Academic Writing

Explaining the Unexplainable: Voter Choice in the 2016 Election

There is plenty that is exceptional about the 2016 election—Donald Trump is the first individual elected president without having served in the military or government, Hillary Clinton and Trump were the least popular nominees in history, and there was notably high turnout. However, there is nothing particularly exceptional about the outcome of the election. By looking at political scientists’ traditional understandings of voter choice and data from the 2016 election, the argument can be made that voters largely acted consistently with accepted voter prediction models. Historical and contemporary data from exit polls show that voter alignment based on demographics such as class, ethnicity, age, and education have been relatively steady throughout the years, an effect that influences and is mirrored in partisan voter choices. Though there was a small—but significant—section of white working class voters who changed their vote from Obama to Trump, even these results can be explained and expected by conventions of elections and campaigns as understood by political scientists. Despite the unusual aspects of this election, the outcome in fact aligns with political scientists’ understandings of elections as far as considering the role that social and partisan identities and campaign strategy play in voter choice.

The most widely accepted model for predicting voter choice in an election is the Michigan Model. The Michigan Model suggests that people form “a psychological attachment to a political party” called party identification. Party identification functions as another kind of social identity such as class or ethnicity, which can, themselves, influence an individual’s party alignment. Sides et al. points out in Campaigns and Elections that in presidential elections, “approximately 90 percent of Democrats and Republicans vote for their respective party’s nominee.” This statistic is important in evaluating the uniqueness of the 2016 election because, in fact, the statistic remained the same for this election. In 2016, Trump, the Republican candidate, won 90 percent of Republican voters, and Clinton, the Democratic nominee, won 89 percent of Democratic voters. From a purely party identification lens, voter choice fell precisely within the expectations of political science conventions. Even among the demographic breakdown of vote share, the Democratic and Republican margins over traditionally held demographic groups stayed mostly the same in 2016. Early analysis of exit polls conducted by the Pew Research Center shows that Trump carried the white vote by 21 points, compared to Romney’s 20-point margin in 2012. Trump’s share of the older vote held strong as well—53 percent in 2016 compared to 56 percent in 2012. As for Clinton, her share of the women voters was 54 percent, reflecting the 55 and 56 percent share Obama received in 2012 and 2008, respectively. These results are not, or at least shouldn’t be, surprising, especially considering party loyalty has generally increased over time. Even those who self-identify as independent “are more loyal to a single party than voters who described themselves as strong partisans were in the 1970s.” Importantly, party identification is not about ideology and platforms, as Sides emphasizes, it is just as much an identity as something as innate as ethnicity. Thus, though many remark on Trump’s inaccurate representation of Republican ideology, as he often “routinely contradicted or dismissed longtime Republican ideals,” he still captured the Republican vote share because he never “disrespect[ed] Republican identities.” Party loyalty is so ingrained, Trump only had to avoid alienating the characteristics of his voter base to receive 90 percent of the Republican vote.

For a majority of the electorate, social and partisan identification accurately reflected which candidate they voted for. However, where voters diverged from such classification became targets in political and popular discussion in the aftermath of the election. The primary source of such conflict arose in the vote share of white, rural, non-college educated, working class voters. Different articles and studies focus on various parts of this conglomerate identity, but they roughly refer to the same voter group. Importantly, this group became known as the Obama-to-Trump voters and accounted largely for the turning of several blue states and counties in the Rust Belt to Trump. One article commented that “one in four of President Obama’s 2012 white working-class supporters defected from the Democrats in 2016.” Similarly, Trump carried a large margin of the rural vote—a 48-point advantage of rural white male voters compared to a 32-point advantage among white men nationally—and rural areas have a larger share (51 percent) of white working class voters than urban areas (22 percent). Additionally, college education tends to be correlated with higher income and class, making Trump’s margin among non-college educated white voters also relevant, as his 39-point advantage was “the largest among any candidate in exit polls since 1980.” Though party identification can explain some of the white working class vote, as those who already leaned or identified as Republican where 11 times more likely to vote for Trump, it does not explain the Obama-to-Trump voters. This voter behavior may seem atypical through the lens of the Michigan Model, but in fact is still accounted for by what Sides calls in The Washington Post the “time for a change” tendency.

In his article, Sides points out the historical preference for the White House to change parties after two terms; in this way, a Republican president seems only natural following Obama’s Democratic term. He explains this voter tendency in terms reminiscent of the retrospective theory of voting, which suggests that voters use past performance, usually economic performance, to decide between candidates. While the incumbency interpretation cannot be directly applied to the 2016 election as there was no incumbent running, voters still used the national economic situation to make judgments about the party in power in the executive capacity, and with low economic growth and presidential approval rating, voters could be predicted to be more likely to favor the Republican candidate in this perspective. Looking at retrospective voting and economic opinions is especially helpful in explaining the white working class votes for Trump. A survey conducted by the Pew Research Center in early 2016 illustrated the “depth of [rural white voters’] financial frustrations with the status quo on some key economic measures” such as jobs, immigrants, and eroding standard of living. Essential in this finding is the push back against the status quo and focus on economic issues that suggest that these rural white voters were engaging in retrospective voting. Interestingly, Nate Cohn presents a comparison between Trump’s and Obama’s campaigns that may explain the Obama-to-Trump phenomenon: “Both Obama and Trump ran as change agents, against the establishment and corporate interests.” That the campaign is suggested to have had any effect in persuading voter choice is another essential element in evaluating the extent to which the voter behavior in 2016 was unprecedented or not.

It is commonly accepted by political scientists that campaigns have very little effect on the outcome of presidential elections, as the candidates tend to be relatively equally matched in skill and resources. However, in the 2016 election, Trump “raised less money, aired fewer ads and had fewer field offices,” a fact that destabilized the equivalence and seemed to suggest an advantage for Clinton. Evidently, based on the election results, the supposed campaign advantage had no discernible effect on the outcome. Despite this demonstrated lack of effect of campaigns, the particular campaign strategies of Trump and Clinton in their geographic focus has implications for the shift of white working class voters. The failure of the Clinton campaign to solidify her white working class base and Trump’s message of change and economic prosperity critically influenced vote proportions among this electorate, a result that brings into question the efficacy of the Bush and Clinton models of campaigning. The Bush model “focuses on mobilizing partisans” while the Clinton model “focuses on persuading independent and weakly partisan voters.” Many post-election analyses suggest that the Democrats “either assumed they didn’t need white working-class voters or took their support for granted,” demonstrating a Clinton model approach, appropriately, in persuading more moderate voters and weak partisans, while largely ignoring the base. The result of this, due to the particular identity and geographic location of the Obama-to-Trump voters, helped enable Trump’s Electoral College win, though the intricacies of this topic are too great to fully investigate here. While Clinton ignored the rural white working class voters in Iowa, Wisconsin, Pennsylvania, Ohio, New Hampshire, Trump’s message that he “would fight for the working class over special interests, and his opponent is bought by Wall Street and would advance the forces of globalization” resonated with voters there who agreed with the Democrats on economics but were discouraged by the performance of the current Democratic office holder. Though traditional opinions on the effectiveness of campaigns may not have held exactly true, concepts of retrospective voting and geographic campaigning methods prove effective in explaining what would otherwise seem to be an anomaly in voter behavior.

Many agree that the 2016 election seemed fraught with twists and turns that seemed to defy the conventions of presidential elections of the past. Yet the outcome and voter choice in electing Trump can be explained and expected by political scientists' accepted understandings. For the most part, voters behaved predictably based on social and party identity, and even among white working class voters who seemed to diverge from this, their votes are explained by Trump's campaign of change and the cyclical tendency of the White House to change parties after two terms. Certainly, political and psychological predictions can only go so far in anticipating the results and effects of campaigns and elections, but despite popular uproar over the unconventionality of this election, it doesn't present anything new that should threaten the prior knowledge currently held to be true.


poli sciSarah Burk