54% of US parents said they felt they had to give their children social media because so many other families already had, according to a Harris Poll of 1,013 US parents conducted for Let Grow in February 2025. That figure measures felt social compulsion rather than retrospective regret — the survey’s nearest direct “wish they hadn’t” item is a separate 39% figure about smartphones, not social-media accounts, and it too is framed as parents who “felt that they had to give in.” This entry covers the social-media-account decision specifically; the companion entry giving-child-smartphone-early-vs-waiting covers the device-provision decision.
The teen and parent perspectives both lean negative. Pew Research’s April 2025 report found that 45% of teens say they spend too much time on social media, up from 36% in 2022. In a separate April 2026 Pew report, parents were more likely to think social media hurt rather than help their teen’s mental health (24% vs 8%), and roughly four-in-ten said it hurt their teen’s sleep and productivity. In general, parents lean more negative than positive about social media’s impact on their teen’s lives — but these are impact judgments, not measured decision-regret about granting access.
Inaction regret — regretting the choice to delay or block social media account access — is harder to measure because no survey has directly asked parents who delayed access whether they regret it. The estimated 10% proxy is an upper bound derived from the same Pew finding that parents lean net-negative on social media (24% say it hurt their teen’s mental health vs only 8% who say it helped), which implies that the large majority of parents who chose to delay are unlikely to second-guess that choice. The American Psychological Association’s 2023 Health Advisory frames potential risks as greater in early adolescence (ages 10-14) and advises adult monitoring of younger teens’ use, reinforcing the inference that inaction regret is uncommon in this domain. The 44-point gap between action (54%) and inaction (10%) is directionally stable across the available evidence, but both figures carry construct caveats — the action figure is a felt-compulsion measure and the inaction figure is a proxy — and both are sensitive to the child’s age, the platforms involved, and the degree of supervision that accompanied any access granted.