Understanding the measurement error introduced by self-reports of smartphone use is essential for making meaningful inferences about smartphone use and its effects. Evidence for the self-reports of smartphone use in repeated measurement designs is largely missing. Based on self- reported and digital trace data from 137 Czech adolescents (41% girls, Mage = 14.95 years), this study examined the accuracy, directional bias, and convergent validity of daily end-of-day and single-time self-reports of screen time and phone-checking behaviour. Overall, the study found considerable discrepancies between self-reported smartphone use and digital trace and low between-person convergent validity for all self-reports considered for the study, which indicates that self-reports are poor measures of smartphone use. Respondents usually underestimated their screen time and phone-checking behaviour in daily and single-time self-reports. The within-person convergent validity of daily reports was low, which shows their poor ability to capture the actual day-to-day fluctuations in smartphone use. The current study shows that the self-reports of smartphone use introduce considerable measurement error. The study reinforces the evidence that conclusions about smartphone use and its effects based on self-reported data should be taken with caution and extends it to daily reports typically used in repeated measurement studies. When possible, researchers should avoid self-reports of smartphone use, and collect digital trace data instead.