February 17th 2020

Kerry Eleveld

Daily Kos

As enthusiastic liberals, it's sometimes easy in the Trump era to view the issues of the day in simple good vs. bad terms because everything Trump is so inherently horrible. But one area where I've been trying to infuse a bit of nuance into the discussion here at Daily Kos as we head toward November is on polling and what it will take to beat Donald Trump in 2020. 

While I don't have the time here to tease out an involved and considered discussion of polling, I would like to offer the simple premise that polling is not inherently evil, and that it's an important part of any informed discussion about the state of the race. The other thing to remember about beating Trump is that it doesn't take a wholesale defection of MAGA hats to determine his demise. Rather, it's half a point here and half a point there because his margins are so slim and he's done absolutely nothing to expand his base since 2016. Consider, for instance, that Trump beat Hillary Clinton in Michigan by less than half of a percentage point, 47.6% to 47.3%. So often in the comments, I see people dismiss diaries by saying it doesn't matter because the MAGA crowd will never change and his supporters don't care. Just remember that we don't need his cultists to change their minds. What we need is some independents here and some late-deciders there to decide they're either voting D this time or staying home because Trump's just too repugnant and purposely divisive.  

Okay, so enough of the programming notes and on with the post! One of my ongoing preoccupations has been questioning the conventional wisdom that Trump's historically low base of support is really as rock solid as people assume. The basic assumption is that if his job approval rating is consistently at about 42%-45%, then all those people are necessarily voting for him. That's an assumption that's worth questioning.

Along those lines, FiveThirtyEight had a story this week suggesting that Trump may be more unpopular than his approval ratings suggest. Spoiler alert: He's still pretty wildly popular among Republicans. But researchers from Cornell University developed a study to measure Trump's popularity beyond the simple question of job approval rating and found that he rates pretty low with independents, a small slice of the electorate that he absolutely must win to be reelected. 

The researchers, Peter Enns and Jonathon Schuldt, tested how respondents rated Trump relative to other notable Republican figures such as John McCain and Ronald Reagan and found that Trump ranked "at the bottom of the list." In the graph below, check out the middle section, where Independents rate Trump's favorability just below that of Sarah Palin, while George W. Bush, John McCain, and Ronald Reagan all rate considerably higher.

Another poll that caught my eye this week focused on Trump's abysmal approval ratings among people who are undecided in head-to-head match-ups between Trump and Democratic frontrunners in Arizona and Iowa. In 2016, Trump won Arizona by 4 points and Iowa by 9 points. When matched up against Democrats today, he's competitive in both states. But PPP writes:

In Arizona more than 80% of the undecideds in every match up between Trump and the individual Democratic candidates disapproves of Trump. And the undecideds also voted for Hillary Clinton in 2016 by anywhere from 45-50 points. If you allocate the undecideds based on whether they approve of Trump or not, all of the Democrats move into leads ranging from 4-6 points. If you allocate the undecideds based on whether they voted for Clinton or Trump in 2016, all the Democrats move into leads ranging from 2-4 points.

In other words, if it's the undecideds that determine the outcome in states that remain very competitive, then it's worth considering where those voters are likely to come down once we have a Democratic nominee. In fact, in the Arizona poll, once the undecideds are allocated according to their approval/disapproval of Trump, all the Democrats win the state. Biden and Trump, for instance, are tied 46-46 in Arizona before allocating the undecideds, but Biden wins the state 53-47 once the undecideds are allocated based on whether they approve of Trump.

Before I post the Arizona and Iowa charts, I'll add a word of caution since this is a narrative we would all like to believe. First, we're using the numbers of one polling outfit. Another poll, for instance, might not have the races so competitive, not to mention that a lot is left to happen between now and November. Second, I'm not exactly sure how big the sample sizes are of undecided voters. That said, they're pretty consistently horrible for Trump across the board, which is worth something. If they jumped around from match up to match up that would be more suspect. 

All that being said, I do think it's worth keeping in mind that the undecideds in these polls aren't necessarily neutral. For the moment, their indecision may be more rooted in the fact that we don't have a Democratic nominee yet. Okay, here's the PPP charts. 

Arizona

Iowa

Posted with permission