EAU CLAIRE -- Charles Franklin, director of the Marquette Law School Poll, has had a rough couple of presidential cycles.
In 2016, polling badly missed in the presidential election, significantly underestimating President Donald Trump’s level of support. Polling organizations, including Franklin’s own, made adjustments after that. And, for a while during the midterms and other elections, it seemed accuracy had been restored.
Then came November.
Polls once again missed this year, albeit by slightly smaller margins than four years ago. But there’s still a big difference between predicting wins for President-elect Joe Biden by 9-11 points and the much closer margins that wound up taking place when votes were counted.
“Errors were widespread and they were pretty substantial,” Franklin said.
Franklin told Friday’s Eggs & Issues audience he suspects the reason is in a comparatively small group of people who support Trump but do not take part in polling at all. He said a group of 4-5 percent of Trump’s overall support would be enough to cause misses in the range that happened this year. And, if those people are more committed to voting for Trump than for other Republicans, that would also explain why polls in other elections were considerably more accurate.
Franklin said polling organizations can make adjustments when they know they have a small sample size from a given group. That’s something they’ve dealt with for decades. When a subset within the voting population is entirely absent, though, it’s much harder to correct for.
That raises a question. Will the people Franklin suspects are missing from the polls remain engaged and willing to support Republican candidates when Trump himself is not on the ballot? If they do, the inaccuracies may remain. If not, a return to more accurate polling could be coming.
“We’re getting elections without Trump on the ballot and we’re getting Democratic candidates,” Franklin said. “But we’re still badly underestimating Donald Trump.”
During the Eggs & Issues series discussion, presented by the Eau Claire Area Chamber of Commerce, Franklin also touched on the redistricting that will follow the 2020 census results.
When one attendee in the online event asked about gerrymandering and the balance of party control in the Wisconsin Legislature, Franklin urged caution. He noted most seats are won with as much as 60 percent of the vote, regardless of which party wins.
“As a result we have a lot of seats that are not contested at all or are contested very weakly,” he said.
In districts with such significant margins, Franklin said the result is clearly the product of a genuine preference on the part of most voters within the district. Gerrymandering isn’t necessary when one side has that kind of dominance.
Where it could make a difference is on the fringe, where neither party has such a clear-cut advantage. Franklin suggested such areas are where the effect of controlling redistricting is strongest.
And, he said, drawing lines to make one district safer for a party can have the effect of costing margins elsewhere.
Franklin used Wisconsin’s 3rd and 7th Congressional Districts as examples. He said Republicans used the redistricting after the 2010 census to tighten the grip on the 7th District.
In November, Democrat Ron Kind won his 13th term in the district, defeating Republican Derrick Van Orden. But the margin was considerably closer than previous elections, and Franklin said he wonders whether Republicans in the Legislature might consider shifting some areas back into the 3rd District to put additional pressure on the district in 2022.