President Trump speaks to reporters after the Congressional Republican Leadership retreat at Camp David, Md., on Jan. 6. (Yuri Gripas/Reuters)

During an Oval Office meeting with senators from both parties last week, a frustrated President Trump lashed out at the idea of allowing immigrants from poorer countries to enter the United States.

“Why are we having all these people from shithole countries come here?” he asked about Haiti, African nations and El Salvador, according to people in the room. The Washington Post’s reporting on his comment immediately began days of analysis of what he might have meant, if he actually used a different word (such as “shithouse”) and, naturally, what he might have meant by that word, if he used it.

The question also started another line of questioning that unfolded on Google. On the afternoon of Jan. 11, after The Post report, searches for “is Trump racist” spiked in the United States.

This is a subjective question, of course. Whether something is racist often exists in a gray area. Nazis and the Klan are racist, nearly everyone would agree, but there are many more subtle gradations of thought that can often be subject to debate — or that are vigorously debated in part to muddy the water over whether they should be considered racist.

In the case of Trump’s comments about the mostly nonwhite countries from which he wanted to curtail immigration (unlike the mostly white country of Norway), we have a new poll from Quinnipiac University that tells us just how likely Americans were to consider them racist.

Because they asked.

Nearly 6 in 10 Americans think that Trump’s comments were racist. That includes majorities of independents and nearly all Democrats. It includes just over half of whites, 7 in 10 Hispanics and nearly 9 in 10 black Americans. White men, Republicans and white people without college degrees were more likely to say that the comments weren’t racist.

Which makes sense. After all, the groups that said Trump’s comments were racist were also mostly the same people that told Quinnipiac that he doesn’t respect people of color as much as he respects white people.

When I say they were “mostly” the same people, that’s an understatement. Ninety-three percent of those who said Trump doesn’t respect people of color as much as he does whites also said the comments were racist; 82 percent of those who think Trump does respect people of color said that the comments weren’t racist.

The overlap between views of Trump’s comments and his broader views of race runs across demographic groups. The diagonal line below marks the place where views of Trump’s comments and his general attitude toward people of color are equivalent. Each dot represents a demographic group. They are all tightly clustered around that diagonal line — meaning that the percentage of each group holding the view that Trump respects people of color less is about the same as the percent saying his comments were racist.

Quinnipiac asked the question about respect in December, too, and the patterns were similar — although the percentage of some groups that said Trump respected people of color less were slightly lower. This is probably a function of margins of error more than anything, although the fact that this most recent poll was conducted over a holiday weekend may have added some variability, as people may have been harder to reach.

What this suggests is that views of Trump’s comments largely overlapped with people’s existing views of Trump. Perhaps that’s because those who viewed him as showing less respect to people of color also then were less charitable in their interpretation of his comments; perhaps it’s because the comments were hard to interpret in a more charitable way.

It also suggests that, barring Trump explicitly endorsing Nazism or white supremacy, interpretation of racially loaded comments by the president will continue to be seen through people’s existing lenses of his attitudes toward race.

Scott Clement contributed to this report.