The CD Media Big Data Poll finds Republican incumbent Senator Ron Johnson leads Democrat Mandela Barnes by just over 3 points in the race for U.S. Senate in Wisconsin, 49.2% to 46.0%. Only 4.5% are undecided and another 0.5% chose "other /write-in".
"The Badger State is a must-hold for Republicans if they intend to retake control of the U.S. Senate," Big Data Poll Director Rich Baris, stated. "That being said, Ron Johnson appears on track to repeat a performance somewhat comparable to his re-election victory in 2016."
By education, voters with a high school degree or less back Johnson 49.4% to 41.4%, though he performs strongest among voters with some college or an Associate degree, 55.6% to 39.8%. Notably, voters with a 4-year degree are nearly split, backing Barnes by only 5 points, 51.1% to 46.4%. Voters who hold advanced or postgraduate degrees heavily favor Barnes, 57.9% to 39.4%.
Men prefer Johnson 54.7% to 43.5% and women prefer Barnes by a much more narrow margin, 48.7% to 43.4%. Further, men (92.3%) are also more "certain to vote" than women (86.1%), setting the stage for a more male-heavy electorate. That has largely been the case in Wisconsin since 2018, when men outvoted women and were 51% of the electorate. In 2020, the electorate was even at 50% male, 50% female.
The electorate in the fall is projected to be more Republican than Democratic. In Wisconsin, machine-learning inference models that use primary vote history, contribution records and scores of other data attributes show Republicans outnumbering Democrats by roughly a 5-point spread—or, 38.58% Republican, 33.95% Democrat, and 27.47% None/Unknown.
The likely voter model used by the CD Media Big Data Poll projects the partisan composition of the electorate slightly less Republican at a R/D/I-O split of 34.5%/31.8%/33.7%. Of those "certain to vote" in November, Johnson leads 50.7% to 45.7%. Of those "extremely enthusiastic" to vote, Johnson also leads 54.2% to 44.3%.
"Our inference models have shown a clear and consistent trend toward Republicans in Wisconsin over the last six years or so," Baris added. "In 2020, the Republican advantage was far larger than most—to include myself—had anticipated."
"The major difference this year is that self-identified independent voters are backing Johnson and other Republicans."
Johnson leads Barnes among self-identified independent voters 52.0% to 38.8%.
As previously stated, men are more likely to vote than women in November. More men (70.4%) are extremely enthusiastic to vote than women (63.8%), as well.
Cost of living and inflation ranked first when asked which issue would be most important to their vote at 26.7%. Abortion was second with 15.9% citing it as their most important issue and the economy and jobs was third with 14.6%. Johnson leads among voters who cited cost of living and inflation, as well as the economy and jobs. Barnes leads among voters who cited abortion. Immigration was fourth, cited by 6.7% of voters who overwhelmingly back Johnson.
President Joe Biden's approval rating is underwater in Wisconsin. While 44.6% approve of the job he's doing as president, to include 23.4% who strongly approve, 54.5% disapprove and that includes 47.9% who strongly disapprove.
"That -24.5% intensity index with only 0.9% remaining undecided strongly suggests Biden will be an anchor weighing down Democratic candidates in November," Baris noted. "This should come as no surprise to anyone."
"Incumbents are difficult to beat even in cycles that favor the party of the challenger, and nearly impossible to beat in cycles such as this one, which favors the incumbent's out-of-power party."
Of those who remain undecided in the race for U.S. Senate, 51.9% strongly disapprove and another 5.0% somewhat disapprove. Only 2.2% strongly approve and 30.5% somewhat approve.
The CD Media Big Data Poll for the Wisconsin Midterm Elections was conducted by Big Data Poll and interviewed 852 likely general election midterm voters statewide via Peer-2-Peer SMS/OSP from September 17 to September 18, 2022. The overall survey sampling error is ± 3.4% at a 95% confidence interval. It’s important to note that sampling errors for subgroups are higher. Results are weighted to represent statewide voter file (Aristotle) demographics to include gender, age, race and ethnicity, education and region. The proprietary likely voter model is determined by both self-reported likelihood and vote history. The full crosstabs can be viewed on MarketSight and a detailed methodology statement can be viewed HERE.
Subscribe to our evening newsletter to stay informed during these challenging times!!