A lot has been written about Florida’s evolution from a razor-thin swing state to a “solidly” red state. Republican voter registration has surged, driven in large part by the post-COVID migration boom.

But headlines only scratch the surface. To understand the underlying mechanics of this political shift, we need to move past the top-line numbers and look at the raw data.

To do that, I extracted three complete, historical Florida voter files (spanning 2018, 2023, and our latest June 2026 snapshot) and began interrogating the data. This post will serve as the master hub for the project. Over the next week, I will update this page daily as we publish the findings for each of our seven core research questions.

Before we begin, A look at the current state

Q:  What is Florida’s Voter Landscape in 2026?

Link:  Florida Voter Registration Snap Shot

Question 1: Voter Removals & Purges

Q:  How many voters have been removed from Florida’s rolls since 2018, and do deletion rates vary significantly across demographic groups and party lines?

Link:  Question 1: Voter Removals & Purges

Question 2: New Voter Additions

Q: How many new voters have been added to the file since 2018, and what do their demographic profiles and party affiliations look like?

Link: Question 2 – Who are the New Voters?

Question 3: Party-Switching Dynamics

Q: How many registered voters changed parties between snapshots, what is the demographic profile of a “switcher,” and did switching behavior change between 2018–2023 vs. 2023–2026?

Link: (coming soon)

Question 4: Demographic Volatility Summary

Q: A comprehensive look at the net impact: How do additions, deletions, and party switches cross-examine by baseline party affiliation and demographic groups?

Link: (coming soon)

Question 5: Active vs. Inactive Electorate (ACT to INA)

Q: An exploration of voters moved to inactive status: Who are the INA voters in each snapshot, and has their profile evolved over time?

Link: (coming soon)

Question 6: Deep Dive into NPA Voters)

Q: Who are Florida’s No Party Affiliation (NPA) voters demographically and geographically (by County, DMA, State House/Senate, and Congressional District), and are they primarily brand-new registrations or party switchers?

Link: (coming soon)

Question 7: First-Time Voters (2020 vs. 2024)

Q: A profile of debut general election voters in the last two presidential cycles: Who are they demographically, how did their party registration shift, and what is their subsequent participation tier?

Link: (coming soon)

Methodology & Data Framework

To maintain analytical integrity, this project relies on three complete data extracts from the Florida Division of Elections voter file to serve as our structural backbone

Data Cleaning & Variables

To ensure clean comparative analysis across cycles, raw fields were standardized and recoded into the following variables:

• party2 (Party Affiliation): Recoded into DEM (Democrat), REP (Republican), NPA (No Party Affiliation), or OTH (all minor parties collapsed for clarity).
• race2 (Race/Ethnicity): Simplified from the state’s complex 9-code scheme into 7 highly readable, distinct categories.
• age_group (Age Brackets): Binned into four analytical cohorts: 18-34, 35-49, 50-64, and 65_plus.
• yr_age (Current Age): Calculated precisely based on the voter’s birthdate at the exact time of the file snapshot.
• yr_reg (Registration Longevity): Calculated number of years a voter has been active on the file at the time of the snapshot.
• fl_dma (Media Market): Florida Designated Market Areas (DMAs) mapped and joined via a county lookup table.
• pri_score & gen_score (Turnout Metrics): Primary and General election turnout scores, calculated as a ratio of actual votes cast to total elections eligible (scaled from 0.00 to 1.00).

Understanding the Turnout ScoresOur primary and general scores measure eligibility-adjusted participation rates, not raw vote history.

Score = Times Voted / Elections Eligible to Vote In

By calculating participation this way, a voter who registered after a specific primary deadline is not penalized as “inactive” for an election they legally could not vote in.  This normalization prevents us from artificially dragging down the scores of recently registered voters, making participation data apples-to-apples across different registration vintages.

(Note: Score ranges operate strictly between 0.00 (never voted) and 1.00 (voted in every eligible election).  Minor anomalies >1.00 represent rare data edge-cases, such as a voter moving counties and updating their registration file mid-cycle after casting a ballot).