President Trump gives two thumbs up to the crowd during a campaign rally Monday in Cleveland, on the eve of the midterm elections. (Carlos Barria/Reuters)

Since news of the “migrant caravan” broke in mid-October, President Trump has made immigration the focal issue in the upcoming midterms, calling this the “election of the caravan.” Trump has claimed that the caravan is composed of criminals, that it includes individuals from the Middle East and that migrants could be shot by the U.S. military if they act aggressively.

The strategy, it seems, is to stoke xenophobic fears to court Republican votes and remind voters about his promise to build a wall along the U.S.-Mexico border, a signature, yet unfulfilled, Trump campaign pledge from the 2016 election.

Will these tactics help Republicans going into Tuesday’s elections? Our data suggests that the answer is no. But more than that, after Tuesday, Republicans in Congress may find it difficult to deliver on an issue that Trump has now doubled down on: immigration reform.

How we did our research

Given Trump’s anti-immigrant rhetoric, we might expect to see that the relationship between immigration attitudes and midterm vote intentions increased during the 2018 campaign season. To test this possibility, we analyzed data from our nationally representative panel survey of likely voters. This survey, fielded for us by NORC at the University of Chicago, is unique because the same respondents indicated their vote intentions and answered the same three questions measuring immigration attitudes in early July and again in late October.

Specifically, we asked the same 957 likely voters whether they support or oppose building a wall along the U.S.-Mexico border, whether recent immigration will take jobs away from people already here, and whether “illegal immigrants” living in the United States should be allowed to stay and apply for citizenship. Because we measured immigration attitudes well before (survey fielded July 3-12) and just after Trump’s attacks on the migrant caravan began (survey fielded Oct. 19-29), if Trump’s comments are affecting immigration attitudes, we should be able to detect it.

To test how immigration attitudes relate to vote intention, we combined responses to the three immigration questions into an overall “anti-immigration” index and used this to predict intention to vote for a Republican or Democrat in the upcoming House elections. Not surprisingly, anti-immigration attitudes were a significant predictor of Republican vote intentions at both time points.

Here’s the key question: Were anti-immigration attitudes a stronger predictor of vote intentions in late October than in early July? If Trump’s caravan rhetoric was in fact leading voters to draw upon their immigration attitudes when deciding their voting preferences, we’d expect a stronger relationship in our October data.

This was not the case. The relationship between immigration attitudes and vote intentions was statistically equivalent at both time points. This means that despite Trump’s recent fixation on immigration, we find no evidence that likely voters changed how much they base their vote choice on their immigration attitudes.

It’s possible that the goal of these anti-immigration statements is to bring out the Republican vote. Our research design also lets us test this possibility. In the late-October survey, we randomly selected 136 respondents who had previously indicated that they were not likely to vote but who now indicated that they were likely to vote. This could be an important group for Trump, who has cast the midterm elections as a “referendum” on his presidency to date.

Strikingly, we find that the relationship between immigration attitudes and vote intention is even weaker for this group. In other words, the vote intentions of those just tuning into the midterms don’t seem to be any more driven by immigration attitudes than those who have been likely voters from the start.

A third possibility is that Trump has made likely voters more anti-immigration. If that’s the case, Trump could still benefit even though the relationship between immigration attitudes and vote intentions has not changed. To test for a Trump effect on immigration attitudes, we compared the percentage of likely voters who endorsed anti-immigration views in the July and October surveys. Across all three questions, anti-immigration views were highly stable; no statistical differences emerged.

Although Republicans expressed much more opposition to immigration than Democrats, the attitudes of both partisan groups remained stable over time. Nor did late-breaking likely voters (those who were in our original sample but didn’t pass our screening criteria until October) express different immigration attitudes compared with early likely voters. To summarize, there is no evidence in our data that Trump’s recent rhetoric is influencing immigration attitudes or how these attitudes relate to vote intentions.

How Trump’s rhetoric may hurt Republicans

But these results don’t mean that Trump’s words have no effect of any kind. In fact, our data shows that immigration has become very salient to many in Trump’s base. In our October survey, we asked: “After the midterm election, what is the single most important issue for Congress to address?” Responses ranged widely — from health care to inequality, education and racism — but immigration-related responses stood out, particularly among Trump supporters.

If immigration is important to U.S. voters, how will this play out? Our findings may spell trouble for Republicans, in that immigration is a salient issue but that attitudes toward immigration have remained stable overall and in relation to vote choice between early July and late October. The GOP has been unable to enact immigration reform even though it controls the House, Senate and White House. Now, despite historical lows in undocumented immigration, Trump has focused his base on immigration — with no apparent electoral benefit.

If the “blue wave” washes over the United States on Tuesday, Republicans in Congress may soon be wishing that the president had rallied behind a different issue.

Peter K. Enns (@pete_enns) is an associate professor of government and the executive director of the Roper Center for Public Opinion Research at Cornell University.

Jonathon P. Schuldt (@JonathonSchuldt) is an associate professor of communication and a faculty affiliate at the Roper Center at Cornell University.