21 days to before November 8 midterm elections: What do we know?

By Dave Andrusko

NRL News Today is providing a countdown to November 8, examining the most recent polling and predictions about the outcome. Twenty one days to go….and counting.

Two items today. One a Wall Street Journal op-ed, the second an analysis from Hotair.com’s Ed Morrissey.

The WSJ piece is by pollster Douglas Schoen and Andrew Stein, who served as New York City Council president from 1986 to 1994. For those of us who follow these things carefully, there is not a lot of new news. But for bringing it all together in one place, the piece is very helpful.

#1. There’s been a 3-point swing to Republicans in generic-ballot polling in just the last month, they write. “Democrats led by as much as 1.3 points in September, but as of Monday, Republicans were up by 1.8.” Which is important but not nearly as important as the results of a “late September ABC News/Washington Post poll” which “found the Republican advantage concentrated where it’s needed most. The GOP had a 5-point lead among likely voters nationwide, but a 21-point margin in ‘competitive’ congressional districts. ”

#2. “Historical trends favor Republicans too,” according to Schoen and Stein. “In midterm elections since 1982, generic polling averages in the weeks leading up to the election have overestimated the president’s party’s vote margin by an average of 3.5 points, according to FiveThirtyEight. When the president was a Democrat (in 1994, 1998, 2010 and 2014), the overestimate averaged 8.6 points.”

You can attribute this to many possible sources, but it is interesting, is it not, that the overestimation was exceptionally high when a Democrat held the oval office?

#3. “The overturning of Roe v. Wade briefly shifted voters’ attention in a way that was advantageous to Democrats,” they observe. “But the media overstated its effect, as the Washington Post’s David Byler notes in a recent column. Citing the analysis of pollster John Couvillon, Mr. Byler notes that relative turnout in primaries has tended to predict midterm election outcomes, with Democrats having the edge in 2006 and 2018 and Republicans in 2010 and 2014. The majority of 2022 primary voters were in Republican races (52%), rather than Democratic races (48%). The Republican edge held up in states that voted after the Supreme Court decided Dobbs v. Jackson Women’s Health Organization, although it narrowed from 8 points to 2.”

#4. “A more familiar predictor of midterm performance is the president’s approval rating. President Biden’s current approval rating of 42% last month is on par with Bill Clinton’s in September 1994 (42%) and Barack Obama’s in September 2010 (45%), as per Gallup polling. In those years Democrats lost 53 and 63 House seats, respectively. Republican gains this year are almost certain to be more modest than that—but only because the GOP already picked up 14 seats in 2020.”

Switching to Ed Morrissey

He places a lot of emphasis, as he should, on the change from polling Registered Voters to the more accurate Likely Voters. Near the end, Morrissey asks

So what does this mean? It portends a significant wave in House races, to be sure. Thanks to the structure of generic ballot polling, Democrats generally need to get to D+4 or D+5 to hold serve in an election. An R+2.2 — if that’s where it stays — suggests a massive set of pickups, at least in a relative sense.

One other point. The Republicans are starting from a position of near-parity in the House, as Schoen and Stein noted. “Don’t forget that Republicans already have 212 seats in the House, so a 30-seat pickup would give them a blowout majority and put them close to 2010 territory,” according to Morrissey. “If this lead widens any further, it may well also point to a wave that could lift marginal GOP candidates for Senate into office as well. That is what happened in 2014, for instance, an outcome that pollsters missed entirely.”

Talk to you tomorrow.