Joe Biden is immensely popular and Democrats are ascendant. That’s the narrative the polling industry would like you to believe. Yet, in an environment where pollsters have routinely published polls with recalled samples of Biden voters +14 or more (i.e. nowhere near reality) to fluff the new administration, this latest effort from the Associated Press really stands out.

We’ll get to the details in a second, but here are the topline numbers.

New @AP poll out this morning: Biden’s overall approval rating sits at 63%. When it comes to the new president’s handling of the pandemic, 71% of Americans approve, including 47% of Republicans.https://t.co/BPBcrNMKtU

— Julie Pace (@jpaceDC) May 10, 2021

There are outliers, and then there’s claiming that Joe Biden has a 63% approval rating. How they got there becomes clear once you dive into the methodology. For example, their sample came out to 46% Democrat and only 17% Independent. That comes nowhere close to any current tracking of political affiliation. The poll also admits to purposely oversampling blacks and Asians for the reason of “analysis.” What that actually means is not elaborated on. In the end, they sampled blacks at 25% of the population when they only actually represent around 12%. Asians were sampled at nearly 20% of the population while only truly representing around 6%.

Later, the poll claims to weigh based on census data, but they conveniently don’t reveal the sample in relation to the 2020 election nor exactly what their ending sample turned out to be. That leaves plenty of room for suspicion that the ridiculousness of their original sample wasn’t remedied by their weighting, thus you get Joe Biden’s approval not matching any other indicators we are currently seeing. Playing games with samples is a favorite pastime of the polling industry these days.

Even just comparing the AP-ORC poll to other recent polls shows how unreliable it is. Looking at the major other polls in May, these are the other results: 50%, 55%, 49% via Rasmussen, Reuters, and YouGov, respectively. The poll is also not of registered or likely voters, making it non-predictive in regards to the political environment.

Lastly, I’ll note that this massive approval, even if it truly exists, is not translating at the ballot box. Democrats were just shut out of a Trump +3 district in Texas via a special election. The Republican candidates took over 60% of the vote and the top two finishers who will go to the runoff are goth in the GOP. And while the turnout was low, as is to be expected, those same dynamics of lower turnout will exist in 2022 because it is a non-presidential election.

These polls are meant to spin a narrative, in my opinion. The samples we are seeing are absurd, and the lack of transparency doesn’t help the case. Regardless, that narrative is not going to save Joe Biden’s party in a year-and-a-half.