Discussion: Sanders Leads By Nine Points In New Hampshire

Discussion for article #245618

At least a dozen delegates are at stake for crying out loud. Meanwhile:

According to our latest polls-plus forecast, Hillary Clinton has a 96% chance of winning the South Carolin a primary.

Polls are all over the place today. Blah!

1 Like

Tis the week between Iowa/NH.

same as it ever was.


Today’s 'What the Huh??"


Different polls, different periods surveyed. But yes, the clash of headlines illustrates the idiocy of overreacting to any one poll.

By the way, an interesting article over at Mother Jones, where David Corn argues that too steeply discounting a big Sanders win in NH based on “neighbor state effect” doesn’t really match up with historical experience.

A little off-topic (but still on the general subject) Bernie just got a ringing (if not surprising) endorsement from Danny Glover:

I’m obviously not paying as much attention to polls as I should be, but if I read this right, we have a poll from the Boston Globe – a respectable newspaper – that showed Sanders with a 9 point lead over Clinton before last night’s debate. From reading all the news sources, including TPM, I thought that Bernie was ahead by 20, 30, 40 points in NH. But now it appears that at least one credible poll had him ahead by only 9 points, and that wasn’t even mentioned in all the high double digit reporting? So, I wonder whether the Boston Globe/Suffick folks are still in the field doing up to date polling, with the debate being in the rear view mirror. Anybody know?

Dates for the Boston Globe poll were 2/2 - 2/4, same as 4 other polls. The Boston Globe poll showed him up by 9 points. Among the other 4 polls done during the same time frame, one shows Bernie up by 31, and the other 3 show him up by 15 or 16 points. So my guess is that he is up by around 15 or 16 points.

Remember, all these polls are now using “likely voter” models, and each poll makes different assumptions and thus has different criteria about who qualifies as a likely voter and gets counted. So even if all the pollsters asked the questions in roughly the same way, use solid methodology, have good random samples, etc, they can still wind up pretty different.

Add in a 5 or 6 point margin of error, on top of the different likely voter models, and the differences really aren’t surprising in the least. Which is why it’s a bit silly for anyone to get too excited about any one poll, and why it’s utterly ridiculous the way some of the polls are presented in the headlines.

For example, today’s headline on TPM that said “New Poll Shows Sanders Obliterating Clinton’s 31-Point National Lead.” Look, I’m a Bernie supporter, and I’d love to believe that was true. But one poll, so far out of whack with most of the polls, is quite likely to be an outlier.

But, people like to get themselves worked up into a lather over click-bait headlined hype-stories (hey, I’ll admit, it’s a guilty pleasure I indulge in far too often myself) and the purveyors of these stories live off clicks. So in that sense it’s a mutually beneficial arrangement, kind of like symbiosis. Ain’t nature great?

A poll will say what you want it to say .