I mentioned in this post that I’d have more to say later about his surprising downturn in Quinnipiac’s new poll, slipping to third place behind Elizabeth Warren(!) at a meager 11 percent. You might dismiss that as an outlier, noting that Sanders was at 22 percent and ahead of Warren by double digits in Morning Consult’s new poll today. Or you might chalk it up to a Biden effect: Since Sanders and Biden are the most well-known candidates in the race, Biden’s entry might reasonably be expected to steal away some “soft” Bernie supporters. For now.

But there’s another possibility. This was tucked away at the end of Quinnipiac’s poll.

A disaster, even though Quinnipiac phrased the question about as charitably as they could to Sanders’s position. They could have emphasized “felons,” “violent felons,” even the Boston bomber by name. Not only is the public against this at a two-to-one clip, even Democrats are evenly split. How many casual Democratic primary voters woke up last week to news that Bernie had gone to bat for Dzhokhar Tsarnaev’s right to vote from death row and concluded that Trump was right — “Crazy Bernie” actually was crazy, or at least close enough that it was time to rule him out? How many progressives, like Ro Khanna, who are squeamish about letting violent felons vote decided that Elizabeth Warren was plenty progressive enough for them instead? How many lefties who are okay in principle with Bernie’s position nonetheless were cowed by the backlash and concluded that, for all his virtues, critics were right that Sanders simply couldn’t win a general election?

After all, if he was willing to lob this political grenade in the heat of a primary, what other sorts of outre positions might he surprise the party with as their nominee next summer?

This isn’t the only recent poll showing strong opposition to letting prisoners vote. Business Insider reported yesterday that 75 percent of people it surveyed online were against the idea, although that poll wasn’t weighted by race or income so it might not be truly representative of public opinion. And it’d be easy to dismiss the Quinnipiac data as one bad survey for Bernie if it had been the only bad poll for him this week — but it isn’t. New from New Hampshire:

Sanders is now tied for second, again right around 11 percent, in a state which his home state borders and which he won easily over Hillary Clinton in 2016. Again, you might see a Biden effect there, preferring to blame Bernie’s slide on Biden’s entry into the race rather than on a backlash to Sanders’s answer on prisoners voting at last week’s CNN town hall. But Biden’s only polling at 20 percent here, not indicative of a campaign-launch bounce. If anything, that’s a bit below where he typically polls in early surveys of New Hampshire. Remember too that Buttigieg also held a town hall on CNN that night and made a point of disagreeing with Sanders about prisoners voting. Now he’s tied for second — although the 12 percent he’s pulling is in line with other pre-CNN surveys of the state. It could be, in other words, that all of this is a coincidence and that Sanders is suddenly struggling for “organic” reasons, like Buttigieg’s surge and growing attention on the left to Warren’s policy slate, rather this his answer on letting prisoners vote. New Hampshire is right next door to the only U.S. states that allow the practice, after all. (Vermont and Maine.)

But maybe it isn’t a coincidence. Maybe Sanders really did confirm suspicions that he’s a wild-eyed radical with his prisoner answer and is suffering as a result. If he did, it’d be half comic and half tragic inasmuch as this is a boutique issue that really has nothing to do with Bernie’s core message. If he’s going to choose a political hill to die on, you would think it’d be related to redistribution, passing Medicare for All, for instance, or taxing the very rich to fund new entitlements. Instead he decided to go with his gut on letting prisoners vote and, at best, it’s done nothing for him poll-wise. At worst, well, just look at the numbers above. Exit quotation: