BACKGROUND: Surveillance of traumatic brain injury (TBI) in the USA has historically relied on healthcare administrative datasets, but these methods likely undercount the true burden of TBI. Self-report measures may help obtain more comprehensive estimates. A fundamental question is whether TBI prevalence estimated through self-report varies by whether and how TBI signs and symptoms are ascertained in the survey question(s). METHODS: Self-report data were collected from 4303 adults in the summer wave of Porter Novelli's 2023 ConsumerStyles survey. Respondents were randomised to receive a question about their 12-month experience of head injury and a grouped list of TBI signs/symptoms or a question about their 12-month experience of head injury and an option to select individual TBI signs and symptoms that resulted from the head injury. All respondents who indicated they experienced a head injury received a set of follow-up questions. RESULTS: A significantly higher percentage of people who received the 'individual signs/symptoms' question self-reported a TBI (8.3%) than those who received the 'grouped signs/symptoms' question (4.3%) χ(2)=19.6, p<0.0001. Among TBI cases identified, there were no statistical differences between question type in relation to whether the respondent reported being evaluated for or diagnosed with a TBI, nor any demographic differences. CONCLUSIONS: These results demonstrate that how TBI is ascertained in a survey can meaningfully impact reporting of TBI prevalence and provides evidence that question wording and structure affect estimates on national surveys.