By Matt Towery, InsiderAdvantage Pollster
Here it comes.
A well-choreographed combination of protests, news stories, and “legacy” polls, all of which will make it appear that Donald Trump’s actions and policies are already becoming unpopular. All “coinciding” with the first Congressional break since Trump was sworn in.
There will be well organized “invasions” of GOP congressional town hall meetings, pointed articles from the press about “the sudden collapse of the economy” and well-coordinated and staged events targeting Elon Musk.
Just this past weekend in my city the local “Patch” community online news site listed along with musical performances and cooking classes in its “Things to do this weekend” list, a “Tesla Takedown” at the local Tesla dealership. Sure. Just another fun event to enjoy on a beautiful weekend, right?
And this is where the pollsters come into play.
There are only a handful of pollsters who have been able to accurately capture supporters of Donald Trump, often called “shy Trump voters” in the three presidential cycles in which he has been on the ballot. And the reward my firm, InsiderAdvantage, and others such as my colleague Robert Cahaly’s Trafalgar Group have received for being accurate has been a concerted effort to hide our performance through convoluted “pollster grades” where raw accuracy takes a backseat to mathematical manipulations combined with subjective grading systems.
Academics and media embraced this nonsense by calling these pollsters who had erred in all three of the Trump contests, “high quality.” Those of us who have nailed the races have been, in essence relegated as second rate.
A few weeks ago, Robert Cahaly and I predicted on our podcast PollingPlus that this onslaught of negative polls for Trump would come rolling in. And right on cue they started arriving.
It was interesting that just a week before the Congressional break, our joint InsiderAdvantage /Trafalgar Group survey found President Trump’s approval rating five points above his disapproval rating. The equally reliable, and deemed unworthy by the polling mavens, Rasmussen Reports showed that same plus five-point number as the weekend approached.
But nevertheless, a host of pollsters who had difficulty polling Trump’s three elections unleashed radically different results at virtually the same time late last week.
I have no doubt there will be many more of these “the sky is falling” polls in the coming days. They will be readily embraced by the media with hardly a thought about the pollster’s past real results in polling Trump.
For example, in recent days both CNN and Quinnipiac produced surveys showing President Trump’s disapproval outstripping his approval by seven and eleven points respectively. Add to that a Reuters/Ipsos which had similar results.
Quite a sudden drop, eh?
The issue of tariffs, off many average voters’ radars, along with one week’s worth of market gyrations simply cannot move that many people from approving to disapproving Trump’s job performance.
And the same CNN poll showed record low approval for Democrats. That makes a lot of sense, right?
According to the RealClearPolitics Pollster Scorecard, which ranks pollsters purely on their error rate, both CNN and Quinnipiac had error rates of seven percent or more in the last Trump election RCP evaluated (2020). In that same election, firms such as InsiderAdvantage and Trafalgar Group, had error rates of less than three percent and ranked at the top for all pollsters. Rasmussen was right there as well. The same firms ranked at the top in RCPs 2016 rankings. And our polls were even more accurate in ‘24.
There is a certain art to finding respondents to a poll who are willing to say they vote for or support Donald Trump. Perhaps that is why firms such as Insider, Trafalgar, Rasmussen, Emerson, Harris, Atlas and a few others have had consistent success in polling Trump elections. And perhaps that would suggest that our surveys on Trump’s job performance provide a more accurate portrayal of the public’s opinion on that issue.
But in the coming days I predict you will see more and more so called “quality” polls with surprisingly negative numbers for Trump. They will enter the mix of protesters and staged events giving us the impression that Trump is suddenly dropping like a rock, and that Elon Musk is some major political liability. Neither which, at present, is the case.
And the media will carry those polls because groups such as FiveThirtyEight gave them, despite their actual error rates, high marks. That led to FiveThirtyEight’s mistaken forecast of a close presidential result in November, with a slight edge going to a Kamala Harris victory! Of course, FiveThirtyEight was shuttered by its owner, ABC News, a few weeks ago. Forecasting has consequences I guess.
If you want to know just how deep this business of promoting pollsters who consistently fail to poll Donald Trump accurately goes, just look to ChatGPT.
On February 5th I randomly asked the AI chatbot why it excludes the straightforward RealClearPolitics pollster rankings when it reports on “who are the top pollsters in the nation?”
The ChatGPT answer then, as preserved in a screenshot, reads in part “RealClearPolitics methodology and transparency differ from..FiveThirtyEight…” So, after FiveThirtyEight was eliminated by ABC News, I asked it the same question of ChatGPT. The answer I received again stated that ChatGPT does not include the RealClearPolitics rankings of pollsters “because RCPs methodology and transparency differ from those of other organizations that provide pollster rankings.” No mention of its beloved FiveThirtyEight.
So, let’s get this straight. The oldest, best known, gold standard for reporting polls and polling averages, which happens to highly rank pollsters who were accurate in three Trump presidential cycles, can’t be part of ChatGPT’s report to the millions who will rely on it. Even when their rival, FiveThirtyEight, no longer exists.
In my opinion, ChatGPT would go through endless algorithmic gymnastics to avoid giving credence to those who accurately poll Trump. And that means something. The more you can marginalize those who can poll Trump accurately, the faster the other pollsters can take him down. Those who can accurately poll Trump’s ratings, be they good or bad, will be shoved aside or may disappear totally and be overwhelmed by the many who never could get his numbers right.
We have no idea where public opinion of President Trump might go in the coming months. But one thing I am sure of is that the polls that most of the media will hype will fail to capture many of his supporters.
They simply will not be accurate.
But don’t ask ChatGPT to back me up on that!
{Matzav.com}The post TOWERY OP-ED: Why Will Polling of Donald Trump Start to be Way Off? Just Ask ChatGPT first appeared on Matzav.com.