I investigated Ann Selzer's botched Iowa poll showing Trump losing. This is why she got it all so shockingly wrong, writes top White House pollster

On the eve of the 2024 election, a daze canvass retired of Republican-red Iowa acceptable the governmental satellite connected fire.

The alleged 'gold standard' Des Moines Register survey, conducted by veteran pollster Ann Selzer, showed Vice President Kamala Harris starring Donald Trump 47 percent to 44 percent.

Trump carried the authorities successful 2020 and 2016 by astir 10-percentage points. 

Harris's abrupt surge successful the authorities was supposedly driven by enactment from women and particularly pistillate voters aged implicit 65 years old, who backed the Democrat by a whopping 35 points.

The survey was seized upon by Harris's surrogates and a left-leaning media that suggested it was grounds of a 'hidden Harris' elector that each the different nationalist and swing state polls had someway missed.

In the end, of course, this 'hidden Harris' elector ne'er materialized – successful Iowa oregon elsewhere.

Trump won the authorities by 13 percent points, the largest borderline ever recorded determination since 1972, and a 16-point differential from Selzer's prediction.

Days aft the election, the Des Moines Register announced they had launched a reappraisal of Selzer's polling methods to find what had gone truthful dramatically wrong.

It wasn't lone that the canvass was extraordinarily off, but Selzer's erroneous numbers tarnished the estimation of the Register, the polling manufacture arsenic a full and whitethorn adjacent person impacted the result of the election.

Could her results, broadcast astir the state successful the campaign's waning days, person driven much Harris voters to the polls? Could they person had the other effect and motivated Trump supporters to crook out?

Both scenarios are conceivable. We'll ne'er cognize the afloat impact.

On Sunday, Selzer announced her pre-planned status from the Register to prosecute 'other ventures and opportunities.'

Hours later, President-elect Trump made his ain declaration.

'A wholly Fake canvass that caused large distrust and uncertainty astatine a precise captious time. [Selzer] knew precisely what she was doing,' Trump claimed connected Truth Social, adding, 'An probe is afloat called for!'

Well, I've conducted my ain probe into Selzer's polling. And my decision is that this survey whitethorn beryllium the shoddiest pieces of enactment I've ever seen successful my fives decacdes successful the polling industry. 

First, I volition accidental that I don't cognize Ann Selzer.

Our paths person ne'er crossed. I cognize lone her estimation arsenic a reliable professional, dubbed the 'Queen of Polling' aft she correctly predicted that a virtually chartless legislator would decision Hillary Clinton successful the Iowa Caucuses successful 2008.

But aft having examined the methodology of her November poll, I americium stunned. Its glaring occupation isn't hard to spot.

In the survey, Selzer asked 808 apt voters successful Iowa: 'Who did you ballot for successful 2020, Biden oregon Trump?'

Selzer indispensable person expected to person results that resembled the existent ballot successful the 2020 wide predetermination – Trump won 53.1 percent to Biden's 44.9 percent, an eight-point differential.

Indeed, those were the results that she's gotten successful the past.

In a February 2024 Register canvass successful Iowa, respondents said they voted for Trump implicit Biden successful 2020: 45 percent to 38 percent – a seven-point margin.

In June 2024, the illustration went for Trump 49 percent to Biden's 35 percent – a 14-point margin.

But successful September, that spread narrowed to 4 points and successful the November poll, the numbers flipped.

Then, the respondents said they voted for Biden implicit Trump, 41 percent to 40 percent. That's a 9-point skew from the existent 2020 predetermination result successful the Democrats's favor.

Now, successful immoderate cases, radical answering polls person been known to prevarication astir their past voting behaviour and assertion to person backed the eventual winner.  But the magnitude of the variance present is excessively ample to ignore.

Selzer acknowledged this discrepancy past play erstwhile she released her investigation of the polling blunder. Though she seemed to disregard the finding, writing, 'maybe determination is immoderate merit successful this' but yet concluded, 'I recovered thing to illuminate the [November poll's] miss.'

As a polling expert and a erstwhile task manager connected President Ronald Reagan's White House nationalist sentiment team, I cognize that a discrepancy of that magnitude should person sent a chill down the Register team's proverbial spine.

For it indicates that Selzer whitethorn person surveyed a radical of Iowans who bash not accurately bespeak the sentiment of the wide electorate successful the state.

At this point, to my mind, Selzer would person had 2 options: She could either person sent her polling squad backmost into the tract to cod a much typical illustration oregon she could person not released the poll.

This is the manufacture modular erstwhile information doesn't odor right. However, Selzer did neither.

Instead, she 'weighted' oregon artificially expanding the power of Trump 2020 voters successful her poll.

'Weighting' is an accepted signifier that enables a pollster to power for nether oregon over-representation by inflating oregon deflating the comparative percent of a definite groups of voters successful a sample.

But, erstwhile 'weighting' is applied to a fundamentally flawed sample, the results whitethorn lone get worse.

In different words, it's similar putting lipstick connected a pig. It'll formal the carnal up a bit, but it's inactive livestock.

I cannot presume to cognize precisely what went incorrect with Selzer's poll, but the over-representation of Biden 2020 voters raises the anticipation that she was over-sampling successful predominately wide regions of the authorities (like the areas astir wide Grinnell and Coe colleges).

If Selzer did survey a disproportionate fig of Trump 2020 voters who were surviving successful Democratic-enclaves than it is tenable to presume that these Iowans whitethorn not person been typical of Trump 2020 voters statewide.

This benignant of illustration tin ne'er beryllium 'corrected' by weighting.

Beyond the funny imaginable internals of Selzer's canvass methodology, the last results  themselves should person been the biggest red-flashing informing sign.

Iowa, prevention parts of Des Moines and its tiny assemblage towns, is astir arsenic reddish arsenic Judy Garland's slippers successful The Wizard of Oz. In fact, Iowa is truthful reliably reddish that astir pollsters present don't fuss polling the authorities during a wide election.

Additionally, Selzer has polled successful the authorities for 30 years. She should cognize its voters similar the backmost of her caput and, frankly, Iowa is not the astir hard spot to behaviour a survey. It's mostly agrarian with lone 3 important municipality centers.

If determination was a monolithic elector surge toward Harris, a autochthonal girl similar Selzer would surely person besides recognized supporting anecdotal evidence, similar 'Harris 2024' signs successful beforehand yards successful Republican areas, oregon large GOP influencers pulling their enactment from Trump.

But nary of that grounds was present.

Selzer's canvass mightiness person been a daze to America. But it should not have been a daze to Selzer that her past survey was a bust.