Share on twitter
Share on facebook
Share on linkedin
Share on email

An Election with Surprises: Inadequate Polls and Arizona’s Flip

The race between former Vice President Joe Biden and incumbent President Donald Trump was tighter than expected as voters turned out in record-shattering numbers.
Share on twitter
Share on facebook
Share on linkedin
Share on email
Related Posts

NATO 2030 Projection Faces Myriad Challenges

FeniksPolitik organized its first in-house event on the future of NATO and the challenges it faces in confronting a changing Middle East, the US pivot to Asia, increasing presence of Russia and China in the region and intra-alliance tensions tied to a Turkey in democratic decline.

Read More »

The race between former Vice President Joe Biden and incumbent President Donald Trump was tighter than expected as voters turned out in record-shattering numbers. According to pre-election polls, Biden was largely expected to win Wisconsin, Michigan, and Pennsylvania and lead slightly in Arizona, Florida, and North Carolina. Yet the margins of Biden’s victory were smaller than predicted.

Despite arguments that the polling industry oversampled rural and uneducated voters, who are commonly seen as the backbone of the Republican Party, polls still generally favored a win for Biden. This overrepresentation of rural and uneducated voters can be seen as a result of pollsters’ undersampling of this group in 2016 and their consequent failure to forecast Trump’s triumph. With this oversampling in mind, some expected Biden to receive even more votes than what was forecast. That is why The Economist, for example, predicted a landslide victory for Biden along with a 91 percent chance of winning. Yet after the in-person votes were counted on election night, Biden was behind Trump in many swing states, including Pennsylvania, North Carolina, Georgia, Michigan, and Ohio. Nonetheless, as mail-in and absentee ballots began to be counted, Biden eventually took the lead in Michigan, Wisconsin, and Pennsylvania, three of the most populous Rust Belt states and important components of the so-called Blue Wall but it was definitely not an “easy win” as pollster predicted.

The Blue Wall is a term used to refer to 18 states in the mid to upper northeast US that have a reputation of consistently voting Democrat in presidential elections, as witnessed in races between 1992 and 2012. The dynamics here changed in the 2016 elections however, as Hillary Clinton lost Wisconsin, Michigan, Ohio, and Pennsylvania. The most important reason for her loss of the Blue Wall is considered to be her failure in attracting the votes of young people and minorities (as Clinton’s campaign primarily focused on winning over suburban swing voters), and her inability to maintain the votes of the white working class. Blind to their weaknesses, Democrats were overly confident that this region was all but secure. In 2020, Biden’s successes in Wisconsin, Michigan, and Pennsylvania can be traced to his more favorable image among the working class electorate and the apparent distaste of voters for Trump. However, despite Trump’s perceived unsuccessful management of the Covid-19 crisis, Biden still did not win here by the expected landslide. 

In 2020, pollsters predicted two scenarios for Biden: a landslide victory or a very close win. It seems the latter is now the case and the race was once again determined by these three Rust Belt states. There are two possible explanations for this.

First, mail-in ballots pushed Biden to victory. Naturally these ballots largely favored Democrats, especially considering that the Democratic Party spent months encouraging its voters to vote early and to vote by mail whereas Republicans and Trump consistently questioned the legitimacy of mail-in ballots. We also know that Democrats tend to be more concerned and cautious when it comes to the Covid-19 pandemic, thereby opting to avoid gathering at polling places. Absentee and mail-in ballots were also received in record-breaking numbers. Altogether, it is projected that a combined total of 161 million votes were cast in this presidential election, the most in US history, and Biden won by over 74.4 million votes, once again, more than any presidential candidate in US history.

The second reason for Biden’s closer-than-expected win, and the inaccuracy of pre-election polling, is the “preference falsification” of the Republican electorate, or in other words, “shy” Trump voters. It is believed that many people might have stated their preference for Biden to pollsters even though they planned to vote for Trump because they saw this preference as more socially acceptable. In 2016, when polls confidently projected a win for Clinton, many Americans were shocked on election night as Trump took the presidency. This year, some scholars once again believe that polls undercounted Trump voters. Nonetheless, there are also scholars who argue against this, claiming that “shy” Trump voters do not appear to follow any discernable geographic or demographic patterns. If “shy” Trump voters indeed skewed the pre-election polls, their effect should have been witnessed equally in both Minnesota and Wisconsin. Even though polls were off by a wide margin in Wisconsin, they were more accurate in Minnesota. A similar pattern is also applicable for Texas and Arizona. Texas polls were less accurate than those in Arizona even though both states have rising Latino populations and similar social and economic characteristics. Similarly, polls were off by a wide margin in Florida, whereas the election results of neighboring Georgia were more accurately predicted. Such realities suggest that there is much more to the story than just “shy” Trump voters, with possible explanations being unpredicted voter turnout and/or Trump’s success in capturing the votes of socially disconnected groups who simply do not participate in polls.

What happened in Arizona?

Democrats were surprised to win Arizona, a critical swing state and traditional stronghold of the Republican Party. The last time it voted Democrat was for Bill Clinton in 1996. In 2004, Republican George W. Bush won the state with 54.6%, and in 2008 it favored its home-state Republican Senator and then-presidential candidate John McCain with 53.8% of the vote. In 2016, Trump prevailed over Hillary Clinton with 48.08% of the vote despite the state’s changing demographics that were believed to favor Democrats. Statistics now show that one in four Arizona voters are non-white and a slim majority are college graduates.  

Nonetheless, despite leaning Republican for over two decades, Arizona flipped Democrat in 2020. Similarly, in the senate race, Democratic senate candidate Mark Kelly won against Republican Martha McSally. Experts point to two major reasons for Arizona’s flip to the Democrats.

First and foremost is the tension between Trump and Arizona’s well-respected Republican Senator John McCain. McCain was a maverick politician and a former US navy officer who served as a senator of Arizona from 1987 up until his death in 2018. Regarded as a hero by many, the highly decorated Vietnam veteran was one of the most outspoken Republican critics of Trump, especially when it came to the president’s attacks on immigrants and attempts to repeal the Affordable Care Act. In return, Trump took public jabs at McCain’s military service, stating “I like people who don’t get captured”, in reference to McCain’s time as a prisoner of war. The attacks continued even in the wake of McCain’s death, when Trump declared that he wasn’t “a fan” of the late-senator and complained that he didn’t even get a “thank you” for giving him “the kind of funeral he wanted”. Due to this, McCain’s wife, Cindy McCain, openly endorsed Biden on social media. This dynamic is predicted to be one of the factors that secured Arizona for Biden. The second major reason for Arizona’s flip is seen in the demographic shifts in the state resultant of the resettlement of Californians to Arizona.

These historic US elections have taught us two important lessons. First, swing states are increasingly significant. The Democratic party should not take Michigan, Wisconsin, and Pennsylvania for granted, the Blue Wall may not be as impenetrable as previously thought. Second, while polling companies did forecast Biden’s triumph, it was much closer than expected. In the end, the political polling and survey research industries continue to face challenges to their tried-and-true methodologies that may ultimately require substantial rethinking.

*Çağla Demirdüzen is a Ph.D. candidate at Arizona State University’s Department of Political Science and Global Studies