33% of adults who hold grudges say they regret doing so, per Trustpilot’s Helping Hands survey of 12,000 adults across six countries. By contrast, roughly 10% of Americans who chose to forgive someone report negative outcomes from that decision — feeling exploited, seeing the offender repeat the behavior, or concluding the forgiveness was premature. That figure is inferred from the University of Michigan’s nationally representative survey of 1,423 adults (Toussaint & Williams, 2001), which described a “small minority” experiencing negative post-forgiveness outcomes but did not publish an exact percentage. The same survey found 70% “acknowledge” that grudge-holding harms their health, but acknowledgment is not regret — the 33% who explicitly reported regretting their grudges is the more appropriate comparison point.
The physiological case is unusually direct. Witvliet, Ludwig, and Vander Laan (2001) measured 71 participants while they alternated between grudge-rehearsal and forgiveness imagery. Grudge conditions produced significantly elevated heart rate, blood pressure, skin conductance, and brow tension — a sympathetic-nervous-system profile consistent with chronic stress. Rasmussen et al. (2019) confirmed the pattern at scale: across 103 samples totaling 26,043 participants, forgiveness was reliably associated with better psychological and cardiovascular health, though the correlation was modest (r roughly 0.14 to 0.18) and causality remains unestablished.
The main caveat is measurement asymmetry. No major survey has directly asked “Do you regret forgiving?” — the 10% is inferred from negative post-forgiveness outcomes described qualitatively, not a regret prompt. The 33% inaction-regret figure is a direct regret question from a consumer panel, not a clinical cohort, and a third of its grudge-holders could not even recall why they were still angry. The directional finding — that grudge-holding generates more regret than forgiving does — is consistent across multiple literatures. The precise magnitude is less certain.