Former Vice President Joe Biden appears to be in a better position to win the November presidential election than ever before, with a number of recently released polls showing him with a significant lead over President Trump.
But despite a clear polling advantage for the presumptive Democratic presidential nominee, there are two key factors that have resulted in many public polls obtaining results that overstate Biden’s actual lead.
The first factor is that some polls use disproportionately Democratic samples.
For example, we can look to CNN’s general election poll released this week, which shows Biden beating Trump nationally by a 14-point margin – 55 percent to 41 percent.
Although Biden holds a national advantage over Trump according to all recent public polling, the CNN poll inflates Biden’s margin of victory by having a disproportionately Democratic registered voter sample.
With 32 percent of people surveyed identifying as Democrats – compared to 25 percent identifying as Republicans – this 7-point Democratic margin does not represent actual voter turnout on Election Day.
Indeed, according to CNN’s 2016 national exit polling data, the actual margin of voters identifying as Democrats compared to those identifying as Republicans favored Democrats by 3 points – 36 percent to 33 percent.
Further, an Emerson College poll of registered voters conducted at the same time still shows Biden ahead, but with his lead cut by more than half – to just 6 points. Accordingly, this poll used a sample that is more representative of turnout by party, with Democrats comprising 35.6 percent of the sample, while Republicans accounted for 32.5 percent.
These figures more closely align with party data from 2016 exit polls, leading me to believe that Biden’s actual lead is likely closer to 6 points, rather than 14 points.
However, there is a second factor that also should give us pause before relying on presidential polling: whether the poll is of registered voters, as opposed to likely voters.
Both the aforementioned CNN and Emerson College general election polls use registered voters. Although the Emerson poll was more accurately representative in terms of party identification, polling registered voters has historically overrated Democrats.
Polls of likely voters have been known in recent decades to be a much more reliable predictor of election results than polls of registered voters.
From 1990 to 2014, polls of likely voters showed virtually no bias towards either party and were largely in line with the ultimate results of the election. In contrast, polls of registered voters showed a median Democratic bias of 2 points, according to FiveThirtyEight.
Furthermore, according to the same FiveThirtyEight analysis, between 1990 and 2014 there has been a nearly 2-point margin favoring Republican turnout in presidential election years. Thus, polls of registered voters would inherently have a Democratic bias, given that more Republicans turn out to vote.
Additionally, public polls fail to factor in the clear enthusiasm gap between Democrats and Republicans. This was evident in Georgia, where turnout in the Republican primary for Trump substantially exceeded the combined vote for Biden and Sanders on the Democratic side. This has been reflected in most national polling done this year, to the advantage of Trump and his fellow Republicans.
Though these factors contribute to exaggerating Biden’s lead, the former vice president still has a clear advantage, meaning it will be an uphill battle for Trump to win reelection.
The president faces clear vulnerabilities in terms of his handling of the coronavirus pandemic, as well as his recent actions relating to the current civil unrest following the death of George Floyd – a black man in handcuffs who wasn’t resisting arrest – at the hands of Minneapolis police.
And some prominent members of Trump’s own party and retired military officers have recently denounced him for some of his comments and actions.
While Trump faces clear electoral issues among critical voter blocs – independent voters, suburban white voters and older voters – Biden also has cause for concern that goes beyond his overstated public polling lead.
According to an analysis of recent polling released this week by The New York Times, in the past few weeks Biden has failed to make inroads with non-white voters. That’s a crucial voter bloc that the Democrats will need to turn out in order to be successful in November.
Indeed, the analysis shows that Biden currently has a 46-point lead among non-white voters, which is less than Hillary Clinton’s 50-point advantage among these voters in the final polls of 2016.
Furthermore, despite the fact that Trump has lost ground with suburban white moderate voters, independents, and older voters – three critically important groups for both parties this year –Biden may find himself in an increasingly problematic situation as he tries to court these voters.
Given the credible threat of Democrats moving too far left with policy initiatives such as defunding the police – which many prominent Democrats have called for in recent weeks – Biden may have a much more difficult time cutting into Trump’s margin with these voters than public polling may make it seem now.
Further, if history is any guide, we can expect that some moderate Republicans and Republican-leading independents will return to the fold – to come home, as is said in politics. This is especially likely if Trump is able to convincingly make the case that he successfully dealt with the fallout from the pandemic and put our economy on the path to recovery come November.
Ultimately, taken together, aspects of public polling that exaggerate Biden’s lead combined with the challenges that both he and Trump currently face will make the result of the election much closer than it currently appears it will be.