A reality-TV-star billionaire businessman tears up the rules and vaults to the brink of the Republican nomination for president, begging the political question of the year: What is support for Donald Trump all about?
The answer, according to statistical analysis of an array of possible explanations tested in the latest ABC News/Washington Post poll, reflects two related but ultimately distinct sets of attitudes: Anti-establishment populism and pushback against outside groups. One is rooted in economic discontent; the other, in a desire among some Americans for traditional authority.
None of these alone explains Trump’s support in the GOP race, nor – especially – do core demographics such as political ideology, gender, age, education or religious belief. While he does better with less-educated adults, and lately, worse among women, evangelical white Protestants and strong conservatives, their explanatory power in Trump’s support is relatively muted, particularly in comparison with the group-based contours of the Democratic race.
Four main explanations –- some overlapping –- have emerged to fill this void. This analysis, produced for ABC News by Langer Research Associates, suggests that each theory has at least some merit, which may help explain the consistency of Trump’s support in the fractured Republican field.
Four Theories of Trumpismo
First, some have suggested that Trump’s support stems from economic discontent, particularly among working-class whites. Indeed, among Republicans and Republican-leaning independents who are registered to vote, 45 percent of Trump supporters say they’re struggling economically, compared with 29 percent of those who prefer either Marco Rubio or John Kasich, with Cruz backers in between. (Rubio and Kasich supporters are grouped together for an adequate sample size, a step supported by the underlying data.)
That said, Cruz supporters are just 7 points less likely than Trump supporters to say they’re struggling, and even among Trump’s backers, more than half say they’re comfortable or “moving up” economically. Clearly there are other factors at play.
It’s also been argued that people who are predisposed to value order, obedience and respect for traditional authority tend to be strongly attracted to Trump, and that this explains his support. The poll measured such orientations with a standard set of questions listing preferred qualities in children, known as an “authoritarianism scale” in academic circles.
Valuing traditional authority does distinguish the candidates, but not Trump from the rest. In fact, 57 percent of Cruz supporters score highest on the traditional authority scale, vs. 43 percent of Trump’s (and 34 percent of Rubio and Kasich supporters). This reflects Cruz’s broad support among evangelical white Protestants, who place a high priority on authority.
These patterns hold in a statistical analysis predicting preference for Trump over the field while controlling for basic demographics. Economically struggling registered leaned Republicans are more likely to prefer Trump than his rivals, but only slightly in the case of Trump vs. Cruz. Inclinations toward traditional authority significantly predict Trump support only in a one-on-one matchup with Rubio. At best, then, these two explanations are just part of the story.
A third explanation –- populism -– suggests that it is not so much preference for established authority that matters, but rather the rejection of traditional politics. Indeed, the idea that Trump’s popularity is fundamentally based on anger against the existing political establishment, and the sense that an outsider is needed to fix it, have significantly more legs.
Eighty-two percent of Trump’s supporters say they want an outsider vs. someone with political experience -– far more than among those among who back the other candidates. Trump’s devotees also are twice as likely as Rubio/Kasich supporters to be angry at the government –- but in this case, no different than Cruz backers. And three-quarters of Trump’s supporters strongly view the political system as dysfunctional, 11 points more than among Cruz voters and far more than among those who support Rubio and Kasich.
The fourth possible explanation also proves quite powerful: pushback against other groups, including Muslims, undocumented immigrants and minorities seen as benefiting unfairly from affirmative action – a set of closely related views reflecting what academics describe as ethnocentrism.
Six in 10 Trump supporters strongly favor deporting undocumented immigrants and banning non-citizen Muslims from entering the country, vs. just a quarter of Rubio/Kasich backers, with Cruz supporters in between. And 37 percent of Trump’s supporters strongly believe that whites are losing out because of preferences for blacks and Hispanics; many fewer of the other candidates’ supporters say the same.
These sets of questions capturing populism and pushback against other groups correlate strongly; as such we combined them as two indices in our statistical analysis of Trump’s support. The analysis also incorporates economic anxiety, valuing traditional authority and many demographic controls. By far, populism and pushback emerge as the strongest independent factors related to support for the frontrunner, all else equal.
While economic discontent and preference for authority lose force as explanations when taking populism and pushback against others into account, that doesn’t mean they don’t matter. Quite the contrary: In a separate analysis of what drives affinity toward populism among the GOP electorate, the most important factor (and one of the few statistically significant ones) is the sense that one is struggling economically. And in another analysis, one of the strongest predictors of pushback against other groups is valuing traditional authority.
These same patterns reappear in a hypothetical general election matchup between Trump and Hillary Clinton. Even after controlling for political party identification, ideology and demographics, populism and pushback are strongly related to backing for Trump, and economic anxiety and valuing traditional authority matter only to the extent that they shape the two former variables. (Gender, income, partisanship and ideology also do matter. Race is not a factor in GOP preferences because leaned Republicans are very predominantly white; in the general election, it’s subsumed by views on outside groups.)
Populism and pushback against other groups relate indirectly as well as directly to support for Trump vs. Clinton. Far and away the strongest variable in general election models is an evaluation of whether things have generally gotten better or worse since Barack Obama took office. And the two variables that most strongly relate to evaluations of the Obama presidency are – again – populist anger and pushback. (Evaluations of Obama do not distinguish GOP primary preferences, since such assessments are almost uniformly negative among registered leaned Republicans.)
These results identify a substantial group of Americans who, seven years into the political order of the Obama era, feel dislocated economically and socially alike. Struggling in the post-recession economy and drawn to traditional authority, they’re attracted to a populist outsider who offers strong anti-establishment credentials, protection against competing groups and a potent stance against the current social and political climate.
This ABC News/Washington Post poll was conducted by landline and cellular telephone March 3-6, 2016, in English and Spanish, among a random national sample of 1,000 adults, including 864 registered voters. Results have a margin of sampling error of 3.5 points, including the design effect, for the full sample; 5.5 points for registered leaned Republicans; and 4.0 points for registered voters. Partisan divisions are 34-25-32 percent, Democrats-Republicans-independents.
The statistical models predicting support for Trump in the GOP primary and general election were estimated in the Zelig package in R using survey-weighted logistic regression; models predicting populism, ethnocentrism and Obama evaluations employed survey-weighted normal regressions. In addition to the variables described above, additional controls included gender, age, education, income, region, identification as an evangelical white Protestant, self-identified social class, ideology and partisanship. Models of the GOP primary included only Republicans and Republican leaning independents who reported being registered to vote; the general election models included all registered voters expressing a preference for Clinton or Trump. The flow charts depicting the models necessarily are simplified. Details of the analysis are available upon request.
The survey was produced for ABC News by Langer Research Associates of New York, N.Y., with sampling, data collection and tabulation by Abt-SRBI of New York, N.Y. See details on the survey’s methodology here.