Measuring the Midterm Intensity Gap

There is a lot of talk about intensity of voters these days, and for good reason. It was the greater intensity of support among Donald Trump backers and the relative ambivalence of many who intended to support Hillary Clinton that likely made the difference in Michigan, Pennsylvania, and Wisconsin, where a combined 78,000 votes determined the winner.

But how much of an intensity gap is there today between the Democratic and the Republican sides? The latest ABC News/Washington Post poll released this past week suggests one thing; the new NBC News/Wall Street Journal poll says something different. The analysis from ABC that came out early Monday morning had the headline, “2018 Vote Margin Narrows as Democratic Engagement Slips,” and the accompanying Washington Post story by Dan Balz and Scott Clement was headlined, “Poll: Democrats’ advantage in midterm election support is shrinking.”

But for the other survey, NBC News’s Mark Murray’s piece was titled, ”Poll shows Democrats have midterm intensity advantage, but no ‘knockout’ yet” with a subhead, “Republicans are behind, but still have a chance of keeping the House, says a GOP pollster.” Janet Hook’s analysis in The Wall Street Journal was headlined, “Democrats Maintaining Advantage Over GOP in Voter Sentiment Poll” with a subhead, “Democratic voters are showing a higher level of interest in midterms than Republicans, but the party’s edge has slipped in who should lead Congress.”

The two polls were taken precisely the same days, April 8-11. In my judgment, neither is an outlier—the statistical 1-in-20 poll that may have been perfectly executed but produces results outside of the margin of error. Occasionally we see polls and wonder which country they are interviewing, but different pollsters employ different methodologies for measuring intensity and likelihood of voting, while different journalists put more weight on one question or approach than another.

The ABC/Post analysis compared a new poll with one in January and looked at the results first among all adults, finding Democrats with a 10-point lead on the generic congressional ballot test, but among just the registered voters the margin fell to 4 points (it was 12 points in January), and among those who were both registered and said they were certain to vote, the Democratic edge was 5 points.

But the NBC News survey write-up observed that “the current poll shows Democrats with a significant advantage in enthusiasm, with 66 percent of Democrats expressing a high level of interest (either a ‘9’ or ‘10’ on a 10-point scale) in November’s elections, versus 49 percent for Republicans.” NBC’s Murray added, “That’s a reversal from the merged NBC/WSJ polling data in 2010—a wave year for Republicans—when 66 percent of Republicans expressed a high level of interest, compared with 49 percent for Democrats. … And among these high-interest voters in this new poll, Democrats lead Republicans in congressional preference by 21 points, 57 percent to 36 percent.”

Both of these polls are top-notch, the ABC/WP poll under the auspices of Gary Langer from the former, Scott Clement the latter, while the NBC/WSJ polls are done jointly with Bill McInturff of Public Opinion Strategies on the GOP side, and Fred Yang and Peter Hart of Hart Research on the Democratic side. All of the journalists involved are among the best in the business—none of these folks are shading or spinning—but even slightly different approaches can produce somewhat different results and conclusions. Starting off with people who have said they are registered to vote and asking either if they are certain to vote, or how likely they are to vote on a scale from 1 to 10, are both perfectly reasonable ways to do it; there isn’t a consensus on which way is better.

The takeaway from both (as well as other high-quality, reputable surveys from CBS News, CNN, and Fox News) is that Democrats are ahead, but the size of Democrats’ lead and intensity advantage have yet to be determined.

One finding in the NBC/WSJ poll that jumped out was that while Democrats had a 27-point lead among registered voters in House districts held by Democrats, Republicans had just a 10-point advantage in districts occupied by Republicans. That is consistent with what those of us who watch individual House races find: Democrats have very few districts in any jeopardy, while Republicans have a lot, which is the reverse of what existed in 2010 and 2014.

No prudent person who knows anything about congressional races believes either that Democrats have control of the House locked up or that the Republican majority is safe. The consensus is that the GOP majority is in danger. That is the conclusion from the macro approach, looking at national polls (which didn’t miss the national popular vote by much in 2016), and it’s also the finding using the district-by-district, micro approach. My bet right now is still a 60-65 percent chance of the House flipping, but we have a ways to go.