How has the percentage of former Trump voters expressing regret changed since 2020?
Executive summary
The share of Trump voters who say they regret their vote has risen from very low levels in the immediate post-2016/2020 period to a larger—but still minority—slice by 2025, with most national polls in 2025–2026 putting personal regret in the single digits to low teens while a wider group expresses disappointment or “mixed” feelings [1] [2] [3] [4]. Differences in question wording, timing and sample mean the trend is better read as a modest increase in expressed regret and unease rather than a mass conversion away from Trump [4] [2] [3].
1. The baseline: historically low explicit regret immediately after early Trump victories
Early empirical snapshots showed only a sliver of Trump voters saying they would change their vote: one post‑2016 poll reported essentially 0–3 percent of Trump voters saying they would have voted differently shortly after the election [1], establishing a low baseline for explicit ballot‑regret that carried into some early polls of the 2024/2025 era.
2. Polls in 2025 show a clear, measurable uptick but still a minority
Multiple 2025 polls found higher, though still modest, shares of Trump voters expressing regret or second thoughts: Navigator Research reported about 11 percent of 2024 Trump voters said they regretted their vote, with another 16 percent “disappointed” [3], while a UMass poll found roughly 2 percent said they would have voted differently and 26 percent had “some concerns” (with 2% explicit regret noted in the UMass writeup) [2] [5].
3. Aggregate analyses describe a small upward trend, not a wave of remorse
Synthesis pieces conclude the “regret” cohort has grown slightly over time but remains a minority: CNN’s analysis described about one in ten Trump voters by mid‑2025 reporting regret, mixed feelings, or wishing they hadn’t voted, and noted that this group had edged up since April of that year [4]. Reuters’ qualitative reporting from interviews with a set of Trump voters found some individuals expressing regret but emphasized most still backed Trump on key issues even if unhappy about others [6].
4. Variation by subgroup and question wording matters for the trend
Different polls measure different things: some ask about firm “would vote differently” regret (UMass/Independent showing ~2%) while others include “mixed feelings” or “disappointed” as partial regret (Navigator/Yahoo/YouGov measures yielding ~11% + 16% disappointed), producing divergent headline rates [2] [5] [3] [4]. Cross‑tabs also show demographic splits—one fall poll reported higher regret among non‑white Trump backers than white backers (19% vs. 5% in that poll) — underscoring that the increase is not uniform across the coalition [7].
5. Anecdotes and reporting amplify perception but don’t substitute for representative trends
A steady stream of human‑interest stories and compilations of “regretful Trump voters” in outlets like BuzzFeed, Daily Beast and other outlets document vocal individuals and local shifts—useful color that signals discontent in pockets but not proof of a national sea change by itself [8] [9] [10] [7]. Journalistic samples and small qualitative studies (e.g., Reuters’ interviews of a subset of voters) capture intensity and reasons but are explicitly not statistically representative, a limitation reporters acknowledge [6].
6. Bottom line and limits in the record
Measured strictly, explicit “would vote differently” regret was very small around earlier benchmarks and rose modestly into the single digits-to‑low teens by 2025 when broader definitions (including “disappointed” or “mixed feelings”) are counted; when polling asks the narrow counterfactual of changing a ballot the figures remain in the low single digits in some polls [1] [2] [3] [4]. The record shows a consistent pattern—more unease and some increased regret since 2020—but exact magnitudes depend heavily on poll wording, timing, and subgroup differences, and the available sources do not provide a single unified year‑by‑year time series from 2020 through 2025 to produce a precise numeric trajectory.