2020 isn't 2016. Because Joe Biden isn't Hillary Clinton.

In a presidential election year that is so unlike any previous one, it’s not hard to compile a list of oddities. But one peculiarity is an almost mystical feeling in each party that is hardly analytical but quite widespread. Many of the most fervent Donald Trump supporters seem convinced that because he pulled a rabbit out of a hat in the last election, winning a race that seemed unwinnable even as late as the first poll closings, he can and will do it again. The polls and pundits were all wrong, they maintain; Trump has some magical ability to defy political gravity and all the experts put together. The mirror-image conviction among many Democrats and passionate Trump detractors comes from the psychological trauma of losing a race they thought they had in the bag in 2016. They fear, even believe, that he will somehow do it again.

This political phenomenon keeps coming to mind when I look through the results of new polls, most recently the Thursday afternoon release of the latest Quinnipiac University poll, conducted June 11-15 among 1,332 registered voters nationwide, followed by a Thursday evening release of a new Fox News poll, taken June 13-16 among 1,343 registered voters. In Quinnipiac, Joe Biden led Trump by 8 percentage points, 49 to 41 percent. That’s a bit narrower than the 11-point margin in the previous (May 14-18) survey, when Biden had 50 percent to the incumbent’s 39 percent. Fox has Biden’s lead at 12 points, 50 to 38 percent, up from the 48-40-percent spread in its May 17-20 poll.

The new Quinnipiac Poll showed 42 percent approving of Trump’s overall performance against 55 percent disapproving. The Fox poll had 44 percent approving to 55 percent disapproving. Both of those sets of numbers are nearly identical to the respective polls’ May numbers.

Looking across the span of all of the polling over the last six months, Trump’s approval ratings started a downward turn in early April, while the gap between Biden and Trump seemed to widen starting in very late May or early June.

Presidential elections when an incumbent is not running are pretty much choice elections, but when an incumbent is seeking reelection, they amount to little more than a referendum on the incumbent. That’s why so much focus is paid to job-approval ratings, the best yardstick of whether an incumbent will be reelected or not. Quinnipiac’s 42 percent approval rating splits the difference between the RealClearPolitics average of 43 percent and the 41 percent in FiveThirtyEight’s averages—not where an incumbent wants to be.

This mythology that because Trump did it once he can do it again ignores the fact that Hillary Clinton is not going to be the Democratic nominee this year, and that is quite consequential. In the late June 2016 Quinnipiac poll, 37 percent had a favorable view of Clinton and 57 percent had an unfavorable view (for a net of minus-20). In the three Quinnipiac polls taken in January through June, she averaged 38 percent favorable, 57 percent unfavorable (minus-19). By comparison, 34 percent had a favorable view of Trump in the late June 2016 poll, to 57 percent unfavorable (minus-23). For the three January-to-June polls he averaged 34 percent favorable, 59 percent unfavorable (minus-25).

But what about this year? In the new Quinnipiac survey, 42 percent had a favorable view of Biden, 46 percent unfavorable (minus-4). In the five polls so far this year, Biden has averaged 44 percent favorable and the same unfavorable. Trump was at 40 percent favorable, 56 percent unfavorable (minus-16) this month. For the year, he is averaging 45 percent favorable, 55 percent unfavorable (minus-14). Biden actually fares much better in the new Fox polling at 53 percent favorable, 44 percent unfavorable (plus-9). Compare that to the 43 percent who had a favorable view of Trump, to 56 percent unfavorable (minus-13).

While I believe this election will be almost exclusively a referendum on Trump, Biden’s favorables compared to Clinton’s at this time four years ago argue forcefully that what was going on in 2016 is hardly the case this year. Partisans on either side can continue to believe what they want, but this time is different.

This story was originally published on nationaljournal.com on June 19, 2020