At the moment — of late, political moments tend to be vanishing so don’t hold it against me if this has changed by the time you read this — the faction that’s driving the evolution of the Republican bill to replace Obamacare is the House Freedom Caucus, a group of staunchly conservative members who feel as though the “repeal-and-replace” effort has too much replace, not enough repeal. After negotiations with Republican congressional leadership and the White House, it appears that the American Health Care Act, as the bill is known, still doesn’t meet their requirements.

But while the Freedom Caucus constitutes a large bloc of votes, it is by no means the only group within the Republican caucus that is balking. In fact, opposition to the AHCA is fairly well-distributed across the Republican political spectrum.

Using the Upshot’s meta-whip-count of opponents to the bill, I pulled data on partisanship of the Republican caucus and those who have expressed opposition. (We only used those identified as opposing the AHCA by two or more media outlets.) This list, too, is evolving quickly, so I’ll timestamp this as of 2 p.m. on Thursday.

I looked at five metrics.

The first was the Cook Partisan Voting Index, a measure of the partisanship of each congressional district. The red dots are known AHCA opponents; the gray circle is the average of the Republican caucus on the whole.

What you want to note is the number of red dots on either side of the gray circle. About half of those who oppose the bill come from less-Republican-than-average districts; about half come from more-Republican-than-average. In one sense, this makes perfect sense: Averages generally land in the middle of populations of people, by definition.

But in politics, it’s usually the case that opposition arises from one end of the political spectrum or the other. That conservatives are balking at something moderates are comfortable with, or vice versa. Here, the temperature is even throughout.

Same with the margin of the vote in 2016. About half of those who oppose the bill represent districts that were more strongly pro-Donald-Trump than average; about half, less. (This data was compiled by DailyKos.)

Considering the partisanship of the members of Congress themselves, using DW-NOMINATE data from VoteView — a bit less even. About twice as many opponents of the bill have a higher score than average — meaning they are more conservative — than have a lower-than-average score.

This is an important factor, of course, and it largely reflects the relative solidity of the Freedom Caucus. Nineteen of the 38 members of the House that we have identified as opponents of the bill are members of the caucus. All 16 who were evaluated by VoteView have partisanship scores higher than the average. Outside of that caucus, three-quarters of those with scores were below average in terms of their partisanship.

We also considered the likely effects of the passage of the legislation. At least one House Republican, Rep. Ileana Ros-Lehtinen (Fla.) announced her opposition based on the likelihood that constituents would lose coverage. The Center for American Progress developed estimates for each congressional district of the number of people who might lose coverage overall and, because data weren’t available for all states in that regard, how many were likely to lose individual coverage.

Again, the results were fairly evenly distributed.

(Ros-Lehtinen’s district is not represented on the first chart above, but is much higher than average in the second.)

The challenge for House Republican leaders is figuring out how they might find a bill that can appeal to enough people to ensure that the bill passes. Solidifying the Freedom Caucus would help, but that’s only part of the vote that Speaker Paul D. Ryan (Wis.) and his team need to secure. The problem that follows is that the Freedom Caucus — on the right side of the spectrum — will want different and perhaps mutually exclusive things than people who are more moderate than average. Adding something to secure their votes may mean abandoning other votes somewhere else.

That’s the challenge. And it’s one that it doesn’t yet seem Ryan et al. have met.