This past week, both Pew Research and ABC News/Washington Post released polls that aim to measure the public’s view of the government shutdown.

Overall, both suggest that more people oppose the wall than support it – ABC/WP by 54% to 42%; Pew by 58% to 40%. But the two polls find very different feelings about intensity. I look at the reason for that disparity below.

Both polls also sought to measure the public’s attitudes about the wall as it relates to the current government shutdown. Essentially, how many people would be willing to give up their positions on funding the wall if that would end the shutdown?

On this issue, the two polls present very different results.

The ABC/WP poll suggests about half the public can be classified as hardliners, those who are unwilling to end the shutdown unless their positions prevail. However, they split about evenly: 27% against the wall, 24% in favor.

By contrast, Pew suggests 80% of the public can be classified as hardliners: 51% against the wall, 29% in favor.

The differences between the two sets of results are no doubt at least partially caused by differences in question wording.

ABC/WP asked its respondents who opposed the wall if Democrats should compromise and asked supporters of the wall if Trump should compromise.

Pew asked its respondents who opposed the wall if it would be “acceptable” or “not acceptable” to pass a bill that would include funding for the wall if that were the only way to end the shutdown. Supporters were asked about the acceptability of excluding funding for a wall.

The results suggest that framing the issue as one of compromise induces more accommodating behavior than when the issue is framed as whether it’s acceptable or not acceptable to agree to their opponents’ wishes.

In polling terms, the word “compromise” may suggest a “socially desirable” point of view, encouraging respondents to choose that option because otherwise they may be judged negatively by the interviewers. Use of this word may underestimate the firmness of people’s opinions. At best, however, this is a subtle dynamic, and it’s difficult to measure.

Still, it’s possible the public may be more strongly polarized than what the ABC/WP poll questions on this issue revealed. A separate ABC/WP question asked:

“Presidents can declare a state of national emergency, giving them special powers to take action without approval from Congress. Do you support or oppose Trump using emergency powers to build a wall along the U.S.-Mexico border? Do you feel that way strongly or somewhat?”

Overall, the public opposed Trump’s possible use of emergency powers by 66% to 31%. The people who felt “strongly” were opposed by 51% to 25%.

These results suggest a highly divided public, with fewer people in the middle than on either end of the spectrum. At least at this point in the shutdown, it appears that a substantial majority of Americans are locked into their own point of view.