11/12/2026
South Carolina's election: What moved the vote?
Earlier this week, I mentioned reading Madeleine Carlisle's analysis of Annie Andrews' Senate victory in The Atlantic.
I found myself agreeing with quite a bit of it. And that made me want to examine my own assumptions more carefully.
It's easy to encounter an explanation that aligns with what we already believe and treat that agreement as confirmation. I've been trying to do something different with this election analysis.
So I've been looking at what voter surveys, election returns, and other available evidence can tell us.
First, some voters departed from their previous voting patterns.
According to ABC News' analysis of the 2026 Voter Poll, approximately one in ten South Carolina voters who reported supporting Donald Trump in 2024 voted for Annie Andrews in 2026. Almost none of those who reported supporting Kamala Harris voted for Republican Darline Graham.
There were substantial differences from the 2022 Senate election, too.
Republican Tim Scott carried white voters by approximately 54 percentage points in 2022. Graham carried them by only about 14 points in 2026.
Women supported Scott by an 18-point margin in 2022 but supported Andrews by 13 points in 2026.
Those changes deserve attention. But so does the continued importance of Black voters, who supported Andrews by approximately 85% to 12%.
Interestingly, ABC estimates that Black voters made up 18% of the electorate in 2026, compared with 22% in 2022. Their strong support remained important even as their estimated share of the electorate declined.
We shouldn't explain this election entirely through voters who changed their preferences while overlooking those who provided a substantial base of support.
These are survey estimates involving different candidates and electorates. And someone who voted for Trump in 2024 isn't necessarily a Republican.
One methodological note: ABC and the Associated Press are reporting findings from the same underlying 2026 Voter Poll conducted by SSRS. They aren't two independent surveys confirming one another.
Second, candidate perceptions appear to have mattered.
The Associated Press reported that approximately two-thirds of South Carolina Senate voters said Graham lacked the right experience to serve.
That's meaningful evidence about voter perceptions, particularly given the unusual circumstances surrounding her appointment and nomination.
But how much of Andrews' advantage reflected concerns about Graham specifically, rather than broader dissatisfaction with Republicans?
After all, Republican Alan Wilson won the governor's race on the same ballot.
That comparison makes candidate-specific circumstances worth examining. It doesn't isolate their effects, though. The governor's race involved an open seat and a Democratic nominee, Jermaine Johnson, who had substantially fewer campaign resources and less apparent visibility.
Several readers have raised questions about fundraising, advertising, endorsements, and candidate recognition. One reader who volunteered phone banking for Andrews described encountering voters who didn't even know Johnson was running.
That's an individual observation, not a representative survey. But it raises questions about the opportunities voters had to learn about the candidates.
I've started investigating campaign resources and party funding decisions, including some surprisingly competitive congressional and legislative races. There's enough there for a separate post, so I'll return to that issue later.
Third, economic concerns were prominent before Election Day.
An East Carolina University poll conducted October 19–24 found that 54% of likely South Carolina voters identified the economy or cost of living as their most important issue.
Among those voters, Andrews led Graham by 56% to 41%.
But among respondents who identified immigration and border security as their top issue, Graham led by 95% to 5%.
And despite Andrews' advantage among voters prioritizing the economy, the ECU poll still showed Graham leading overall, 51% to 47%.
Different groups of voters emphasized different issues, and their candidate preferences reflected those differences.
We also have to distinguish a relationship between issue priorities and candidate support from evidence that a particular issue caused someone to vote a certain way.
Fourth, the organizing explanation remains particularly interesting to me.
For months, I've argued that the Democratic Party's decision to recruit candidates for all 124 State House districts could matter beyond those individual races.
Contested elections create opportunities for local organizing, voter engagement, and conversations that might not happen otherwise.
The Atlantic discusses that explanation, too.
And now that the recounts are complete, Democrats have flipped ten State House seats, ending the Republican supermajority.
One of those victories was confirmed Wednesday. Democrat Joe Madge defeated Republican Micah Caskey in House District 89 by just 106 votes. Caskey, who had been chosen as the next House Republican majority leader, won his 2024 election by about 20 percentage points.
Seven of the ten Democrats who flipped Republican-held House seats are Black. Several regained districts Democrats had lost in recent election cycles.
Those victories reflect the work of individual candidates, volunteers, and supporters. And the broader recruitment effort included many candidates who invested their own time and resources in races they ultimately lost.
But here's where I need to challenge my own interpretation.
South Carolina Democratic Party officials report that their House candidates performed approximately ten percentage points better, on average, than previous Democratic performance across districts. I'd like to examine the underlying comparison more closely.
If that pattern holds up under independent examination, broader statewide political conditions may have contributed substantially to the results.
Organizing could have helped mobilize voters within that environment. It could also have benefited from political conditions already moving in Democrats' favor.
To distinguish those possibilities, we'd need to compare organizing activity, campaign investment, district characteristics, and changes in turnout.
We know extensive organizing and electoral gains occurred together. We don't yet know how much one contributed to the other.
One more point about the pre-election polling.
Andrews' victory was historically unusual. South Carolina hadn't elected a Democrat to the U.S. Senate since 1998.
But by late October, the race was already competitive.
A New York Times/Siena poll conducted October 19–26 showed Andrews and Graham tied at 48%.
Other polls differed, but there was evidence before Election Day that the outcome was uncertain.
An election can be historically significant without having been entirely unforeseeable.
So what do we know, and what are we still trying to understand?
We have evidence of changing voting patterns, concerns about candidate experience, prominent economic issues, and substantial Democratic organizing.
What we can't yet establish is the relative contribution of those factors to the election results.
I've appreciated how many readers have offered additional explanations, challenged assumptions, and suggested research questions throughout this series.
Those conversations are expanding my research agenda. I'm tracking questions about campaign funding, migration, endorsements, negative advertising, party divisions, and even whether differences in campaign tone affected voters.
Some will become separate posts. Others may have to wait until more evidence is available.
What interests me now is how these different forces interacted—and which of them will continue to matter in future elections.
The election results tell us who won. Understanding why requires more evidence, more questions, and a willingness to reconsider our assumptions.
Sources, including my Atlantic gift link, voter-survey methodology, and additional reporting, are available in the comments.
Research and AI disclosure: This analysis draws on public-opinion surveys, election results, campaign-finance reporting, and political commentary. AI assisted with research, source comparisons, organization, and drafting. Interpretations and conclusions are my own.
Graphic credit: I designed and created the accompanying graphic in Adobe Express on November 8, 2026.
A note about comments: Thoughtful disagreement and differing interpretations are welcome. Personal attacks, inflammatory comments, and unsupported allegations presented as fact are not. Let's keep the discussion respectful and focused on the evidence.