JMIR Publications

ABSTRACT

Background:

Reliable measurement of daily smartphone screen time is critical for research examining associations with health outcomes. While objective tools overcome biases inherent in self-reported methods, it remains unclear how many days of monitoring are needed to reliably estimate typical use.

Objective:

This study aimed to determine the minimum number of days, within a 7-day measurement period, required to achieve reliable estimates of weekly smartphone screen time using two objective methods.

Methods:

We analyzed 7-day smartphone screen time data collected via passive sensing on Android devices (Android Chronicle, n = 71) and daily screenshots on iOS devices (iOS Screenshot, n = 64). We computed one‐way random‐effects, single‐measure ICCs (ICC₁,₁) with 95% bootstrap confidence intervals (CI; 500 resamples) and applied the Spearman–Brown prophecy formula to project the number of days needed to achieve a target reliability of ≥0.80. Projections were conservatively based on the criterion that the lower bound of the ICC’s 95% CI must exceed this threshold.

Results:

Single‐day ICC₁,₁ values were 0.66 (95% CI: 0.58 to 0.75) for Chronicle and 0.72 (95% CI: 0.63 to 0.79) for screenshots. Using the conservative lower-bound criterion to ensure 95% confidence in achieving reliability ≥0.80, passive sensing from Android Chronicle users required four days, and screenshots from iOS users required three days of monitoring. Average daily screen time was lower on weekends than on weekdays.

Conclusions:

To reliably measure daily smartphone screen time duration (ICC ≥ 0.80), a monitoring period of three to four days is required. Due to screen time differing on weekends, we recommend including at least one weekend day. In a rapidly evolving field with diverse, often proprietary measurement tools, these conservative benchmarks, and transparent processing pipelines, are helpful for producing reproducible, comparable smartphone-behavior data.