THE LAST FEW weeks have featured a dizzying array of analysis from people using polls to show President Biden is either within striking distance, out of striking distance, still in the race, definitely out of the race, competitive in swing states, not competitive in swing states, about as good as alternative candidates, much worse than alternatives, and everything in between. This has mushroomed into a battle royale for party leaders, interested super PACs, reporters, armchair forecasters, and even a stressful pastime for voters trying to figure out whether they think Biden should stay or go.
At the moment, the side trying to edge Biden off, stage left, seems to be winning the argument and voters seem to be on their side. But polls can’t really settle the argument over whether that’s the right decision for three main reasons.
First, analyzing polls in a close election can show you whatever you want to see and support whatever you want to say. There is no cutoff, no threshold a candidate must clear to definitely have a path to victory and no point of no return below which a candidate is doomed. Those still in Biden’s corner can adopt a familiar set of analyses to support their argument that he should stay in. Campaign pollsters are paid to find and highlight a path for their candidates to win and are very good at it. This kind of analysis helps candidates and campaigns make decisions, but as someone who has done it a lot, a lot of it is also non-falsifiable guesswork.
Analysts on Biden’s side would point to a mostly stable race where only a few things have to go right in a few places to push Biden over the edge and back to the White House. They can (and do) argue that Biden’s policies are popular, his values shared by more voters, and say they just need to get the word out to get credit from more voters.
They can say voters haven’t tuned into the election as a clear two-way choice yet. They could say Biden is universally known and still close, and alternatives are not well known. As Dan Kanninen, the Biden campaign’s battleground-states director, told The Atlantic this week, “Once we do define that choice on the issues—on the record of Joe Biden, on values, on what Donald Trump represents—we move voters to our camp. We know the voters whom we have to win back favor us quite a bit more on all those fronts.” And they can say many candidates trailing by larger margins later in the process have gone on to win (including Donald Trump).
Those wielding the stage crook are using polls to show there is no path for Biden and that he should drop out, a point which can also be made with the same data and style of analysis. They can point to Biden trailing in swing states, running behind other Democratic candidates, record low approval ratings, the apparent immovability of most views of Biden and Trump, and the very widespread view that Biden is too old to do the job effectively for another four years. They can also show lesser known replacement candidates apparently polling better than Biden and new messaging avenues that would be opened with a replacement.
Second, replacing Biden would be without historical precedent, with no reference points available for what might happen as the campaign unfolds. In thinking about a new candidate four months from Election Day, there is no model to run poll numbers through, no way at all to predict outcomes following a once in a lifetime event. Even models that have historical outcomes to rely on are built on just a few dozen presidential elections. Such as they are, the models do not show Biden’s cause as lost, with FiveThirtyEight’s model actually giving him a slight edge. A new candidate at this point introduces unpredictable variability in potential outcomes. Such variation could potentially go in either direction, with support for a new candidate soaring or plummeting.
The size of the shockwave a candidate replacement would unleash makes it impossible to assess ahead of time the impact on the candidates’ polling. Voters themselves have no idea how they would react to a new candidate and can’t tell pollsters what they would think even if they wanted to. They also don’t know what the events would look like, what the new and (by definition) lesser known candidate would do or say. It would also turbocharge media attention to the Democratic ticket throughout the rest of the cycle and immediately change the subject away from Biden’s age and the related palace intrigue.
Third, a new candidate would introduce the very real prospect different likely voters would show up in polls, voters whose frustration with the current options is pushing them to the sidelines. Turnout is about more than how many doors each campaign knocks or which side prints effective mailers. It’s also about candidates whose very presence on the ballot pulls different groups of voters to the polls. Looking at Congresswoman Ayanna Pressley’s 2018 Democratic primary victory illustrates the impact it can have when a completely different electorate shows up. As of now, marginal and disaffected voters can’t predict what candidate might pull them from the couch and to the polls this year.
When different voters opt into (and out of) the process, it doesn’t just change the candidate support figures in the crosstabs of the polls folks are obsessively poring over to make their points. It carries the potential of changing the size of different demographic groups who choose to participate. These changes can’t really be assessed by just asking whether voters would turn out if the contest included a candidate they know much less about.
These imprecisions make Democrats’ crossroads all the more perilous. They must choose a course based on a map that really only shows where they are now and a blurry, half erased outline of what might be the path if Biden stays in. People with different interests are drawing the rest of the map in whatever shape they want, and saying it’s what the polls show. None are necessarily wrong, but none are clearly right.
Even at the campaign’s end, all of the predictions and pronouncements made at this stage can’t really be proven one way or another. Consultants, pollsters, party insiders, and media on the losing side of their assessments can always blame – or credit – the campaign, candidate, or external events for why their version of the map didn’t pan out. The almost unfathomable pace of news events also offers ample excuses for predictions gone wrong.
Of course, with such a plethora of confident forecasters, some will almost certainly happen to guess mostly right and will be rewarded with years of consulting contracts and cable TV appearances. But for Democratic Party leaders trying to choose a path forward, there’s really no way of knowing who that might be.
Steve Koczela is president of the MassINC Polling Group.
The post Polls can’t really tell us whether Biden has a better chance than anyone else appeared first on CommonWealth Beacon.