This study used accelerometers to examine SB patterns of US children during different time periods throughout the school day and outside of school hours. Our findings revealed that the children in this study participated in significantly more SB outside of school hours versus during school hours, as well as on weekend days compared to weekdays. Girls participated in significantly more SB on weekdays than did boys. Importantly, all of the children in this study, both boys and girls, as well as children from both low and high SES schools, demonstrated significant increases in SB on weekend days and outside of school hours compared to weekdays and during school hours. Finally, the majority of children were classified as participating both in high amounts of SB and high amounts of MVPA (sedentary/active).
Our finding that girls spent a greater proportion of their weekday time in SB than boys is consistent with recent studies suggesting that boys spent slightly less time in SB than girls [19, 20, 22, 27]. Additionally, the finding that our children participated in significantly more SB outside of school hours was consistent with findings from Steele et al. . However, Fairclough et al. found that differences in SB from weekdays to weekends were dependent on the child’s sex and habitual level of activity . Specifically, Fairclough et al. found that SB increased on the weekends in all boys except those in the highest PA quartile, who demonstrated minimal variation in SB across weekdays versus weekends . In contrast to our findings, girls in the middle two quartiles of PA participated in greater amounts of SB on weekdays compared to weekends, with girls in the highest PA quartile participating in similar amounts of SB on both weekdays and weekends . We used quartiles to examine our data (data not shown) in order to compare our findings to those of Fairclough et al. and found that differences still existed when using quartiles of PA. It is possible that the differences are due to the populations studied, as our children were from the U.S. and Fairclough et al.’s children were from England. Nonetheless, additional research is required to better understand children’s SB patterns on weekdays versus weekends, with a particular need for examining sex differences in SB patterns during these time periods.
In addition to examining differences between weekdays and weekend days, the use of accelerometers allowed us to classify the proportion of time children spent in SB during distinct time periods throughout the weekdays. Our results showed that children achieved their lowest percent of time spent in SB (21.7 %) during the recess/lunch time period. While we were unable to identify other studies examining SB patterns during specific time periods throughout the weekdays, this finding was not surprising, as other studies have demonstrated that children spend a large percentage of their recess time in MVPA [29, 30]. As such, one would expect correspondingly lower levels of SB. Our data also revealed that children only spent 36.7 % of their time at school in SB, while they spent a significantly greater proportion of their before and after school time in SB (60.4 and 45.8 %; p < 0.001). These findings imply that interventions aiming to decrease SB should consider focusing on before and after school time periods, as well as weekends, since these time periods appear to have the greatest room for changing SB.
In conjunction with our findings that SB increases outside of school hours, we also found that increasing SB corresponded with decreasing light PA. Specifically, children in our study spent a significantly greater proportion of their time in light PA on weekdays compared to weekend days (48.6 and 44.3 %, respectively), while time spent in MVPA varied little on weekdays versus weekend days (8.0 and 7.7 %, respectively). These findings were true regardless of sex or SES, with the exception of low SES children, who spent a significantly greater proportion of their time in MVPA on weekdays compared to weekend days (8.0 % vs. 6.3 %, respectively, Wilcoxon ranks test: p = 0.005). When examined as a whole these findings indicate that children shifted away from light PA during weekdays towards greater SB on the weekends. This shift from light PA to SB is concerning due to growing evidence that the health consequences of SB, including a less healthy cardiometabolic profile, are independent from participation in MVPA [1, 3, 4, 13], and due to the fact that light PA is independently associated with cardiometabolic health benefits [14–18].
The shift from light PA to SB is also concerning considering recent research indicating that children do not compensate for SB by increasing PA levels , which contrasts with the previously described ‘activitystat’ idea [10–12]. Indeed, our findings also contrast with the ‘activitystat’ idea, as our children did not compensate for increased SB on the weekends by increasing either light PA or MVPA. These contrasting findings could be due to methodological differences regarding how Wilkin et al. (2006) and Fremeaux et al. (2011) categorized their accelerometer data. Wilkin et al. categorized their data as ‘low’, ‘medium’ and ‘high’ intensity , while Fremeaux et al. categorized their data as total physical activity or MVPA . In both cases the researchers neglected to identify SB separately from light PA. Combining SB with light PA may result in inaccurately representing SB as light PA, thus preventing a clear understanding of children’s activity behaviors at the low end of the activity spectrum (e.g., SB and light PA). This also ignores the differing health consequences of SB versus light PA. The lack of a distinct SB category prevents a direct comparison with our data. Regardless, our finding that children shift away from light PA and towards SB (while maintaining MVPA levels) suggests that perhaps, instead of targeting MVPA, interventions should potentially focus on decreasing SB by shifting towards light PA, as suggested by other researchers as well [6, 14–18, 20, 31].
Additional data from our study support the notion that, instead of targeting MVPA, future interventions should consider focusing on replacing SB with light PA. We categorized children based on the amount of time spent in both SB and MVPA. Children were classified as either ‘sedentary’ (≥360 min/day in SB) or ‘non-sedentary’ (<360 min/day in SB) and as either ‘active’ (≥60 min/day in MVPA) or ‘inactive’ (<60 min/day in MVPA). We found that the majority of children (65.3 %) were classified as sedentary/active, 25.3 % of children were classified as sedentary/inactive and the remaining 9.4 % of children classified as non-sedentary/active. These findings indicate that, while a large proportion of children were achieving recommended amounts of MVPA , they also spent a significant amount of their day in SB. Our findings contrast somewhat with those of Herman et al. . In their study, a smaller percentage of boys and girls were classified as active/sedentary (27.6 and 5.2 %, respectively), with a slightly larger percentage of boys and girls classified as inactive/sedentary (34.3 and 38.7 %, respectively) and active/non-sedentary (18.5 and 10.1 %, respectively) . These differences could be due to the classification criteria, as Herman et al. used screen time to classify sedentary behavior (>2 h/day = ‘sedentary’, ≤2 h/day = ‘non-sedentary’), while the current study used total minutes spent in SB. Despite these differences, it is concerning that the majority of children in both studies were classified as sedentary. Due to the fact that the health and obesity risks associated with SB exist independently from MVPA [1, 3, 4, 6, 8], greater emphasis should be placed on decreasing children’s SB, as opposed to simply increasing MVPA. The fact that PA interventions in children have been largely unsuccessful in achieving long-term changes in BMI  and PA  also suggests that future interventions should perhaps focus on a potentially more achievable target- transitioning children from SB to light PA, as has been suggested by other researchers [6, 14–18, 20, 31].
There were some limitations to this study. We classified awake time as 6 AM to 11 PM. This may have resulted in misclassifying sleep time as SB, which would result in overestimating SB. Future studies should include records of sleep time to prevent misclassification of sleep time as SB. We used 360 min of SB per day to classify children as either sedentary or non-sedentary based on a review determining that children aged 6–11 years spend an average of six hours per day in SB (Pate 2011), and due to the fact, at the time of data analyses, no other studies we found had dichotomized children based on SB. Additional research is required to determine the best cut-point for identifying children as sedentary or non-sedentary. We did not classify the type of SB in which children participated, which prevented a more specific representation of children’s overall activity behaviors. Some researchers have advocated for the concurrent use of objective measures of SB (accelerometers) with self- or proxy-report tools in order to identify both the type and amount of SB [35–37]. The original purpose of this study did not include specifically classifying the type of SB; however, future studies should include the type of SB in order to better understand the specifics of children’s sedentary behaviors, thereby allowing more targeted interventions [7, 9]. The students in our sample attended a fairly affluent school district in the Denver-metro area; therefore, the generalizability of our findings is limited to this and other similar districts throughout Colorado and the US. Finally, this study should be replicated in a larger sample of children and in a variety of socioeconomic settings. In spite of these limitations, we believe that our findings contribute to the literature because we are one of very few studies that have used accelerometers with short epoch lengths (15-s) to examine SB patterns in US children during specific time periods throughout the school day.