Screen Time Doesn’t Just Steal Exercise Hours, Landmark Study Reveals

Every hour you spend binge-watching television or scrolling through a tablet may not be adding to your overall sitting time in the way scientists have long assumed. Instead, a new analysis suggests, recreational screen time appears to cannibalize other forms of sedentary behavior, quietly reshuffling the composition of a person’s day rather than simply stacking more sitting on top of it. The finding, published in the Journal of Activity, Sedentary and Sleep Behaviors, comes from one of the first studies ever to track screen time and the full suite of daily movement behaviors continuously and objectively, and it challenges a foundational assumption in the science of sedentary behavior.

Recreational sedentary screen time, abbreviated rSST by researchers, refers to discretionary time spent watching television, playing inactive video games, or using computers, smartphones, and tablets for enjoyment rather than work or education. It is the most prevalent form of discretionary sedentary behavior outside of work and sleep, consuming more than four and a half hours per day among American adults. Unlike sitting at a desk or commuting, screen-based leisure has been repeatedly linked to elevated risks of cardiovascular disease, type 2 diabetes, certain cancers, and all-cause mortality. Alarmingly, even adults who exercise at two to three times the levels recommended in current physical activity guidelines show only a partial reduction in the mortality risk associated with five or more hours of daily screen time.

Yet a crucial question has remained unanswered: when screen time goes up, what exactly does it replace? Previous studies typically relied on self-reported screen use and weekly averages, leaving researchers unable to see how behaviors trade off against one another from day to day within the same person. The new study, led by Kristina Hasanaj of Arizona State University and Northwestern University Feinberg School of Medicine, together with colleagues including Matthew P. Buman and Sarah K. Keadle, set out to close that gap using an unusually rigorous measurement strategy drawn from the baseline phase of the StandUPTV randomized trial, an intervention designed to reduce screen time in adults.

The technical approach was distinctive. Ninety-four adults, with an average age of 42.3 years, wore a thigh-mounted activPAL3c accelerometer continuously for seven or more days to capture posture and movement, distinguishing sitting, standing, light physical activity, and moderate-to-vigorous physical activity at the minute level. Sleep was measured with a wrist-worn GENEActiv accelerometer processed through the open-source GGIR software package. Crucially, screen time itself was not self-reported: Wi-Fi-enabled smart plugs monitored the power state of each participant’s primary television, while a pre-loaded tablet tracked app usage, with participants logging or adjusting any screen time that occurred outside the home or on other devices. A minute was classified as recreational sedentary screen time only when two conditions coincided: the accelerometer identified a sedentary posture and the smart plug or tablet identified concurrent recreational screen use.

With these continuous streams merged into a daily 24-hour behavioral profile, the team applied multilevel modeling, a statistical framework capable of separating two distinct kinds of association. Between-person effects describe how a person’s average screen time relates to their average levels of other behaviors compared with other participants. Within-person effects describe how a day on which an individual exceeds their own typical screen time relates to that same day’s other behaviors, controlling for the between-person pattern. This dual decomposition is what allows the analysis to speak to the dynamic, reciprocal nature of daily time use rather than static weekly averages. Models were adjusted for age, sex, body mass index, chronotype, education, and whether the day was a weekday or weekend day, and all coefficients were scaled hourly for interpretability.

The headline result was striking. Each additional hour of average total recreational screen time was associated with roughly 45 minutes less of other sedentary time at the between-person level, and about 43 minutes less at the within-person level, with the relationship running in both directions. Standing time also fell, by around 9 to 10 minutes per hour of screen time, while light physical activity and moderate-to-vigorous activity showed small but statistically significant displacements within individuals, on the order of two minutes and one minute per hour respectively. The reverse associations held as well: each additional hour of other sedentary time displaced total screen time by roughly 36 to 44 minutes, and an hour of moderate-to-vigorous activity displaced screen time by about 50 minutes within individuals.

Timing mattered. When the researchers split screen time into daytime and evening components, the displacement effects grew larger in the evening hours. Each hour of average evening screen time was associated with a remarkable 85-minute reduction in other sedentary time at the between-person level, compared with 66 minutes for daytime screen time. This suggests that evening television viewing and tablet use do not merely add to the day’s sitting total; they elbow out other seated pursuits such as reading, hobbies, and time spent with family and friends. In other words, the composition of sedentary time shifts toward screens rather than the overall volume of sitting simply expanding.

Perhaps the most surprising null result concerned sleep. Despite a substantial body of literature linking screen use to shorter and poorer sleep, the study found no statistically significant associations between any of the screen time variables, whether total, daytime, or evening, and sleep duration on either the subsequent or the previous night. The point estimates were tiny, on the order of one to three minutes per hour of screen time. The authors caution that the sample consisted of adults motivated to reduce their screen time, which may have restricted the variability needed to detect sleep effects, and that sleep duration alone may not capture the quality and continuity disruptions that screen use is thought to produce. Moderator analyses revealed scattered interactions by age, sex, chronotype, and weekday status, hinting that these relationships are not uniform across the population.

The findings carry real implications for how interventions are designed. If screen time displaces other sedentary behaviors rather than adding to total sitting, then simply cutting screen time will not automatically reduce sedentary time overall, because the displaced hours may be backfilled with other forms of sitting. The authors argue that total sedentary time and its specific contexts should be treated as distinct behavioral targets, and that interventions aiming to reduce recreational screen time should simultaneously include strategies to lower total sedentary time, rather than assuming that one goal accomplishes the other. This reframing matters for public health guidance, which has traditionally treated sedentary behavior as a single undifferentiated exposure.

The study is not without limitations. The sample was predominantly female, non-Hispanic White, and highly educated, and participants were required to report more than three hours of daily screen time at enrollment, limiting generalizability. Smartphone screen time could not be captured directly, so some recreational screen use may be underrepresented, and the smart plug and tablet proxies may imperfectly reflect actual viewing. Still, the combination of continuous device-based posture measurement, objective screen monitoring, and day-level statistical modeling represents a methodological leap forward. As streaming continues to claim an ever-larger share of leisure time, understanding precisely how those hours reshape the rest of the 24-hour day, and designing interventions that account for that reshuffling, may prove essential to turning back the tide of sedentary living.

Subject of Research: Bidirectional associations between recreational sedentary screen time and 24-hour movement behaviors in adults

Article Title: Bidirectional associations of recreational sedentary screen time and 24-h behaviors: a dynamic cross-sectional multilevel model analysis

Article References: Hasanaj, K., Leonard, K. S., Sears, D. D., Yu, F., Petrov, M. E., Keadle, S. K., & Buman, M. P. (2026). Bidirectional associations of recreational sedentary screen time and 24-h behaviors: a dynamic cross-sectional multilevel model analysis. Journal of Activity, Sedentary and Sleep Behaviors, 5(1), Article 4. https://doi.org/10.1186/s44167-026-00096-0

Image Credits: AI Generated

DOI: 10.1186/s44167-026-00096-0

Keywords: screen time, sedentary behavior, physical activity, sleep, television viewing, activPAL, multilevel modeling, StandUPTV trial, 24-hour behaviors, public health, displacement, chronotype