Soccer moms, security moms, Zoom moms.
These catchphrases are a metaphor for one voter group often mentioned as the key to deciding elections: suburban women.
Anointing an all-important swing voter group is a time-honored tradition of election coverage. More often than not, the picture painted of these voters is white, suburban, usually college educated, Independent, “voting for the person, not the party.” Trump’s first election challenged the traditional view of swing voters. Those who swung heavily in his direction were overwhelmingly white voters without a college degree, hailing from rural or small town areas, not upscale suburbs.
Still, the suburban bias in election coverage remains strong. It persists in everything from the precincts the media covers on Election Day to where voter focus groups are usually conducted — populated areas that our urban media can easily parachute into. The problem is that when everything in politics revolves around population density, going to relatively dense areas paints a misleading picture if there’s even bigger movement happening in more spread-out places that are harder to chronicle.
The focus on the suburbs does make sense thinking of them as a midpoint between urban areas which have always voted more Democratic and rural areas that have historically voted Republican. That’s where the usefulness stops. Collectively, the suburbs may better reflect the country’s overall political balance. But that doesn’t mean that individual voters in the suburbs are more prone to being swing voters and that you should focus on them at the expense of urban or rural groups. In fact, swing voters are all over place, even in the most lopsidedly Democratic and Republican groups.
According to the 2016 exit polls, white voters without a college degree moved to backing Donald Trump by 37 points after backing Mitt Romney by 25 points, a net margin swing of 12 points. In the other direction, white voters with a college degree moved from supporting Romney by 14 points to Trump by 3 points, a net swing of 11 points. Both groups were pretty swing-y in 2016.
White college voters are close to the political median, and white non-college voters are to the right of it. Yet they moved in different directions in roughly equal amounts. One’s initial leanings to the right or left does not make one less subject to being persuaded to vote differently than you did in the last election.
African Americans, a group considered a lock for Democrats, moved from supporting the Democratic ticket by a net of 77 points in 2004 to supporting Barack Obama by 91 points, a net movement of 14 points, higher than the shift seen among each segment of white voters in 2016. Their turnout also increased substantially, but that doesn’t explain the Republican vote collapsing from 11% to 4% in just four years. Because the margins were so large to begin with, any additional movement doesn’t seem to matter, but in fact it was responsible for a more than 1 point shift in the national popular vote. Removing Obama from the equation, and thus unwinding this extra advantage with black voters, possibly cost the Democrats the election in 2016.
All of these pale in comparison to the recent historic shifts among Hispanic voters, another Democratic leaning group. The net Democratic advantage shrunk from 51 points in 1996 to 27 points in 2000 to just 9 points in 2004, a swing of 42 points in just four years. While Democrats have vigorously disputed these 2004 numbers, there’s no questioning the fact that having a more Latino-friendly face to the Republican Party in George W. Bush caused shifts in voter support more substantial than those that have since been seen among white or black voters, or any “swing” voter group.
What’s happening here is not turnout. It’s not base mobilization. It’s persuasion. Most of us are subject to it to one degree, even if it doesn’t cause a complete shift in political alignment. It might show up in a Trump voter not being more or less enthused than they were in 2016, or a liberal casting a protest vote in 2016 because Clinton had it in the bag and she didn’t seem that inspiring anyway.
Also: High-income, educated groups — those happen to cluster most in suburbs — are somewhat less likely to shift their preferences from election to election. In Presidential elections from 1980 to 2016, white college voters’ average movement from the previous election was 7 points, compared to 10 points for white non-college voters. Hispanic voters moved an average of 13 points in each election, and African Americans moved 6 points.
Trends among whites are consistent with evidence from polling that lower educational attainment correlates with lower voter turnout and a relative lack of political alignment. These voters may lean in one direction for cultural or historic reasons, but they are less loyal and harder to pin down than someone in the suburbs with a graduate degree who is highly politically engaged.
This can be seen in individual-level model scores, which are essentially polling results expressed probabilistically for each voter. On the y-axis is below is the likelihood of a person turning out in the 2020 election, and on the x-axis is the probability that they approve of Trump’s job performance. The people in the middle, who are neither certain to support or oppose Trump, also have lower likelihood of turning out. Lower turnout correlates with age and party registration, with young people and political independents likely to turn out less. But it also correlates with lower levels of educational attainment. There are simply fewer entrenched voters at the lower end of the educational spectrum, and by definition, more swing voters.
Our conception of swing voters as a finite group is flawed. As voters, we share similar DNA and a compelling-enough set of facts is likely to move any group of voters — not just suburban voters or political independents — off their existing baseline. No matter where on the political spectrum this movement happens, it matters just as much. Some groups may see bigger swings from election to election, but these groups are not just suburban voters or those with voting preferences close to the national average.
While 2020 may be very different from 2016, the election is again likely to be about white non-college voters. Not only are they more likely to move based on political information and advertising, but they hold strategic importance in that they overrepresented in the Midwestern swing states. In politics, it’s easier to recapture something you recently had than to take something that’s been long lost. And both sides have good reason to believe they can recreate performances with white non-college voters that were key to decisive Electoral College wins: for Republicans in 2016 and for Democrats in 2012.