By the end of this week you’ll have seen two headlines built off the same survey, and neither one is going to be straight with you.
One will read something like “Majority Backs AR-15 Ban.” The other will read “Half the Country Defends the AR-15.” Both will cite Marquette Law School, both will be technically sourced to the same survey, and both are junk. The reason is a number in the fine print that nobody quoting this poll at you has any intention of mentioning.
So let’s take the thing apart.
Marquette put the question to 1,023 adults between September 2 and September 9, using an online panel recruited off a probability sample. That’s the good kind, by the way — actual random recruitment, not a click-here-to-tell-us-what-you-think banner on a website that already agrees with you. On the AR-15, 52 percent said the Supreme Court should find there’s no constitutional right to possess one. Forty-eight percent said the Court shouldn’t.
Four points. That’s the whole “majority.”
Now the fine print. Of those 1,023, exactly 864 were registered voters, and Marquette puts the margin of error on that group at 3.6 percentage points. Among registered voters the split was 51 to 49. Among the 581 they scored as likely voters, margin of error 4.3, it was 50 to 49.
Margin of error is one of those phrases everybody’s heard and almost nobody’s had explained, so here’s the counter version. The poll did not hand you a number. It handed you a range, and the number is just the middle of it.
Think about a three-shot group at a hundred yards. The center of that group is your best guess at where the rifle shoots. It is not where the rifle shoots. Send three more downrange and the center moves, and if somebody told you he knew his rifle’s zero to a quarter inch off three rounds you’d smile and change the subject.
That’s what 51 to 49 with a margin of 3.6 means. The real answer could be 54 to 46 in one direction. It could just as easily be 48 to 52 in the other, which is the same poll telling you the opposite story. Both of those sit comfortably inside the group. The likely-voter number, 50 to 49 with a margin of 4.3, isn’t even a lean. It’s a room somebody counted twice and got a tie both times.
And every one of the splits underneath the topline is built on a smaller slice of that same 1,023, which means their margins are wider still. Men came in at 59 percent backing AR ownership rights. Women went the other way, 63 percent rejecting them. Republicans were 77 percent for Second Amendment protection, Democrats 74 percent for upholding bans, and independents landed at 51-49, which in a sample that size is another way of saying they didn’t land anywhere.
None of that is nothing. It’s just nowhere near what it’s about to get used for.
Here’s the finding I actually can’t stop chewing on. That same poll asked about Bruen, the 2022 decision protecting the right to carry a handgun outside the home for self-defense. Sixty-nine percent favor it.
Same survey. Same week. Same people.
Better than two out of three Americans will tell a pollster you have a right to carry a loaded handgun down a public sidewalk to protect yourself, and then the country splits straight down the middle over a rifle that mostly sits in a safe. Same amendment. Same constitutional question, more or less.
I’m not going to pretend that’s a win and I’m not going to pretend it’s a betrayal. It’s information, and it’s the most useful thing in the whole poll. It means the argument we keep having was never really about the Second Amendment in the abstract. Most people settled that one. It’s about one rifle, and largely about what that rifle looks like, and whatever you think of that, it’s worth sitting with honestly instead of cheering past it.
Marquette also found Supreme Court approval down to 43 percent in September from 47 in July, which the Court will care about roughly as much as the weather in Tulsa.
Because this is the part that actually decides anything, and it’s the part the poll has nothing to do with.
Viramontes v. Cook County gets oral argument on December 2. The reporting says a lot of people who follow this closely expect the Justices to strike down AR-15 bans, and that expectation may well be right — but it’s an expectation, not a holding, and oral argument isn’t a decision either. The ruling lands whenever it lands.
Nine people hear that case. They don’t stand for election, they don’t run cross-tabs, and there is no mechanism anywhere in the process by which 52 percent of 1,023 adults in early September touches Cook County’s ordinance.
A poll tells you what your neighbors think. That’s genuinely worth knowing, and I’d rather know it than not.
It tells you exactly nothing about December 2. Anybody who hands you this thing as a forecast is selling you something, and it isn’t rifles.
