Objective Smartphone Screen Time Monitoring Among Caregivers of Preschool-Aged Children: Reliability Study of Monitoring Duration

Abstract Background Reliable measurement of smartphone screen time is important for research examining associations between digital media use and health outcomes. Although objective monitoring tools reduce bias associated with self-report, the number of monitoring days required to reliably estimate typical smartphone use remains unclear. This question is particularly important for study design because longer monitoring periods increase participant burden and the likelihood of missing data. Objective This study aimed to determine the minimum number of valid monitoring days within a 7-day measurement period required to reliably estimate average daily smartphone screen time during the same 7-day period among primary caregivers of preschool-aged children using 2 objective methods. Methods This secondary analysis included caregivers participating in 2 observational studies conducted in the Eastern United States. Android users installed Chronicle, a passive sensing application that records raw Android UsageEvents data, which were preprocessed into smartphone session events (pickups where the device was unlocked). iOS users submitted daily screenshots of Apple Screen Time summaries. Reliability analyses included 100 Android Chronicle participants and 129 iOS Screenshot participants with at least 7 consecutive complete days of data. The primary outcome was day-level total smartphone screen time. For Android Chronicle, this was defined as the summed duration of device-unlock sessions across each calendar day, whereas for iOS Screenshot, this was defined as the daily total reported by Apple Screen Time. One-way random-effects intraclass correlation coefficients (ICC [1,1]) with 95% bootstrap CIs (500 resamples) were calculated separately for Android and iOS data. The Spearman-Brown prophecy formula was used to estimate the number of monitoring days required to achieve target reliability thresholds (ICC≥0.70, 0.80, and 0.90). Sensitivity analyses using the combined durations of sessions and glances produced similar conservative monitoring-day recommendations. Results Single-day ICCs were 0.69 (95% CI 0.63-0.76) for Android Chronicle and 0.71 (95% CI 0.65-0.77) for iOS Screenshot data. Point-estimate projections indicated that 2 monitoring days were sufficient to achieve an ICC≥0.80 for both data sources. Using the more conservative criterion that the lower bound of the projected 95% CI exceeded 0.80, 3 monitoring days were required for both Android and iOS data. Screen time was significantly lower on weekends than on weekdays, suggesting that monitoring periods should include at least 1 weekend day. Conclusions Among caregivers of preschool-aged children, relatively short objective monitoring periods may provide reliable estimates of average daily smartphone screen time within a 7-day monitoring period. Two monitoring days were sufficient under point-estimate projections, whereas 3 days were required using a conservative lower-bound criterion. These findings support the feasibility of brief smartphone monitoring protocols for estimating the 7-day average duration, while emphasizing the importance of including weekend representation and avoiding the extension of these estimates to longer-term habitual smartphone behavior or screen time metrics beyond duration without further evidence.