JACKSONVILLE, Fla. – A University of North Florida poll of likely Florida midterm voters found that most supported the idea of phasing out property taxes until they learned of potential effects to local budgets.
The Public Opinion Research Lab at the University of North Florida surveyed 848 likely midterm voters from July 8–17.
Recommended Videos
Initially, 61% said they supported a proposed state constitutional amendment that would begin phasing out some homestead property taxes.
But when voters were told about the budget impact, they decisively rejected the proposal.
After respondents were told estimates indicate its passage could create a roughly $11.8 billion municipal budget shortfall over two years, support fell to 45%, and opposition rose to 47%.
“These numbers show how important the final ballot language will be in determining the outcome of this amendment,” said Sean Freeder, faculty director of the UNF Public Opinion Research Lab.
On other issues, Floridians were split about the general increase in use of artificial intelligence over the past year — 49% supported it and 49% opposed it — but opposed building AI data centers near them: 28% said they would support building a data center in their area and 68% opposed it.
The poll also found Republican Byron Donalds leading Democrat David Jolly by five points and Republican U.S. Sen. Ashley Moody holding double‑digit advantages in paired matchups.
When respondents were asked to choose between Donalds and Jolly as the next Florida governor, 46% said they would vote for Donalds and 41% for Jolly; 6% were undecided.
In U.S. Senate matchups, Moody led Alexander Vindman 50% to 40% and led Angie Nixon 50% to 42%.
President Donald Trump’s overall job approval in the poll was 48%, with 51% disapproving; 46% said they strongly disapproved. Respondents rated Trump’s speech, personality and demeanor less favorably: 39% approved and 60% disapproved.
The UNF PORL poll used a mixed‑mode design of telephone interviews and web surveys and weighted results by education, party registration, age, race and ethnicity, sex and geography. The overall margin of sampling error is plus or minus 3.8 percentage points. For full crosstabs and methodology, visit unf.edu/coas/porl.