Throughout the race, the media has been pushing a narrative that Trump’s support amongst both Republicans and Independents has slid since his election in 2016.  As poll after poll rolls in, the results say that Biden has a mounting lead on Trump.  Just since Septemeber first, not one poll in on RCP has Biden leading by anything less than 5 points, with only one (Rassmussen – one of the most accurate of 2016) giving Trump a 1 point lead.  Topline results may suggest this result, but the crosstab data does not support this suggested outcome.

In comparing two polls, one from this time in 2016 and one current poll, we can see that if anything, Trump has gained support amongst Republicans and Independents.

In an Economist/YouGov poll from September 10-13, 2016, Trump garnered support from 83% of Republican voters and 34% from Independent voters.  Also, 6% of Republican voters and 37% of Independent voters pulled for Hillary.  When factored for ideology only 74% of Conservatives leaned towards Trump (14% for Hillary) and 29% of Moderates slid towards Trump (49% for Hillary).

In The Economist/YouGov poll from September 13-15, 2020, Trump earned 92% of Republicans support (with Biden only 5%) and 37% of Independents’ support (with Biden at 41%).  Again, factoring for ideology, Trump earns 86% of Conservative support (compared to Biden’s 11), and 32% of Moderates’ support (53% for Biden).

These polls have two different topline results.  The 2016 poll had a 42 to 40 split for Clinton, while the 2020 poll had a 50 to 41 split for Biden.  One may begin to ask the question that while Trump’s support amongst Republicans increased from 83% to 92% and his support amongst Independents increased from 34% to 37%, how could these polls be right?  Additionally, when we look at the ideology results from both polls, Trump’s conservative support increased from 74% to 86%, and his moderate support increased from 29% to 32%.  The way these poll samples are being weighted is obviously incorrect.    For starters, the 2016 polling sample had a 38 to 28 split (10 point advantage) Democrats to Republicans.  The 2020 sample gave an impossibly wrong 40 to 25 split (15 point advantage) Democrats to Republicans.  When you over-sample Democrats by at least 11, and your poll gives the preferred candidate of that over-sampled group an 11 point advantage, you know your poll is wrong.

What this simply means, and what has been repeated time and time again, is that these polls can’t be trusted.  The media’s own polls say that Trump has gained support amongst both Independents and Republicans.  How can they possibly say that Trump has lost support? Even worse about this is that they know just how wrong their 2016 polls were and now they are doubling down on the same mistakes?  Trump’s candidacy certainly is polarizing, but if anything it has forced people who would otherwise have been soft support, into a more solidified position for Trump.