Americans largely agree on at least two points on the Charlottesville riot and response, according to a new CBS poll taken Monday through Wednesday. First, 63% of respondents agree that the car attack on the crowd that took the life of one woman and injured 19 others qualifies as domestic terrorism, a finding that has majority support among all three partisan demographics. Second, majorities in all three agreed that Donald Trump needed to do more after his initial statement on Saturday.

After that, it goes off the rails a bit for the president. The poll shows Trump at 34/55 for his response to the Charlottesville incident:

He gets majority disapproval overall for his response to the events, while most Republicans approve. Republicans interviewed following Tuesday’s press conference also feel Mr. Trump is assigning blame accurately in the matter, while Democrats and Independents, and the country overall, disagree.

Post-Tuesday afternoon, views on the president’s description of events are tightly tied to overall views of his handling of the matter. Independents and Democrats, and the country overall, feel his description of blame in the events is inaccurate; Republicans feel it is accurate.

Specifically, Democrats disapprove in the extreme, 10/82, which is no surprise. Independents also strongly disapprove, 32/53, not far off the overall result, but Republicans are more supportive. Still, at 67/22, it’s not exactly a slam dunk among GOP voters, and the timing of the poll in relation to Tuesday’s press conference is likely to raise at least some question as to whether that’s an accurate depiction of current support. The numbers are almost identical for the question of whether Trump laid blame for the incident accurately:

In a new CBS News poll, 55% of Americans say Pres. Trump’s view that “both sides” are to blame is inaccurate: https://t.co/pVELgSkxhk pic.twitter.com/hxoyfoDtxp

— CBS News (@CBSNews) August 17, 2017

Note that the CBS graphic says that the sample for this was from “Aug 15-16,” with a ±4% margin of error. The toplines don’t make that distinction; that’s apparently from the whole sample, which has that MoE, and not just from the respondents reached after Tuesday’s presser. Even if the toplines broke down the responses by day (it doesn’t), the sample taken on Tuesday evening and Wednesday would probably be to small for any reliable observations. The attempt to rely on Wednesday’s results alone is problematic at best, especially without seeing the entire daily breakdown. Perhaps they should have scrapped Monday and Tuesday’s responses and started fresh on Wednesday.

With that in mind, the results for Monday’s statement are not very promising. As noted above, majorities in every political demo thought Trump needed to do more after Saturday’s “all sides” statement, including 50/40 among Republicans. Fifty-one percent of Republicans also agreed that the car attack was a case of domestic terrorism. Overall, 60% of Republicans thought that Trump’s statements were “about right,” but two-thirds of the poll had been conducted before Tuesday’s statements. One has to wonder how much that may have eroded on Wednesday, but it seems pretty clear that most people probably got what they wanted on Monday, and that revisiting the issue was almost certainly an error.

In some ways, though, this is the same as it ever was:

Overall, Americans are more apt to say Mr. Trump’s policies have encouraged racial division rather than racial unity, but there are big partisan splits here, too, as Democrats see division and Republicans see no impact. Views on this appear to be entrenched for now, as these percentages are similar to what this poll found in April.

On that point, the question on Trump’s statements is also instructive. Among Democrats, 47% said Trump didn’t go far enough, while 34% said he went too far, which … doesn’t leave a lot of room for approval under any circumstances. (Among independents, it was 15% and 35% respectively.) In other words, there seems to be a lot of narrative validation going on, but even if that’s the case, then Trump isn’t succeeding in selling his narrative outside of his base.

For those keeping score, the weighted partisan D/R/I breakdown of the 1,223-respondent sample is 31/24/45 for a reasonable-ish D+7, with independents somewhat overweighted.

Update, 12:16 ET: CBS appears to have pulled this release from its poll. The link now goes to other questions in the same survey relating to North Korea. There is no explanation for this change at the link, either.

Update: They had just tweeted the story out an hour earlier, but the link now dead-ends.

https://twitter.com/CBSNews/status/898202765289291777

In case they delete the tweet, here’s a screen-cap of it:

Update, 2:11 pm ET: Without any explanation, the link now directs to the original poll results on Charlottesville. The new story offers a bit more detail on the differences between Mon-Tue polling and Wednesday’s results. Strange that CBS News didn’t just update the post in place.