Donald Trump once quipped that his supporters were so loyal that he could “stand in the middle of Fifth Avenue and shoot somebody” and not lose any votes. Thankfully, we do not have the data to test this claim.
But Trump’s recent ambiguity on his signature issue, illegal immigration, is imposing a loyalty test nonetheless. After branding his primary opponents as weak on immigration, Trump has now signaled that he may abandon his proposal to deport millions of undocumented immigrants, or at least significantly “soften” it. How will voters respond to this bizarre shift?
Our new research suggests that Trump’s repositioning, if he follows through, has little chance of persuading many people to change their vote.
In an online survey administered by Survey Sampling International and conducted just before the Republican National Convention, we asked roughly 3,000 people to choose one of two hypothetical presidential candidates in repeated head-to-head matchups.
In each matchup, we randomized the candidates’ personal attributes (party, gender, race/ethnicity) and positions on health care, terrorism, immigration and the economy. These attributes and positions reflected those of Trump and Clinton as well as some others. This experiment allowed us to estimate the effect of each issue position on candidate support, including whether a candidate supporting a path to citizenship for undocumented immigrants garnered more support than one favoring mass deportation.
The results suggest that Trump’s latest rebranding is unlikely to matter very much. In the sample overall, which was roughly 50 percent Democrat and 50 percent Republican, there was no detectable change in people’s choices when we presented candidates who supported mass deportation vs. ones who favored a path to citizenship.
Still, it’s possible that Trump could see losses and gains within various subgroups. So we simulated the chances that different types of people would vote for candidates who favored either deportation or a path to citizenship, but who resembled Trump on all other issues and traits:
Favoring a path to citizenship lowers the chances that Trump supporters would vote for a hypothetical presidential candidate by about 10 points. But even after this penalty, these voters still have a 76 percent chance of supporting the Trump-like candidate.
Among non-Trump supporters, the opposite happens: Supporting a path to citizenship brings a 10-point boost in support. But these people still have only a 27 percent chance of supporting this Trump-like candidate.
The bottom part of the graph above suggests that is also unlikely. Favoring a path to citizenship boosts support among these groups, but neither shows more than a 36 percent chance of voting for someone like Trump even with the new immigration position.
One last question is whether reactions to Trump’s potential shift would depend on people’s view of deportation. Perhaps those strongly in favor of deportation would abandon Trump, or those strongly opposed to deportation would start to support him? There is little evidence of this:
Changing the immigration position does alter people’s views in predictable fashion. But even among die-hard supporters and opponents of deportation, a “softening” on immigration would still leave them far apart in their chances of voting for — or against — someone like Trump.
To be sure, there are some caveats here. Our experiment featured unnamed, hypothetical candidates, so it may not capture how voters would respond to Trump himself. Our survey was also fielded in July, and by now voters’ preferences may be even more solidified, making it harder for shifting or “softening” positions to change people’s minds about the candidates.
But there are also reasons to suspect that Trump’s shift would actually produce a larger penalty than our data let on. Our experiment did not explicitly test the impact of “flip-flopping,” where candidates change the positions they previously held. Prior work on flip-flopping shows that candidates tend to lose support when they reverse their positions.
Our experiment also did not test whether candidates’ positions on issues could cause people to abstain from voting altogether. If Trump supporters upset with his shift come to feel they have been duped and choose to stay home on Election Day, Hillary Clinton’s already strong chances of victory would grow even stronger.
Altogether, if Trump’s “softening” really develops, we think it will be too little, too late. Such a shift would only weaken his support among his base, with little chance of bringing new voters aboard.
By adopting a hardline stance during the primaries, Trump painted himself into a corner. Now he’s walking across the wet paint.
Jonathan Mummolo is a PhD candidate in Stanford University’s Department of Political Science.
Sean Westwood is an assistant professor of Government at Dartmouth College.