Chapter One

United States of Trump

An inside look at the voters who took over the Republican Party

Who were the Donald Trump voters who took America by surprise on Tuesday and what did they want from the election? Following the conclusion of the Republican primaries, NBC News talked to Trump supporters all over the country and crunched the numbers for a comprehensive look at his voters.

Many of the same dynamics that led Trump to victory in the primaries also applied to the general election. Over the course of the GOP race, Trump started with an enthusiastic base of blue-collar voters in downscale rural counties and only gradually expanded his appeal to bring in more affluent conservatives in suburbs and cities. On Tuesday he pulled off the same trick nationally, winning thanks to a combined turnout of fired-up rural voters along with party stalwarts in the suburbs and exurbs who had been slower to accept him as their nominee. NBC News is presenting this reporting once again following Election Day. It was originally published over the summer.

Joe Roberts and Matthew Douglas, auto mechanics from Indianapolis, took off early from their job to line up for a Trump event in nearby Carmel. They came in their work clothes: a black t-shirt and an Indianapolis Colts hat for Roberts, 39; a powder-blue shirt with the sleeves cut off for Douglas, 36; and dirty, ripped jeans for both. Douglas has a tattoo of a flaming panther on his right arm and a Chevy logo on his left. Roberts’ hands are rough, weathered by his auto work and two decades of landscaping jobs.

Roberts laid out his reasons for backing Trump: As a billionaire, Roberts said, Trump can’t be bought. He likes Trump’s plan to slap tariffs on goods from American companies that outsource. He thinks abortion should be legal, and doesn’t like the way candidates obsess over the topic instead of the economy and national security. He thinks people in his neighborhood are getting government benefits while drawing income under the table.

“I dated a girl on disability who said she couldn’t sit for eight hours straight,” he told NBC News. “She could sit on a barstool eight hours.”

Both Roberts and Douglas said they had registered to vote for the first time in order to support Trump in their state’s primary.

We’ve been waiting for someone worth a damn,” Douglas said.

In many ways, the pair embodied a type of Trump voter political observers have learned to recognize: Blue collar, disconnected from politics, indifferent to the conservative doctrine that has defined the Republican Party for the last several decades.

Trump fans recognize the type, as well. A picture of the two auto workers this reporter tweeted out instantly went viral among the candidate’s vast network of supporters online. “Why we are winning,” a fan commented under a post of the picture on Reddit, a popular online message board with a large pro-Trump following.

The next day, Trump won 53 percent of the vote in Indiana’s primary, including 46 percent of Hamilton County, the site of his Carmel rally. With Trump’s path to the nomination all but assured, rival Ted Cruz dropped out that night and John Kasich left the race shortly afterward.

Republicans since then have struggled to wrap their heads around their new reality. Over the course of 10 months, Trump became the GOP’s standard-bearer despite offending large swaths of voters, antagonizing movement conservatives on a variety of issues and defeating well-funded and experienced Republican rivals from every wing of the party. Only two of the party’s five living presidential nominees – Bob Dole and John McCain – are publicly supporting Trump in the general election. And even as he routed his opponents in Indiana, a full 42 percent of Republican voters in the state said in an exit poll they were “concerned” or “scared” about a Trump presidency.

Who voted for this guy?

The Trump Takeover

To answer that question, NBC News interviewed Trump voters at rallies across the country, consulted polling data and academic studies and analyzed how counties voted in every state that held a competitive contest.

What we found was a distinct movement of Americans alarmed by economic trends, unsure of their place in a more diverse nation and convinced that the major parties no longer have their interests in mind.

In telling their story, we tried not only to describe Trump’s support, but to trace its evolution from a relatively small plurality of Republican voters to a powerful majority that overcame a nationally organized effort to stop their candidate.

Trump struggled in the middle of the country and in Utah. Big wins closer to the coasts helped drive his success.

Bottom 20th percentile of Trump support

20 to 40th percentile

40 to 60th percentile

60 to 80th percentile

Top 20th percentile of Trump support

Less

More Trump vote

Top 20th percentile of Trump support

Less

More Trump vote

60 to 80th support percentile

40 to 60th percentile

20 to 40th percentile

Bottom 20th percentile of Trump support

Dave Leip’s Atlas of U.S. Presidential Elections

Today, a description of who supports Trump is simple enough: Republicans.

In the final seven races before party leaders crowned Trump the presumptive nominee, he took every county save for six, while winning almost every demographic slice of GOP voters. His county victories included places like the aforementioned Hamilton County in Indiana, which many local politicos had assumed was too upscale and traditionally conservative for Trump to win based on his performance in similar counties earlier in the race.

There are guys like Joe Roberts and Matthew Douglas at every Trump rally, but there are also voters like B.J. Cobb, 64, for whom it took time to come around to supporting Trump.

“I thought he was crazy when he started talking about walls and border security,” she said at a Trump rally in Charleston, South Carolina in February.

One of the most dramatic predictors of Trump’s success early in the race was how much a county’s income had grown – or not – from 2004 to 2014.

Cobb is a lifelong Republican who outwardly fits the definition of establishment – she fondly recalls working for the Republican National Committee in the 1970s and name-drops top Republican strategists she has befriended over the years. But when she saw the energy at Trump’s events, Cobb said she became convinced he had tapped into something the party needed.

“He had the foresight to see what was coming,” she said. “He speaks to the people. He’s able to stand up for the country. I haven’t been as excited about a candidate since Reagan.”

If national polls are to be believed, there are a lot of voters like Cobb who warmed to Trump gradually. In a March 2015 NBC News/Wall Street Journal survey, 74 percent of Republican voters said they could not imagine supporting Trump. The number was only marginally better, at 66 percent, when he launched his campaign that June. By April of 2016, however, the number had plummeted to 38 percent.

By the time the primary race ended, Trump also drew close to 50 percent support from GOP primary voters in national polls. It wasn’t always that way: For almost all of 2015, he polled in the mid-20s, leading many Republicans to speculate that there was a ceiling to his support.

Trump’s Support in the Polls Grew Steadily Over Time Other candidates jostled for position but couldn’t compete

60%

45

TRUMP

30

15

0

June 2015

November 2015

May 2016

60%

45

TRUMP

30

15

0

November 2015

June 2015

May 2016

Real Clear Politics Republican Presidential Nomination Average

Trump’s support was a modest plurality within a large Republican field when voting began in February, as well. His first big win, in New Hampshire, came with just 35 percent of the vote. By April, as his popularity grew and the field of candidates shrank, Trump won states like Pennsylvania with 56.7 percent of the vote and Delaware with 60.8 percent.

The story of Trump, then, is about a committed minority of Republican voters, whose existence was routinely ignored, gradually taking over the entire party. As you’ll see in our data, the people who backed Trump early on – and the places they live – were markedly different than the unified GOP bloc he commands now.

The World of the Trump Voter

One way to describe Trump’s support is by geography. Polls of voters may send mixed messages depending on how questions are worded and the sample size, but we know exactly which counties voted for Trump and by what margins.

In examining Trump’s county-level support, we broke the election down into two time periods.

The first period stretches from the Iowa caucus on February 1 to the five primaries held on March 15, which were headlined by Ohio and Florida and ended with Sen. Marco Rubio dropping out of the race. The second period starts with the Arizona primary and Utah caucus on March 22 and ends with the May 3 Indiana primary, which effectively finished the Republican primary contest.

It’s important to make these distinctions, because the markers that best distinguished Trump’s initial voters shifted as Trump’s support expanded.

Trump seemed to grasp intuitively that his base lived in a different world than that of other candidates. He boasted in speeches that he did well with “poorly educated” voters and cited economic frustrations over trade and jobs, especially among blue collar workers, as essential to his appeal.

“When people say, ‘It’s unbelievable what’s happening to Trump,’ it’s really not,” he told a crowd in Pennsylvania in April. “You have people who did better 18 years ago, you have people who have two jobs and they did better 17 years ago and they’re working harder now.”

The data suggest Trump was onto something. According to our county analysis, one of the most dramatic predictors of his success early in the race was how much a county’s average pay had grown (or hadn’t grown) from 2004 to 2014. Another major predictor was the percentage of whites who participate in the labor force. Still another was whether residents were more or less likely to hold at least a high school degree.

The below chart plots Trump’s county-level support in the first period – up until the March 15 contests – against a composite metric of local distress, particularly focused on whites, who made up close to 90 percent of self-identified Republicans in the previous presidential election cycle. The larger the circle, the larger a county’s population.

In counties with higher white distress, Trump tended to receive more of the vote

Distress metric is weighted among the five-year 2014 American Community Survey estimates for white poverty (non-Hispanic), white labor participation (non-Hispanic) and high school degree holders and the inflation-adjusted difference in average annual income from the BLS 2004 and 2014 Quarterly Census of Employment and Wages surveys.

See that medium sized circle on the far left of the chart? That’s Midland County, Texas, the site of an oil boom that sent incomes through the roof in the 2000s and kept unemployment well below the state average during the Great Recession. Cruz won 47.9 percent of the vote there versus Trump’s 20.9 percent, and did well in the small neighboring Texas counties clustered around Midland on the chart. Look slightly to the right and you’ll spot Alexandria, Virginia, the affluent D.C. suburb where unemployment peaked at just 5.6 percent in 2010 and where Rubio performed especially well.

Move to the upper right of the chart and you’ll notice a number of counties in Appalachia where jobs are relatively scarce, poverty is high and Trump enjoyed strong support. This includes coal-heavy Knox County in Kentucky (Trump vote: 55 percent) and Buchanan County in Virginia (Trump vote: 70 percent), where he secured his highest winning margin in the period covered.

As time moved on, though, Trump started picking up votes in more diverse places, including the kind of well-heeled counties that shunned him earlier on in the race. It was this late shift that allowed Trump to knock out his rivals in the final stretch.

In the simulations below, toggle between the two primary periods to see how Trump’s appeal grew over time in counties with different distinguishing traits. To start, we’ll look at income growth.

Switch to
Period 1
Period 2
Percent of counties won by Trump in Period 1 with..
Data computed from inflation-adjusted difference between the BLS 2004 and 2014 Quarterly Census of Employment and Wages survey.

In the February 1 to March 15 time frame that makes up Period 1 above, Trump performed significantly better in counties that had below-average growth in average annual pay from 2004 to 2014, but won just 46 percent of counties that had experienced the most improvement. Of the 100 counties with the highest gains, he won just 23.

Click over to Period 2, though, which includes contests from March 22 to May 3, and Trump won 77.9 percent of these high-growth counties, including 77 of the 100 where growth was strongest.

The same dynamic played out with education, as you can see below.

Switch to
Period 1
Period 2
Percent of counties won by Trump in Period 1 with residents..
Data computed from five-year 2014 American Community Survey estimate for high school degree holders.

In Period 1, 88.2 percent of counties where residents were least likely to hold a high school diploma went for Trump, while 42.5 percent of the counties where high school graduates were the most common went for one of his opponents. Trump won a dominant 91 of the 100 counties with the lowest percentage of high school graduates during this stretch. The relationship was similar for college as well: Trump took just 36 of the 100 counties where residents were most likely to hold at least a bachelor’s degree.

Click on Period 2, though, and you can see how Trump maxed out his support among low-education counties to 100 percent while expanding his appeal to the most educated counties, 74.2 percent of which voted for him.

Finally, the last simulation below looks at white labor participation. In Period 1, Trump only won a majority of counties in the two slices with below-average labor participation rates. Click over to Period 2, though, and it’s a clean sweep, with moderately better labor participation counties 41 points more likely to support him than they were in Period 1.

Switch to
Period 1
Period 2
Percent of counties won by Trump in Period 1 where there was..
Data computed from five-year 2014 American Community Survey estimate for white labor participation (non-Hispanic).

The county level data lend some context to one of the oddities of the GOP race: Why voters rallied around a burn-it-all-down populist message while unemployment dropped to 5 percent nationally and Americans increasingly described themselves as happy with their own economic situation. In many parts of the country, the income trajectory wasn’t great even before the crash, the damage of the Great Recession was more extensive and the benefits of recovery accrued unevenly. Many of the gains nationally went to the top 1 percent of earners, whose investment income bounced back more quickly than jobs and wages for ordinary workers.

Health statistics also speak to pressure on the white working class voters who have been most drawn to Trump’s candidacy from the start. An analysis by two Princeton economists last November found a frightening deterioration among middle-aged white Americans, especially those without a college education. Death rates rose dramatically from 1999 to 2013, even as they fell for other ethnic groups, and the leading drivers of the shift were suicide and substance abuse. A follow-up analysis by the Washington Post found the death rate had risen particularly sharply for middle-aged white women in rural communities.

The degree to which Trump supporters are struggling with these issues themselves isn’t entirely clear. Solid data on individual voter income, wealth and health is hard to come by.

What our county data does show, however, is that Trump’s most enthusiastic primary voters were more likely to live in places where they were exposed to rougher economic conditions. For them, these issues were not something they read about in a newspaper.

What Trump’s Supporters Want

If you go to a Trump rally and ask people why they support him, what’s the most common answer? You might be surprised.

It isn’t the border wall or his plan to ban Muslims from entering the country or his position on trade, although those come up regularly. It isn’t that he defies “political correctness” or “says what we’re all thinking, but afraid to say,” two phrases that come up often, as well.

In interviews with NBC News, Trump supporters discussed why they support Donald Trump.

Instead, it’s an issue that’s been almost entirely ignored by the Republican Party in recent years: Money in politics.

“He can’t be bought,” Eleanor Crume, 72, said at a South Carolina rally. “He’s not going to be bought by the lobbyists.”

“He can speak his mind because he’s not backed by these donors who say what he can and can’t say,” Travis Klinefelter, a 39-year old Iowa nurse, said.

“He’s not bought and paid for by special interests,” Dominic La Rocca of Ft. Lauderdale, Florida said. “Insurance companies, the banks, they get the law that they want.”

Unease with the corrosive influence of campaign donations and super PACs probably wasn’t something GOP leaders expected to influence their primary. Until now, the loosening of campaign finance restrictions after Citizens United and a subsequent flood of hundreds of millions of dollars into conservative causes was widely assumed to be an unmitigated good for the right.

Trump, however, seized on a number of issues that challenged expectations about how much sway the right’s governing agenda held over its base. It turns out that Republican voters were much less concerned with conservative orthodoxy (as defined by the party’s big donors, activists and elected officials) than party elites had thought.

Surveys taken just before Trump joined the race in June 2015 showed Trump was intensely disliked by Republican voters. But he rocketed to first place in the crowded GOP field immediately after his announcement speech, where he attacked illegal immigration and accused the Mexican government – against all evidence – of deliberately pushing “crime” and “rapists” into America.

Early polls found few distinguishing traits to Trump’s supporters, leading some skeptics to dismiss his strength as the product of low-information voters responding to whatever was on TV.

“If you would give some other candidates time from eight in the morning ‘til eight at night – all day long, every day for three weeks – I’m guessing some other candidates might rise as well,” Sen. Rand Paul, then a candidate, complained on CNN in July.

As time went on, though, it became clear that while Trump’s support included many corners of the party, his most committed backers had an ideological profile all their own.

In national polls leading up to the first contests and in exit polls after they began, Trump regularly performed best with voters who considered themselves “moderate,” a descriptor that had become almost a slur in the tea party era. He did worst with voters who considered themselves “very conservative.” They leaned more toward Cruz and his campaign’s focus on ideological purity.

Trump dominated among moderates but struggled to win over very conservative voters.

Trump’s success..

.. among moderate voters

.. among very conservative voters

IN

NY

CT

MS

MA

PA

NV

MD

TN

NH

MO

AL

MA

PA

VT

FL

VA

MD

GA

NV

AL

OH

NC

FL

GA

WI

MI

SC

TN

IL

IN

AK

IA

SC

MS

OK

NH

IL

NY

OK

MI

AK

MO

CT

VT

NC

VA

IA

OH

WI

-40

-20

0

20

40%

-40

-20

0

20

40%

Margin behind leader

Margin over next best candidate

Margin behind leader

Margin over next best candidate

Trump’s success..

.. among moderate voters

IN

MS

PA

NV

TN

MO

MA

FL

MD

GA

AL

NC

WI

SC

IL

AK

IA

OK

NH

NY

MI

CT

VT

VA

OH

-40

-20

0

20

40%

Margin behind leader

Margin over next best candidate

.. among very conservative voters

NY

CT

MA

MD

NH

AL

PA

VT

VA

NV

OH

FL

GA

MI

TN

IN

SC

MS

IL

OK

AK

MO

NC

IA

WI

-40

-20

0

20

40%

Margin behind leader

Margin over next best candidate

NBC News exit polls

“Moderate,” in this case, meant a specific kind of profile recognizable to political scientists. Rather than simple middle-of-the-road voters, “moderates” tend to combine hardline, even fringe, views from opposite ends of the political spectrum, leaving them without a clear home in either party or any specific political movement.

Trump’s brand of nativism and populism is a classic example: He is ultra-conservative on deportations and border security, leans left within the GOP on social spending and trade, and offers a mix of isolationist and militarist views on foreign policy. It turns out these positions are popular among many GOP voters, but not among the intellectual leaders who have shaped the party’s platform.

“I like that he gets heat from Republicans,” 46-year old Pennsylvania Trump supporter John Bishop said. “He’s extreme on some things, but on others he’s a moderate. I like that mix.”

One of the most detailed analyses of this effect was the RAND Corporation’s Presidential Election Panel Survey (PEPS), which tracked 3,000 voters at different points in the election. NBC News reviewed their data with assistance from political scientists Michael Tesler, Lynn Vavreck and John Sides, who helped oversee the study.

To assess voters’ outlook on economic policy, respondents were asked questions like whether they support labor unions, higher taxes on the wealthy, a higher minimum wage and government-funded health care. The further left their answers were, the more likely they were to support Trump.

As time went on and Trump’s overall support increased, the connection between a voter’s economic views and their affinity toward Trump faded. Trump’s base included more traditional economic conservatives in RAND’s March poll than it did in its December/January poll.

More of Trump’s early supporters were economically liberal. Later in the primaries, his support coalesced to include more economically conservative voters.

Economic ideology

50%

Intend to Vote for Trump

40

30

20

10

0

More liberal views

Less liberal views

Dec./Jan. Survey

March Survey

RAND Corp’s Presidential Election Panel Survey

RAND wasn’t the only one to pick up on this trend. An October 2015 study by political scientists Doug Ahler and David Broockman found that voters who leaned left on taxing the rich and leaned right on immigration were far more likely to back Trump.

Cutting federal spending to achieve a smaller government has long been an essential part of Republican orthodoxy. In interviews, Trump supporters expressed concern about rising national debt, but they were often more upset by a perception that government spending went to the wrong places, such as undocumented immigrants, foreign governments, foreign wars and undeserving welfare recipients.

“We have friends with disabled children who need help, veterans who need help, but we’re spread too thin trying to help everyone,” said Krista Adams of Madison, Wisconsin, who attended a Trump event with her husband and young children.

Trump’s support also challenged assumptions about Republican views on social issues. Research by the University of Pennsylvania’s Institute for the Study of Citizens and Politics suggests Trump’s voters tended to be more supportive of abortion rights than other Republicans at a time when many Republican politicians have moved further to the right on that issue. An NBC News|SurveyMonkey poll of 1,007 registered voters who support Trump in late February found that 45 percent of respondents identified as “pro-choice.”

Trump nonetheless performed well with evangelical voters, although church attendance (or lack thereof) also proved to be an important predictor of support. In Indiana, for example, exit polls found Cruz won 61 percent of voters who attended church more than once a week, even as Trump won self-described evangelicals. By contrast, Trump won 69 percent of voters who attend church only a few times a year and 66 percent of voters who never go to church.

National polling pointed in the same direction. An analysis of an American National Election Studies survey of Republican voters in January 2016 found that 36 percent of evangelicals who attended church weekly supported Trump, versus 52 percent of evangelicals who seldom or never attended church.

Trump hasn’t relied on traditional TV advertising to spread his message. Instead, he draws huge crowds to his rallies and takes advantage of the media coverage that follows.

That doesn’t mean Trump and his voters are always in sync. He often tries to have it both ways on taxes, for example, boasting about his willingness to make the rich pay more while running on a tax plan that would reduce the top 1 percent’s tax bill by an estimated average of $1.3 million. While he’s downplayed social issues, he still favors appointing pro-life judges, banning abortion with limited exceptions and defunding Planned Parenthood. Whether it’s accurate or not, though, he’s left an impression among his supporters that he’s closer to them on issues.

An important aspect of Trump’s appeal isn’t about individual issues at all, but rather a sense that his unusual background will allow him to bypass partisan gridlock and run things more effectively than conventional politicians.

“Trump will take measures that may seem drastic, but we’re going in a circle,” Jacob Elias, a 22-year old supporter from Racine, Wisconsin, said. “I think he can break out of that.”

Given the heterodox views of Trump’s supporters, it’s not surprising that many feel less beholden to political parties. A New York Times analysis of research from Civis Analytics in December that included more than 11,000 Republican-leaning respondents over several months found Trump did especially well with voters only weakly connected to the party, including registered Democrats who tended to vote Republican in national elections.

At rallies, it’s common to run into Trump voters who say they used to vote for Democrats or had supported other “outsider” candidates like Ron Paul or Ross Perot in the past. Some even say they admire Bernie Sanders for challenging the usual two-party system.

Dan Powers, a 51-year old Trump supporter from Wisconsin who runs a business installing electronic systems, described himself as “more Democratic than Republican,” but complained that the Democrats he supports “never make it to the front of the line.”

Powers did not vote for Obama, whom he considers hostile to gun rights, and he doesn’t like Hillary Clinton either. Republicans usually left him cold, as well. When the tea party movement took off, Powers said he had little interest.

When Trump came to Janesville for a rally, however, Powers donned an oversized paper mask of the candidate’s face to cheer him on.

“We need to lock the borders down,” Powers said.

The Resentment Factor

One interpretation of Trump’s early support was that his voters felt economically abandoned and were lashing out. The less charitable interpretation, espoused by Trump’s critics on the right and left, is that Trump finally gave a voice to their hostility toward immigrants and minorities.

Trump’s own appeals to prejudice are well-documented: During the campaign, he argued a federal judge was biased due to his “Mexican heritage,” spread false rumors about American Muslims celebrating 9/11, proposed an indefinite ban on Muslim travel and retweeted fake statistics from white supremacist circles claiming black criminals preyed on whites.

Whether Trump’s racially inflammatory remarks and positions are the primary driver behind his support is harder to answer. There is strong evidence, however, that Trump’s supporters are characterized by intolerance more than other voters.

The RAND PEPS surveys in December/January and in March that found Trump supporters to be more economically liberal than other Republicans also found that they were more likely to show signs of resentment toward minorities and immigrants based on questions designed to measure these qualities. Trump also did better with voters who showed stronger signs of white ethnocentrism, a trait based on how favorably respondents rated whites relative to other racial and ethnic groups.

The charts below show the percentage of Republican voters in each survey who said they intended to vote for Trump divided into four quartiles, measuring their views on immigration, race and ethnocentrism. It’s a stark gap and one that grew as Trump consolidated support in March even as the divide over economic issues shrank.

A national poll found that GOP voters who feel more resentment toward immigrants more often intended to support Trump.

Dec./Jan. Survey

March Survey

Racial Resentment

60%

50

Intend to Vote for Trump

40

30

20

10

0

More resentful

Less resentful

Immigrant Resentment

60%

50

40

30

20

10

0

More resentful

Less resentful

White Ethnocentrism

60%

50

40

30

20

10

0

Less ethnocentric

More ethnocentric

RAND Corp’s Presidential Election Panel Survey

A March Washington Post/ABC News poll found that Republicans who saw themselves as “struggling economically” were more likely to support Trump. Even more drawn to his candidacy, however, were Republicans who felt whites were “losing out” compared to other racial groups. Economically-minded supporters, the Post found, backed Trump no matter their views on race, while racially-minded supporters backed Trump regardless of their views on the economy.

Pew Research conducted an online survey in April and May that produced similar results. Republican respondents who believed immigrants threaten American values rather than strengthen society, that Islam was more likely to encourage violence than other religions, and that America’s trend toward majority-minority status was bad for the country, were much more likely to have warm feelings for Trump in comparison to other Republican voters. By contrast, questions that tested their economic views – like whether poor people deserved more benefits or whether the country is tilted too far in favor of the economically powerful – produced much smaller splits in Republicans’ feelings toward Trump.

Trump’s hard-line immigration policies loom large over the discussion, but the issue’s centrality to his campaign is nuanced. Exit polls indicate that Trump performed better with Republican voters who want to remove all undocumented immigrants versus those who want to offer them a path to legal status. But he often won a substantial share of the latter group, too, which made up a majority of Republican voters in almost every state contest where the question was included. Fifty-five percent of respondents in February’s NBC News|SurveyMonkey poll of Trump supporters said they favored deportations versus just 29 percent of all voters, but a significant 44 percent of Trump supporters favored legalization.

Exit polls found that Trump did much better among voters who wanted to deport illegal immigrants.

Trump’s success among voters who believe illegal immigrants should..

.. be deported to the country they came from

.. be offered a chance at legal status

NY

NY

PA

PA

MS

AL

IN

FL

TN

NH

NH

OH

FL

MI

AL

IN

MI

IL

GA

GA

IL

OK

NC

SC

AK

VA

SC

TN

MO

MO

VA

NC

OK

MS

WI

AK

OH

WI

-60

-40

-20

0

20

40

60%

-20

0

20

40

60

80%

Margin over next best candidate

Margin behind leader

Margin over next best candidate

Margin behind leader

Trump’s success among voters who

believe illegal immigrants should..

.. be offered a chance at legal status

NY

PA

MS

IN

TN

NH

FL

AL

MI

GA

IL

NC

AK

SC

MO

VA

OK

WI

OH

-60

-40

-20

0

20

40

60%

Margin over next best candidate

Margin behind leader

.. be deported to the country they came from

NY

PA

AL

FL

NH

OH

MI

IN

IL

GA

OK

SC

VA

TN

MO

NC

MS

AK

WI

-20

0

20

40

60

80%

Margin over next best candidate

Margin behind leader

NBC News exit polls

One explanation is that immigration symbolizes different things to different people. In interviews, a number of Trump’s supporters stressed that they believe in or even encourage legal immigration. Others described legal immigrants — not just undocumented workers — as competition for jobs and wages. Some looked at the issue through the lens of national security, imagining scary scenarios of ISIS infiltration.

“I want to control immigration,” said Marge Stafford, a 51-year old nurse in Media, Pennsylvania. “I’m all about legal immigration, which is fine.”

The more hostile talking points came from voters who viewed migrants as a sponge on resources.

“They come here, they have babies, then they have the right to stay illegally,” Sandy Murray, a 50-year-old Trump voter from Dubuque, Iowa, complained. “We need the money from their handouts for education and veterans.”

“Illegal aliens get free health care and we pay for the interpreter,” Lynette Phillips, a 48-year old real estate agent from Elkhorn, Wisconsin, said. “Our voters don’t get free health care.”

Social Security and Medicare, the largest and most popular entitlement programs for senior citizens, are the primary drivers of the long-term budget deficit. Undocumented immigrants are ineligible for both, along with almost all other federal benefits. Still, for many Trump supporters, it feels like the government’s attention is going to every group except the ones they believe need it most.

This belief isn’t inherently linked to race or immigration status, but at times it can resemble an ethnic competition.

“They’ve had their candidate for eight years,” as one tattooed trucker put it outside a Cincinnati-area Trump event. Asked what he meant by “their candidate,” he clarified that he meant black voters.

There are other stray data points that point toward the resentment factor. In Illinois, a state Trump won by a large margin, voters were required to select individual delegates on their ballot with the names of the candidate they supported next to them. Two delegates submitted by the Trump campaign, Nabi Fakroddin and Raja Sadiq, received several thousand fewer votes than the other Trump delegates in their districts for no apparent reason other than their names.

Online, a dedicated fan base – one that Trump sometimes encourages with retweets – engages in more overt expressions of racism and anti-semitism, often bombarding his critics with slurs and offensive images. It’s hard to tell how much Trump’s internet following is representative of the average Trump voter, however.

Donald Trump signs copies of his TIME magazine cover. For the magazine’s photoshoot, Trump posed in his office with an eagle one arm while giving a thumbs-up with the other.

In interviews, supporters frequently bristled at accusations of racism, and Trump himself has made a show of highlighting black, Latino and Asian supporters at his events. A February NBC News|SurveyMonkey poll of Trump voters found that only 21 percent believed “racial prejudice” drove support for their candidate either “somewhat” or “a lot,” while 79 percent said they believed “anti-immigrant sentiment” was a motivating factor.

Still, a tension hangs over Trump events. Edgar Green, an African-American Trump supporter who works as a dishwasher and convenience store clerk in Maryland, offered an illustration of the complicated relationship between Trump voters and race.

While waiting in line for a Trump rally in West Chester, Pennsylvania, Green, who voted for Obama previously, insisted that “Trump is not racist, he’s America first.” He complained that Trump’s critics had distorted the candidate’s message. His plans to ban Muslims and build a wall were rooted in national security, not bigotry, Green said.

“It’s high time we stop having business outsource to other countries,” Green said. “We got a high unemployment in this country and we need to do something about it. … All these people in line feel Donald Trump is the answer.”

Just then, a car drove by waving a large Confederate flag, which prompted a round of whoops and hollers from the long procession of Trump supporters waiting to get into the rally.

Green was among those cheering.

“Hey, rocking the Confederate Flag, all right!” he said.

The Trump Surge

Trump’s support has challenged so many conventions about the Republican Party and politics more broadly that right up until the last minute, observers were unsure whether it was real. The Trump campaign’s weak ground game fueled hopes among his opponents that his supporters wouldn’t bother to show up and vote.

They were wrong. In key states throughout the race, turnout soared. By the time Trump became the presumptive nominee, about 7 million more votes had been cast than in the entire 2012 Republican primary, even before big states like California and New Jersey had a chance to weigh in. In Florida, a state that had been a critical battleground for Mitt Romney in the 2012 GOP race, close to 700,000 more voters swamped the polls and helped power Trump past the state’s native son, Marco Rubio. Research by Politico suggested the national surge of voters came primarily from general election voters who normally skipped primaries rather than habitual non-voters, but their presence nonetheless reshaped the primaries.

To Trump’s opponents, these voters were falling for an act. Rubio famously labeled Trump a “con artist.” Other rivals complained that he was ignorant of foreign policy, that his trade wars would cause a recession, and that he frequently made false statements.

Voters heard plenty of criticism of Trump, mostly late in the race. But they had their pick among a wide range of Republican hopefuls from every corner of the party. Whatever alternative message these candidates had to offer didn’t resonate.

“I’m tired of people promising the moon and not giving us anything,” Jeff Marshaus, 53, of Shreve, Ohio, said. “I’m a lifelong Republican and I got tired of the same old, same old.”

Republican voters had their pick of more than a dozen candidates, but no one resonated more clearly than Trump.

Credits

Words Benjy Sarlin

Data editor Sam Petulla

Photographs Mark Peterson/Redux

Additional reporting Dante Chinni

Senior editor Beth Fouhy

Copy editor Liz Johnstone

Photo editor Rachelle Klapheke

Director of photography Amy Pereira

Designer & Developer Ian Rose

Project Editor David Taintor

Video Robert Adashev, Annie Cruickshank, Michael Sperling

More from NBC News Specials

The Year in Pictures

From an unexpected hurricane and deadly wildfires to migrants flocking to the U.S. border, 2018 was full of big stories with powerful images to accompany them.

Apollo 11 50th Anniversary: The history of the first moon landing

It's been 50 years since the first humans walked on the moon. What comes next?

“It never crossed my mind that I would not be able to get pregnant.”

Your stories about egg freezing, ages 19 to 60