A follow-up to last night’s new benchmark in RINO poll analysis. Skip to 1:00 of the clip for the key bit.

Cruz is making two points here that were also made in last night’s comment thread: First, the sample was 43D and 32R, which is a mighty big D spread, and second, “20 percent of the people polled were government workers.” On the first point, Sean Trende of RCP assured me last night that it’s not crazy to have a partisan spread like that in a poll of American adults, especially at a moment when the GOP is in momentary disfavor with the wider public. NBC itself says, “That 11-point Democratic advantage is consistent with the combined data for all of the NBC/WSJ polls this year – a 13-point Democratic edge.” Of course, partisan identification is a dynamic thing; if Healthcare.gov ends up being taken offline for months due to necessary repairs and there’s a bureaucratic nightmare over O-Care enrollments, rest assured that the +11 Democratic advantage will shrink. If you want to attack the partisan spread, a better attack is to note that polls of adults are the least accurate way (versus polls of registered voters and likely voters) to predict elections. Although, true as that might be, I think it would miss the point. The takeaway from last night’s poll isn’t that the GOP is doomed next November, it’s that the current shutdown over ObamaCare isn’t moving more people to oppose O’s pet boondoggle and isn’t benefiting Republicans more broadly.

As to the second point, that “20 percent of the people polled were government workers,” it’s literally not true. Overall, around seven percent of the population works for the government — a number that, contrary to popular belief, hasn’t risen much over the last 40 years. But the poll didn’t claim that 20 percent of the sample was employed by the government. It claimed that 20 percent is either employed by the government or lives with an immediate family member who is.


If your spouse, mom, dad, son, or daughter works for government at any level and still lives with you (and, thanks to the magic of the Obama economy, adult children living with their parents is less uncommon than it used to be), then you’re in the 20 percent. And that includes all manner of government jobs, according to NBC, from military to local public-school teacher. I can’t find demographic data that precisely matches this question but 20 percent doesn’t seem wildly implausible as a measure of how many families include some sort of government employee. And even if the sample is an overestimate, it’s surely not so great an overestimate that the results of the poll would differ drastically had a proper sample been used. We’d be looking at a modest adjustment of the data, not some new finding that ObamaCare’s popularity had totally collapsed.

One other point. Cruz begins by saying “the polling has varied,” which is true when it comes to the precise numbers but less so when it comes to trends. ObamaCare’s unpopularity varies a bit from poll to poll but there’s no recent poll I know of showing the bottom dropping out. On the contrary, there are two polls that suggest it’s less unpopular this month than it was last month. Virtually every poll shows the GOP losing popularity after the shutdown and, what’s more, losing it at a steeper rate than either O or congressional Democrats. Both WSJ/NBC and Gallup showed Republicans at historic lows this week, even though of course their exact numbers diverged. The one metric that has been bouncy is Obama’s job approval, which the AP had as bottoming out at 37 percent a few days ago and WSJ/NBC had climbing slightly in last night’s poll. A result like that on the GOP’s or ObamaCare’s popularity would be noteworthy variance. But like I say, we’re not seeing that. (Yet?)

To reiterate, though: None of this means the GOP’s going to lose seats next year in the midterms. All I’d add to Nate Silver’s piece today making that same point is that, as noted last night, the last time Republicans were below 40 percent on the generic ballot in a WSJ/NBC poll was October 2009. A year later, thanks to ObamaCare, they crushed Democrats in November to take back the House. Parties can have bad spells and then recover almost overnight depending on events. The lesson from the poll isn’t that “WE’RE DOOMED!!1!”, it’s that we’re doing some political damage to ourselves here and getting nothing major for it on the Hill. The last time the Democrats decided to gamble significantly on a policy gambit, at least they got universal health care in exchange. I’ll leave you with this from John Nolte of Breitbart, since I’m on the same page he is: