Both Dems and GOPs are fired up in Wisconsin for 2020. So why are polls counting more GOPs?

Craig Gilbert in the Milwaukee Journal-Sentinel gave his rundown of the primary election results for the Wisconsin Supreme Court, and focused on the fact that there were a lot more votes than we are used to seeing for this time of the year.

Wisconsin’s February election was the first hard indicator of what many people in both parties have long suspected — that this could be a year of unusually high voter turnout.

More than 700,000 votes were cast in Tuesday’s state Supreme Court primary, an exceptional total for a nonpartisan statewide February race. It’s easily the highest turnout of the seven Supreme Court primaries that have occurred since the year 2000….

The six other Supreme Court primaries that have been held since 2000 generated an average of a little over 400,000 votes, with highs of 567,038 votes in 2016 and 534,980 in 2018.

The top finisher in Tuesday’s court primary, conservative candidate Daniel Kelly, got more than 350,000 votes, according to unofficial returns. In the six previous court primaries, no single candidate received more than 253,000 votes.

Tuesday’s second-place finisher, liberal candidate Jill Karofsky, also topped that previous high, attracting more than 260,000 votes. Another liberal candidate, Ed Fallone, received about 89,000 votes.

Some of this higher turnout reflected the contested local races in both the city of Milwaukee and Milwaukee County as a whole, but that was counteracted with the competitive Republican primary in Wisconsin’s 7th Congressional district, as that WI-7 race drove a massive jump in votes from Northern Wisconsin.

The last time Wisconsin held a February primary in 2018, only 3,363 people voted in Oneida County.

This year, 9,469 people cast ballots in state and local primary races marking a more than 280% increase.

Hartman said the increase in votes is all about who's on the ballot.

"We were expecting an increase with the Congressional seat primary on there, we knew the voter turnout would be better. But this, I think, was higher than what any of the clerks were expecting," said [Oneida County Clerk Tracy] Hartman….

Marathon County also saw high voter turnout Tuesday with 23,908 ballots cast. That's a more than 252% increase from 2018.

If you want an indicator of Dem engagement, note that Dane County had almost no local races of note, but still had more than 87,000 votes on Tuesday vs slightly more than 77,000 in 2018, and 80% of those votes went to the more liberal candidates. So if anything, Dems are even more engaged than they were at this point in 2018.

And yet, despite the 50-50 nature of the vote totals and the inflated turnout in Northern Wisconsin, recent polls in Wisconsin only seem to account for this high voter turnout among Republicans, and not among Democrats. This has been reflected in a consistent slant toward the GOP in the Marquette Law School Poll, which had an electorate of GOP +4 in its most recent release.

It was even worse in Quinnipiac University’s “swing state” poll of Wisconsin. I’ll use the closest matchups of Joe Biden and Bernie Sanders vs Donald Trump as the examples here, although all the cross-tabs with all Dem candidates will tell a similar story.

Quinnipiac Poll of Wisconsin, president. Feb 2020
Dems- Biden +91 vs Trump, Sanders +88 vs Trump
GOPs- Trump +90 vs Biden, Trump +91 vs Sanders
Independents – Biden +1, Sanders Even

Seems like it would be a close race in a 50-50 state, right? Nope.

Trump defeats Senator Bernie Sanders 50 - 43 percent;

Trump is ahead of former Vice President Joe Biden 49 - 42 percent.

Why? Here’s why.

party breakdowns in Quinnipiac samples are:
plus 6 Rep in WI
plus 3 Dem in MI
plus 5 Dem in PA, which would go a long way in explaining the divergent results of the 3 states ...

— Craig Gilbert (@WisVoter) February 20, 2020

Former Russ Feingold press secretary and Progressive writer Jud Lounsbury points out that this is a simple flaw of doing polls via phone in 2020.

It's why the oldest age group is very consistent from poll to poll and the youngest age group varies wildly from poll to poll. It's because the ACTUAL sample size of the older group is very large and the ACTUAL sample size of the youngest group is very small.

— Jud Lounsbury (@JudLounsbury) February 21, 2020

Then add in the fact that many younger voters aren’t registered until Election Day in Wisconsin, and that many highly educated, mobile individuals that moved to Wisconsin still have phone numbers from other states, and you have the makings for garbage inputs that don't reflect who will actually be voting in both April and November.

I don't understand why pollsters are allowing themselves to publicly release these things when prior and current events indicate there isn't going to be a pro-GOP electorate in Wisconsin for 2020. I understand that it's hard to control who picks up the phone from weird phone numbers these days, but when you're not adequately adjusting your sample to match reality, you're going to have a faulty product.

Which I'd blow off, except that lazy media reports these faulty polls as fact, and lazy low-info voters are influenced by these polls, as they like being associated with a winner. And that's why we have to call out any poll that is so badly slanted, and put public pressure on them to correct themselves as the election gets closer.