For now, the 2020 Democratic presidential nomination race has settled into two tiers. The top four candidates are former vice president Joe Biden, Mayor Pete Buttigieg, and Sens. Elizabeth Warren (Mass.) and Bernie Sanders (Vt.). The second tier of, essentially, everyone else has difficulty getting media attention and becoming part of the public discussion.

That’s what makes recent attention to Rep. Tulsi Gabbard (Hawaii) so notable. Former Democratic presidential nominee Hillary Clinton charged that Gabbard is being “groomed” by the GOP to run as a third-party candidate, to drain votes from the Democratic nominee. Gabbard tweeted back that Clinton is “the queen of warmongers, embodiment of corruption, and personification of the rot that has sickened the Democratic Party for so long.”

In the Atlantic, author Tom Nichols suggested that this exchange boosted Gabbard, helping her stay in the race until the nominating convention. And when Gabbard announced that she would not seek reelection to her House seat, some speculated that she might indeed run as a third-party candidate.

Clinton’s comments may have done what she intended: signaling to other Democrats that Gabbard is untrustworthy. By drawing attention to Gabbard, did Clinton change how Democrats think about the candidate from Hawaii?

I can answer this, at least partly. In partnership with Civiqs, I am conducting polls using Civiqs’ panel in Iowa. We have conducted two of six planned online monthly surveys that include 598 Iowans who are likely to attend the caucuses in January. The second came just after the October debate, as Clinton’s comments were hitting the news. Using these surveys, we can better understand who supports Gabbard, who opposes her and how these people have shifted.

What do we know about Gabbard supporters?

To start with, there aren’t that many. In the September survey, Gabbard was at 4 percent. That’s higher than in any national poll or other Iowa poll. In October, that support was down to 2 percent.

Gabbard’s supporters are overwhelmingly male. She got 7 percent of the men — and less than 2 percent of the women — in the September sample. In October, Gabbard was the first choice of 3.3 percent of the men and just 1.9 percent of women. Less educated voters are more likely than others to support her; she had 15 percent support from those with only a high school degree. And her supporters are less interested in politics than other likely Democratic caucus-goers, and were more likely to have backed Sanders in 2016.

But there’s a bigger difference: Gabbard’s supporters are less positive about the rest of the field. Our poll asks respondents their first and second choice, giving them a full list of candidates. Next, we asked them two open-ended questions: Who are they considering, and who do they not want? Respondents could list as many as they want — but most listed their first and second choices and then, on average, a bit less than one additional candidate. Further, they were opposed to, on average, 1.4 candidates. Gabbard’s supporters, like the average respondent, were considering one additional candidate beyond their first two. But Gabbard supporters easily listed the highest number of candidates they did not want to be the nominee, with an average of 2.28.

Finally, our full sample also included Iowa Republicans; we asked them to name their first choice of candidates for the Democratic nomination. In both waves, their top choice was Gabbard, giving her 13 percent in September and 10 percent in October. She appears to be more popular among Republicans than Democrats in Iowa.

Who opposes Gabbard?

In both waves of our survey, we asked respondents to list any candidate they “DO NOT want to be the nominee.” In September, only 7 percent of likely caucus-goers said they did not want Gabbard, well behind the 31 percent who did not want Biden.

But by October, a much larger percentage opposed Gabbard: 17 percent. That’s a smaller group than opposes Biden, with 31 percent, and Sanders with 19 percent — but it’s at least 5 percent more than opposed any other candidate.

In the October survey, we interviewed a random selection of a little over half of our September respondents. A sizable portion — over 10 percent — of the respondents who were not opposed to Gabbard in September did say they were opposed to her in October.

Who are these Iowans? Three things stand out. First, they are more interested in politics than the rest. Gabbard is opposed by 20 percent of likely caucus-goers who are interested in politics “most of the time.”

Second, the more candidates the person is considering, the more likely he or she is to oppose Gabbard. The more candidates a respondent finds acceptable or even likes, the more likely he or she is to oppose Gabbard.

Finally, the day the person took the survey matters. Our survey went into the field on Oct. 18, the day Clinton’s comments made news. We can split the sample into the first day of interviews vs. all other days.

The figure above shows the percent of respondents who opposed Gabbard for three groups: respondents in the September wave; respondents interviewed on Oct. 18; and respondents interviewed after Oct. 18. On Oct. 18, 14 percent of respondents did not want Gabbard to be the nominee. After Oct. 18, a quarter of respondents opposed Gabbard. In this group she is the candidate named most frequently, barely surpassing Biden for this dubious distinction.

Finally, the respondents who do back Gabbard have dug in their heels. Remember that in September, Gabbard supporters reported that they were considering an average number of other candidates, or just under one. In October, the average Gabbard supporter was only considering 0.2 other candidates, far fewer than average — meaning most weren’t considering anyone else at all.

So might Clinton’s comments alter the race?

Clinton’s attack does seem to have tarnished Gabbard’s image among Iowa Democrats, signaling that Gabbard is not credible. If Gabbard backtracks on a recent pledge and decides to run in the general election as a third-party candidate, her support among Republicans and, now, opposition among Democrats means it’s possible she could pull more voters from President Trump than from the Democratic nominee.

In other words, Clinton might have damaged Gabbard — and helped the eventual nominee — more than some people originally assumed.

Read more:

David A.M. Peterson is the Whitaker-Lindgren Faculty Fellow and professor in the department of political science at Iowa State University and co-author of “Ignored Racism: White Animus Toward Latinos” and “Mandate Politics.”