New polling from NYT/Siena College finds Angie Craig with a big advantage over Jason Lewis in Minnesota’s 2nd Congressional District

Angie Craig
MinnPost file photo by Craig Lassig
The CD2 poll indicated a clear advantage for Angie Craig across key voting blocs.
A new round of polling in Minnesota’s 2nd Congressional District, a battleground in the fight for control of the U.S. House of Representatives in these midterms, shows what most observers of the race already knew: Rep. Jason Lewis is facing a very strong challenge from DFL candidate Angie Craig in what has long been considered a “toss-up” race.

But this week’s poll from the New York Times and Siena College — a live survey that was still under way on Tuesday — suggests this race may no longer be a toss-up: With responses from more than 400 voters in CD2, the poll found Craig with a 14-point advantage over Lewis, 52 percent to 38 percent. Ten percent of voters were undecided.

If that result holds, Craig’s lead would be greater than the poll’s 10-point margin of error, a standard that, in the eyes of most pollsters, indicates a clear advantage for one candidate. Craig’s 14-point advantage would also be one of the biggest leads for any candidate among the 30 U.S. House races nationwide surveyed by the Times and Siena, a group that includes Minnesota’s 3rd and 8th Districts. (In CD3, the poll gave Democrat Dean Phillips a nine-point lead over GOP Rep. Erik Paulsen; in CD8, Democrat Joe Radinovich led Republican Pete Stauber by one point.)

The CD2 poll indicated a clear advantage for Craig across key voting blocs: Among voters 65 and older, Craig led Lewis 53 percent to 34 percent; among women, Craig led 56 percent to 34 percent; and among voters who identified as independent — a group that made up a third of those surveyed — Craig led 49 percent to 35 percent.

Typically, challengers find themselves above water in polls’ favorability ratings, but most people report they don’t know much about them — meaning their opinions could be influenced by negative ads. Incumbents, on the other hand, may be disliked by more voters, but are known quantities in their districts.

The NYT/Siena poll had some notable findings on that front: 31 percent of voters polled didn’t know if they approved of Craig, while 27 percent said the same about Lewis — a roughly similar number, which isn’t especially surprising given that this is a rematch between two candidates who ran in 2016. (Lewis won that contest by 1.8 points.)

But among voters who did have an opinion of the candidates, the results could not be more different: 38 percent of respondents gave Craig a favorable rating, and 31 percent gave her an unfavorable rating. Lewis, meanwhile, was 15 points underwater: 44 percent of respondents had an unfavorable opinion of him, while 29 percent had a favorable opinion.

Craig’s best result yet

The NYT/Siena survey is the best showing yet for Craig out of the four polls on the race that have been made public so far. A SurveyUSA poll, commissioned by KSTP and conducted in mid-September, showed Craig narrowly leading Lewis, 48 percent to 45 percent, which was well within the poll’s nine-point margin of error. Another mid-September poll, conducted by the Democratic firm Public Policy Polling for the advocacy group Protect Our Care, found Craig leading by the same margin.

The only public poll showing the Republican with a lead is a survey from August, commissioned by the Lewis campaign, which had him leading by one point.

Jason Lewis
MinnPost file photo by Steve Date
Poll: 44 percent of respondents had an unfavorable opinion of Jason Lewis, while 29 percent had a favorable opinion.
With just over a month until Election Day — and with Minnesotans already participating in early voting — the NYT/Siena poll results suggest that the race could be breaking in Craig’s favor.

While prior polls in CD2 had President Donald Trump’s approval rating roughly even, the NYT/Siena poll finds him decisively underwater: 37 percent of voters in CD2 approve of Trump, per the survey, versus 56 percent who disapprove. (Trump won CD2 in 2016 by 1.2 points.)

The poll also asked voters their opinions regarding Judge Brett Kavanaugh, the U.S. Supreme Court nominee whose confirmation battle has become a national controversy over allegations that he sexually assaulted women in high school and college. Lewis has vocally defended Kavanaugh, issuing a statement last week — a rarity for a House member on a Supreme Court fight — calling the episode a “sad circus” and arguing it shows that “no individual will be safe from deliberate smears solely designed to attain political power” if Democrats win in the midterms.

The NYT/Siena poll found that 49 percent of those surveyed in CD2 oppose Kavanaugh’s confirmation, while 43 percent support it. Thirty-three percent said they believed the allegations against him, while 24 percent said they did not believe them. (A plurality, 44 percent, said they did not know.)

In the past month, two of the leading election handicappers — the Cook Political Report and Inside Elections — moved CD2 from the “toss-up” column to the “lean” or “tilt” Democratic columns. (Those outlets did the same for the CD3 race.) This week, Politico reported that deep-pocketed GOP outside groups are cutting off certain incumbents who are falling behind in order to shift resources elsewhere; Lewis and Paulsen were mentioned as being in danger.

The poll is ongoing — the NYT/Siena have typically polled anywhere from 500 to 550 people in House races — with responses posted in real time online. The CD2 poll should be wrapped up by Tuesday night.

You can also learn about all our free newsletter options.

Comments (22)

  1. Submitted by Frank Phelan on 10/02/2018 - 01:13 pm.

    If Lewis looses, he can always go home and run in his own district.

  2. Submitted by joe smith on 10/02/2018 - 02:05 pm.

    As we know from all the Liberal polling sites, they are never wrong. Just look at how many had Trump winning in an electoral landslide, zero!!!

    • Submitted by Eric House on 10/02/2018 - 03:12 pm.

      Ahh, there it is. Attacking the messenger- that didn’t take long.

      No one is saying polls are perfect. But, the history of political polling in this country generally shows a lot of accuracy. Of course this doesn’t this mean Angie Craig has won, and we can all go home. At this point, the poll only indicates that Rep. Lewis is very, very vulnerable. But you go right ahead and keep assuming that information that doesn’t fit your preconceptions is “liberal”.

      • Submitted by joe smith on 10/03/2018 - 06:48 am.

        No, it’s called stating a fact.

        • Submitted by Pat Terry on 10/03/2018 - 11:19 am.

          No, you are completely wrong. The final polls had Clinton with about a 2 point lead. And she finished 2 points ahead of Trump. The polls were exactly right.

          • Submitted by joe smith on 10/03/2018 - 01:29 pm.

            Pat, as I stated before , you didn’t respond, name the liberal polling places that had Trump beating Hillary in a electoral landslide.

          • Submitted by RB Holbrook on 10/03/2018 - 04:53 pm.

            Does it make a difference now? It’s 2018–why the continued bickering about the polls in 2016?

            Polls predict probability. Those who report on them imply that they have a greater predictive value than they do.

            Incidentally, I wouldn’t be kvelling too much about Trump’s “landslide” victory. His margin in the Electoral College is the 46th largest out of the 58 elections we have held. I won’t mention the popular vote, because I know perfectly well that that is not how we elect our President (but in only two of the elections in which the popular vote was recorded did the winning candidate get a lower percentage of the popular vote).

          • Submitted by joe smith on 10/04/2018 - 10:12 am.

            RB, Trump won 2,626 counties, Hillary 487…. That along with 306 electoral votes was a message to DC, elites don’t run our country, regular hard working, blue collar folks do.
            The original point was, polling has been wrong before (not one left leaning poll gave Trump a chance) and will be wrong again.

          • Submitted by RB Holbrook on 10/04/2018 - 11:10 am.

            Counties don’t vote. Counties are arbitrary demarcations of local authority. They do not vote.

            “[E]lites don’t run our country, regular hard working, blue collar folks do.” First, on some level, an intelligent fellow like yourself has to know that’s a load of crap. I would be insulted that you would raise that line with me but I am too busy laughing about it.

            Second, the elites are American citizens just as much as “regular hard working, blue collar folks,” like my parents. Their votes count just as much as anyone, and even more than the imaginary votes of counties.

          • Submitted by joe smith on 10/04/2018 - 12:46 pm.

            Counties represent blocks of voters. Hillary won blocks in major cities and along the upper East coast and West coast to the tune of 487. Trump on the other hand won 2,626 blocks of voters, mostly in the heartland. Limousine liberals have been blocked by regular folks in that fly over country where we cling to our bibles and guns. Trump winning 30 states also shows that more regular folks believed we needed a change from the past 8 years rather than 4 more with Hillary.
            I find it ironic that folks on the Left tout Hillary winning the popular vote but down play Trumps total dominance in Counties won. In our Electoral system the goal is to reach the most people living in the 50 diverse States, counties make up States and yes they did vote…… Overwhelmingly for President Trump.

          • Submitted by RB Holbrook on 10/04/2018 - 02:03 pm.

            What are the first words of the Constitution, Mr. Smith? That’s right: “We, the People.” Not “We, the Counties,” or “We, the Blocks of Voters.” The United States is supposed to be about self-governance, meaning government by the people. Tracts of land, however huge, are not in it.

            “We, the People” (now) includes everyone. The academic in Berkeley or the hipster in Bushwick are “real Americans,” just as much as a truck driver on the Range is a real American. Saying otherwise is, at best, as much a form of elitism as anything practiced by tofu-eating urbanites. At worst, it reinforces cultural and racial* stereotypes that should long ago have been consigned to the dustbin of history. There is no one paradigm for a “real” American (what would you say if I told you that I thought the “hicks” in “flyover land” are not real Americans?). Like it or not, most Americans live in urban areas, and they are concentrated especially in large cities (about 1/7 of the population of the US lives in the NY, LA, or Chicago metropolitan areas). Like it or not, they are as entitled to representation as anyone living anywhere else.

            *Yes, racial. If you think Trump supporters are the true representation of “real Americans,” you are talking about people who are overwhelmingly white.

      • Submitted by joe smith on 10/04/2018 - 05:22 pm.

        RB, not sure you remember who called 1/2 of Americans deplorable. We have an electoral college system, so reaching folks in all 50 states is the goal. America is a center right country, no matter what CNN tells you. Hillary and the left’s message of division just didn’t reach enough folks who don’t live in big cities. Not racial at all, just a fact!

        • Submitted by RB Holbrook on 10/05/2018 - 01:13 pm.

          My point was that if you assume that Trump supporters are the “real Americans,” you are painting a picture of a people that is overwhelmingly white.

          Our Electoral College has non-representation built into it.

        • Submitted by Jonathan Ecklund on 10/09/2018 - 10:05 am.

          She called half of Trump voters deplorable. Trump got 62.9 million votes (Hillary got 65.8 million), so she called approx 31.5 million people deplorable, out of an eligible voting population of approx 240 million.

          So, she called 26 percent of Americans deplorable.
          I don’t disagree.

    • Submitted by ian wade on 10/02/2018 - 03:21 pm.

      78,000 votes spread across three states. Trump was an aberration.

    • Submitted by RB Holbrook on 10/02/2018 - 05:18 pm.

      I’m pretty sure Trump is not running in the Second District.

    • Submitted by Tory Koburn on 10/02/2018 - 05:51 pm.

      Even Lewis’ own internal polling only had him up by one percentage point. That should be alarming, as the 2nd has been a fairly solidly republican district since redistricting years ago. But for it even to be this close, or for Lewis to lose, suggests that he is either a worse-than-average republican candidate, or that the district’s demography is becoming less republican. Maybe both.

    • Submitted by Pat Terry on 10/03/2018 - 11:21 am.

      Any polls that had Trump winning in an electoral landslide would have been wrong. So your comment speaks to the accuracy of the 2016 polling.

  3. Submitted by Curtis Senker on 10/03/2018 - 12:16 pm.

    “The margin of sampling error on the overall lead is nine points, roughly twice as large as the margin for a single candidate’s vote share. One reason we’re doing these surveys live is so you can see the uncertainty for yourself.”

    Oh yes, I see the uncertainty pretty clearly.

  4. Submitted by David Therkelsen on 10/03/2018 - 10:58 pm.

    Why would MinnPost report on a poll with a 9-point margin of error? That is wildly below the standard for polling on any topic.

  5. Submitted by Paul Udstrand on 10/08/2018 - 11:35 am.

    I just think it’s funny that Republicans are suddenly anti-wealth/anti-corporate executive guys. Do they not know who their President is? After decades of claiming we need to give the business boys (and girls) a crack at government, they’re suddenly all about trashing Democrats for being wealthy business boys and girls. Whatever, it’s just another sign of demise.

Leave a Reply