exclude the non Marxist candidate.
Rigged elections, not rigged opinion polls, are the matter of real importance.
For many elections are rigged – many of the major cities are utterly twisted with electoral corruption (mass “mail-in-ballots” fake registration of voters, and other fraud), winked at (indeed encouraged) by State governments – such as the despicable regime in California (including its corrupt judges – whose corruption goes back many years, to their over turning the effort to stop the handing out of taxpayer funded benefits and services to illegal immigrants – in a different time the corrupt judges who did that would have been dragged from their seats, stripped, tarred-and-feathered, and hung from the nearest trees – what they did has destroyed California) – but also several other States as well, including the Republican Governor of Georgia – who did nothing about the rigging of the 2020 Presidential election in Georgia – and the rigged United States Senate elections in Georgia.
To ignore rigged elections – and to go on about the antics of this-or-that polling company – is absurd.
There is nothing lower than a leftist judge – their leftism betrays the fundamental principles of law (of jurisprudence) – the very principles a judge is supposed to defend to the death.
There is something just as low as a leftist judge who ignores law, and rules according to his preferences.
It’s a rightist judge who does the same, for his own preferences.
But there does seem to be more of the former than the latter.
It would be interesting to learn why there are so many rabidly leftist judges, and relatively fewer rabidly rightist judges.