Analysis of 2023 Chicago Mayoral polling vis a vis election results
The polling in the 2023 Chicago Mayoral election was not “off.” It was spot-on.
Some “experts”claim the Mayoral polling was “off.” It wasn’t. Polls are not predictions. It’s a common misconception even among those who know better. Turnout matters. Weather matters. Last minute messaging matters. A candidate’s “ground game” matters, including their skill at utilizing every aspect of elections, including early voting, ballot harvesting, a mail-in ballot program, and an effective election day field organization. All are difficult to poll because they happen post-poll.
Polls are not done in a vacuum. Time is not frozen when a poll is done. Campaigns react to their own poll results, to bolster the positive aspects of their campaign and correct the areas of their campaign that need improvement. It is, in fact, the entire reason political polls are conducted.
Before we dive into the numbers, let’s stipulate that every published poll in the Mayor’s race showed the race within the margin of error, which led every media outlet to report the race was “too close to call.” Because it was. No published poll “predicted” a Vallas victory, or even showed Vallas at 50%. I was quite explicit in saying this every time we released a poll.
Victory Research’s final poll showed a narrow margin of 4.2% for Paul Vallas (49.6%-45.4%), within the margin of error. WGN’s poll, released eight days prior to the election, showed a wider Vallas lead of 5.7% (46.3%-40.6%), also within the margin of error.
What could happen in the final days of an election to change a 49.6%-45.4% or 46.3%-40.6% race to a 52-48% margin the other way? In multiple media interviews, I outlined four possible paths to victory for Johnson:
Larger than usual turnout of early voters, among which he was leading. In our poll, Johnson led among people who voted early 52%-48%. The day after our poll came out, “The city of Chicago set a new record for early ballots cast in a single day, as voters headed to the polls with the possibility of severe weather looming on Election Day,” according to NBC Chicago.
As WGN put it when discussing their poll, “To make up ground, Johnson would likely need a surge of younger voters casting their ballots before polls close on April 4” According to data from the Chicago Board of Elections, there was an increase of more than 24,000 voters under the age of 35 years old from February to April. Our poll showed Johnson leading voters from 18-30 years old with 64.5%. The WGN poll had Johnson leading among voters under 35 years of age with 58.2%. Johnson’s margin currently in raw votes is 26,448. Using WGN’s numbers, 24,000 new voters under 35 years of age would produce nearly 14,168 additional more votes for Johnson, or more than half his current margin. Using VR’s numbers, 24,000 new voters under 35 years of age would net Johnson 15,702 more votes.
Johnson needed to roll up big margins among Black voters. On WLS AM 890, I said he needed a “tidal wave of votes” in Black wards. In the WGN poll, Blacks favored Johnson 55.4%-24.6%. Actual results show that in the 15 wards with 55% or more Black population, Johnson won 76.6%-23.4%, or as the Johnson campaign said the morning after the election, “We outperformed on the South Side in terms of the margin. It could be the case that some of the voters on the South Side who felt like they couldn’t vote for us didn’t vote at all.” There were more than 175,000 voters from Black majority wards. Using the WGN poll numbers, the raw vote total difference between 55.4% and 76.6% is 37,247 votes, remembering again Johnson’s margin stands now at 26, 448. VR’s final poll had Johnson winning Black voters with 78.2% of the vote (as stated he got 76.6% in Black majority wards).
As world-renowned political analyst Tom Bowen, the architect of spectacularly well run campaigns such as Sam Royko for 1st Ward Alderman, and Eileen Dordek for State Representative, said on April 2nd (two days before the election) when analyzing the WGN poll, “You could have quite a few voters change their mind in the closing week.” In the WGN poll, there were 13% undecided. In VR’s final poll, 5% were undecided, with an additional 7% saying they could still change their mind. While less so than earlier in the campaign, with that many voters not locked into voting for one candidate or the other, the race remained volatile.
One media critic of Victory Research polling, tweeted that Victory Research polling had other goals “ than accuracy.” This reporter works at WGN, whose poll showed a larger margin for Vallas than Victory Research’s poll. Their poll, including “leaners” said “Vallas’ lead increased to six points, 53% to 47%.” But….but….but…accuracy. Can’t make this stuff up. Makes one wonder if he started drinking again.
Every knowledgeable analyst in town correctly praised the “ground game” of the Johnson campaign, as I did. One Johnson campaign leader, Alice Yin, tweeted about the “vast ground game operation for Brandon Johnson.” The Tribune called it “a formidable ground game.” Politico used the same description. Here is how the Sun-Times’ Fran Spielman described it, “Johnson has an ace in the hole: a 1,000-person-strong ground game, using CTU and SEIU members and volunteers from United Working Families. They’ve already knocked on over 200,000 doors and made thousands of calls and text messages and will be a get-out-the-vote army on Election Day.”
Johnson’s campaign manager Jason Lee accurately described the effect “You can pick up to 2 to 4 percentage points with a robust ground game,” Lee said. “That could make the difference.”
So, everyone agrees. The Johnson ground game was “robust” and “formidable.” An “expert” can’t have it both ways. If the polls were “off” and Johnson was leading going into election day, then he didn’t really need or have a “robust ground game.” Considering his final margin is just over 4%, under the theory that Johnson was already leading, his ground game was ineffective. Nobody believes that.
As mentioned, the WGN poll, with leaners, said Vallas’ lead was 53-47. For the sake of this analysis, let’s assume the 5% undecided in VR’s poll split down the middle. That would make the Vallas lead 53.1%-46.9%. As noted, the Johnson campaign estimated their ground game could make a 2-4% difference in the outcome. Let’s split the difference and say that the ground game made a 3% difference in the outcome. That changes the WGN poll to 50-50 between Vallas and Johnson and the VR poll to Vallas 50.1%-49.9%. If the ground game made a four percent difference, then the WGN polls becomes 51-49 Johnson and the VR poll 50.9%-49.1% for Johnson. His current margin according to the Chicago Board of Elections is 52.2%-47.8%. As I said, the polling was “spot-on.”
Geographically, VR’s poll showed Vallas getting 71.0% on the Northwest side. In the white ethnic wards of the Northwest Side, his current total is 71.2%. We had Vallas at 50.8% on the North Lakefront. His current total is 50.9%. We had him among Latino voters at 50.6%. In wards with more than 30% Latino population, his current total is 52.1%. We had Johnson at 78.2% among Black voters. In the 15 Black majority wards, his current total is 76.6%. Doesn’t seem “off.”
While we didn’t have access to the final internal tracking polls of either candidate, judging from their public comments, it seems that both campaigns also believed it was “too close to call” going into election day. Vallas campaign insider former 44th Ward Alderperson Tom Tunney said on election night “We knew it was going to be a nail-biter,” when Vallas was still leading by 1,000 votes early in the vote counting.
The morning after the election, Johnson campaign manager Lee said “We outperformed on the lakefront. The turnout was good. We outperformed on the South Side in terms of the margin. It could be the case that some of the voters on the South Side who felt like they couldn’t vote for us didn’t vote at all.” The question is, “Outperfomed” what? Their own polling obviously. A campaign as well-run as Johnson’s would not have relied on outside polls to plan their last-minute strategy. Their polling also clearly showed the race “too close to call.” The difference in the two campaigns’ closing strategy is the Johnson did a more effective job bolstering where all the polling said they needed help-among Latino voters, the north lakefront and the South Side.
The polling was spot-on