As the two candidates for the Wisconsin governor race faced off on television Wednesday night, Democrat David Crowley and Republican Tom Tiffany criticized each other’s records and laid out their governing priorities.
We’ve already recapped key takeaways. Here, we use data to examine how the candidates spent their time, which topics drew the most attention and how they criticized each other.
School funding dominated the debate
Each candidate spent over four minutes explaining his take on school funding, including positions on private school vouchers, making it the debate’s most covered topic. Taxes ranked second, with candidates covering several issues: income taxes, state budget surpluses and property taxes.
The two candidates spent roughly the same amount of time on most topics. One notable exception came when Tiffany spent three minutes and 37 seconds — two minutes more than Crowley — explaining why he voted against certifying the 2020 presidential election results.
“One of the reasons that I did that is Zuckerbucks,” Tiffany said. “If you remember, Mark Zuckerberg, the Facebook billionaire, he put $15 million into Wisconsin, and so he was trying to privatize our elections via electioneering. The people of the state of Wisconsin rejected that a few years later.”
“Zuckerbucks” refers to grants funded by Facebook founder Mark Zuckerberg and his wife, Priscilla Chan, that helped local election officials run elections during the COVID-19 pandemic. Wisconsin voters approved a constitutional amendment banning private funding for election administration in April 2024, after Republicans argued the grants disproportionately benefited Democratic strongholds.
Crowley spent more time attacking Tiffany
Although Tiffany spoke for about 200 seconds longer overall, he spent much of that time explaining and defending his own positions.
Crowley, by contrast, took a more aggressive approach, spending about four minutes more than Tiffany in directly criticizing his opponent.
For this analysis, we classified a statement as an attack when a candidate criticized his opponent or made him the subject of a statement.
Crowley spent the longest time criticizing Tiffany’s positions on taxation, school funding and border control.
“We’re talking about somebody who actually supported the elimination of the United States Department of Education. We’re talking about someone who, when he served as a senator, voted for one of the largest cuts to our UW systems as well as K-12 through education,” Crowley said. “I went to Milwaukee Public Schools. My daughters go to public schools, and I know that they need an increase in funding from the state of Wisconsin because of the fact that we have been dealing with decades of cuts ever since he was in the Legislature.”
Related Stories
Tiffany spent the most time criticizing Crowley when discussing how Crowley would appeal to those who voted for democratic socialist Francesca Hong in the Democratic primary.
“What he said about a week or two before the primaries, he said there was very little difference between he and Francesca Hong,” Tiffany said. “In fact, one of the only things he could think of is that he was supportive of data centers.”
The candidates spent similar amounts of time criticizing one another’s records on data centers — less than a minute each.The candidates spent roughly 50 seconds apiece criticizing their opponents’ record on data centers.
Crowley accused Tiffany of supporting tax breaks for data center operators. (Tiffany voted for President Donald Trump’s tax and spending package that included broad tax breaks benefiting data center developers along with many other businesses, but lawmakers did not specifically vote on a data center tax break.)
Tiffany countered by criticizing Crowley’s past support for subsidies and calling Crowley’s proposal to power data centers with wind and solar energy unrealistic.
Word clouds show how the candidates framed their attacks
After we removed verbal fillers and words used to address the other candidate, the word cloud of attacks shows that Tiffany used the word “taxes” the most often when criticizing Crowley. This reflected Tiffany’s argument that Crowley’s policies would increase tax burdens on Wisconsinites.
“Taxes have went up every single year that David Crowley has been county executive,” Tiffany said. “You know what he’s going to do as governor.”
Crowley most often used the word “work” when attacking Tiffany, repeatedly questioning Tiffany’s track record in following Trump’s decisions and whether Tiffany would represent Wisconsinites if elected.
“Congressman Tom Tiffany, it doesn’t sound like you want to work for the small guys,” Crowley said after Tiffany said he was open to raising income limits on private school vouchers. “It sounds like you want to work for those who have the money and have the means to continue to pay their way. This is what this election is about.”
The pattern was less pronounced in the word cloud analyzing defensive statements. Tiffany most often said the word “Wisconsin” while Crowley’s most frequently used word was again “work.”
How we analyzed the debate
Wisconsin Watch transcribed the gubernatorial debate audio through Parakeet MLX, an automatic speech recognition model, then used the artificial intelligence tool Pyannote to assign each line in the transcript to a speaker. Automated transcription and speaker recognition can produce errors, which may affect the results.
We manually reviewed the transcript to correct as many speaker-identification errors as possible. We also tagged each line spoken by the candidates by topic and tone. We categorized a line as “aggressive” if a candidate referred to his opponent as the subject of a sentence or otherwise clearly criticized him. We categorized all other statements as “defensive.”
We calculated speech duration using the transcript’s start and end time stamps.
The SpaCy tokenization algorithm we used to generate the word cloud misidentified “taxes” as “taxis,” so we manually corrected that error.
Brittany Caroloni contributed reporting.
Republish our articles for free, online or in print, under a Creative Commons license.
