Last week, the Washington Free Beacon’s Lachlan Markay wrote about how Sen. Bernie Sanders and Hillary Clinton were running neck-and-neck in the upcoming Democratic Nevada caucus this Saturday (via WFB):
The poll shows Hillary Clinton and Bernie Sanders tied at 45 percent each among likely caucusgoers. Clinton narrowly edges Sanders among those who have completely made up their mind. But undecided caucusgoers and those who might change their mind say recent scandals involving Clinton make them significantly less likely to support her.
The survey was conducted between Feb. 8, the day before the New Hampshire Democratic primary, and Feb. 10, the day after. It shows no bump for Sanders after his 22-point victory in the nation’s first primary; by a single point, respondents the following day said they supported Clinton.
The demographic breakdown in caucusgoer support reflects the trends of the previous two Democratic presidential contests. Sanders enjoys overwhelming support among young people; by 63-16, voters between the ages of 18 and 29 support the Vermont senator.
Clinton leads by seven points among women, but women under 30 back Sanders by a 40-point margin.
The survey was conducted by Targetpoint Consulting, and includes responses from 1,236 likely Democratic caucusgoers. The margin of error is 2.9 percent.
Yet, MSNBC’s Rachel Maddow cried foul, noting that the Free Beacon is a conservative outlet (for which they offer no apology and rightfully so). And Targetpoint is a Republican polling firm. Oh, and they only sampled the Democrats, but didn’t ask for a poll for the Republican caucus which is held the following Tuesday. She hinted that the Beacon might be trying to influence the election, or something, by putting a poll in the field that only sampled Democrats, which turned out to be pretty much correct; CNN’s poll found that Sanders and Clinton are split 48/47 respectively. Over at Hot Air, Ed Morrissey makes the point that the Free Beacon poll might actually be more accurate since their poll sampled 1,236 likely caucus goers; CNN sampled 1,006 people, of which only 282 are likely to caucus. The minimum benchmark for an accurate poll sample is 350 people for a congressional race, at least 600 for a statewide race, and 1,000 people (minimum) for a national poll:
Now the grain of salt: the sample size on this CNN/ORC poll is remarkably low, at least when it comes to likely Democratic caucus-goers. Despite a polling cycle of six days (February 10-15 inclusive), the pollster only made contact with 1,006 adults, and only 282 of those were likely participants for Saturday’s Democratic contest. That leaves the poll with a 6.5% margin of error, which seems like a rather large amount of uncertainty for a poll published four days before an electoral event.
However, it does align rather well with another poll taken about the same time. TargetPoint conducted a survey for the Free Beacon of 1,236 likely Democratic caucus-goers, as reported by Jon Ralston, and found almost exactly the same topline — a 45/45 tie. If the numbers in CNN’s poll are correct, Hillary has lost a 16-point lead since October in Nevada.
If true, this is some serious bleeding being exhibited by Team Clinton. For starters, it should raise some red flags that she’s dropped by double-digits (again) in a year that was supposed to be smooth sailing to the nomination. I’m not saying she won’t be the 2016 nominee; I think she will, albeit a very, very bloodied one. Yet, one area to look is how much of the nonwhite vote Sanders gobbles up this Saturday. This is his campaign’s Achilles’ heel heading into Nevada and the primaries below, in, or around the Mason-Dixon line, where nonwhite Democrats should form the core of Hillary’s so-called “firewall.” If Sanders can expand past his young, white liberal base, which doesn’t represent as much of the primary electorate (South Carolina’s 2008 electorate was 55 percent black) in these regions, it could be a huge red flag for the Clinton campaign. We shall see what happens.