The aim of this research was to describe sedentary behaviours among NZ young people aged 10-18 years. Results from both data sources indicate this group spend much of their waking time in sedentary pursuits, consistent with reports from other developed countries [3, 5, 30, 31].
SST (most commonly television watching) has been reduced successfully in previous interventions, with concurrent improvements in weight indices . SST is thought to have adverse effects on body composition via displacement of physical activity, reduction of metabolic rate, exposure to food advertising and promoting snacking , and adverse psychosocial effects through exposure to violence and adult content . NZ young people used screen technologies for 181 (SE 3.91) minutes per day on average, somewhat lower than an Australian estimate of 230 (SD 114) minutes per day in adolescents aged 9-16 years in a survey that also used the MARCA . However, these estimates are still markedly lower than recent data from the US, which found that 8-18 year olds reported spending 324 (no SD reported) minutes per day in SST as it is defined in this study .
Though SST is an important target for intervention, it should be noted that the daily time reported in NSST was nearly twice that of screen-based activities. NZ young people engaged in non-screen sedentary behaviour for 340 (SE 5.22) minutes per day on average. This estimate is very similar to a MARCA-derived NSST estimate of 345 (SD 105) minutes per day in Australian adolescents aged 9-16 . Some non-screen activities such as school may be considered less discretionary and others such as reading may be considered more socially desirable than screen time. However, it is still possible to intervene to break up extended periods of sedentary time and increase energy expended. Modification of the school environment to allow standing lessons and opportunities for physical activity has been shown to significantly increase movement in children . In adults, breaking up sedentary time has been associated with favourable body composition and metabolic risk, independent of total sedentary time and physical activity level . It may also be appropriate to encourage displacement of sedentary socialising and transport with more active alternatives, as each accounted for approximately 45 minutes of daily activity in this sample.
The accelerometry data indicated that young people spent 420 minutes (SE 4.26) in sedentary behaviour per day. This estimate is broadly consistent with surveys from other developed countries conducted in adolescents using the 100 count per minute cut-point. Estimates for total sedentary time from other countries include 452 (SE 6.0) minutes per day in 12-15 year olds and 482 (SE 4.8) minutes per day in 16-19 year olds in a representative US sample , 512 (SD 54) minutes per day in a longitudinal US sample aged 13 years , 428 (SD 66.4) minutes per day in a longitudinal UK sample aged 11 years  and 496 (SD 80.6) and 471 (SD 84.3) minutes per day in a Spanish boys and girls, respectively, aged 13-16 years . Consistent with the well-documented decline in physical activity from childhood through adolescence to adulthood , the accelerometry data also indicated an increase in sedentary behaviour with age. Children aged 10-14 years engaged in 60 minutes less daily sedentary behaviour than those aged 15-18 years (400 [SE 4.95] minutes or 50% of monitored time versus 460 [SE 7.64] minutes or 58% of monitored time, respectively). This indicates a disturbing trend in health-related behaviour across adolescence. However, contrary with reports from other countries [2, 3], this study found no differences in engagement in total or specific types of sedentary behaviour between overweight and obese participants and their healthy or underweight counterparts.
The results of this study suggest that the most appropriate targets for intervening to reduce screen time in NZ are males, those in their late teens, and those of Asian ethnicity. Although television watching was the predominant screen behaviour, computer and video games were used for approximately 45 minutes per day. Moreover, interventions to reduce NSST in NZ young people may be more appropriate for females, who report greater levels of NSST than males, and those of Asian ethnicity who reported high levels of NSST as well as high levels of SST. For Asian participants, the MARCA data indicated a greater engagement in computer- and school-based sedentary behaviour. This likely reflects an academic rather than a leisure orientation in sedentary behaviours. Culturally-appropriate interventions should consider the nature of sedentary behaviour in this ethnic group.
The specific strengths of the study are the use of a large, representative sample of NZ young people and the use of the survey methodology for analysis. To the best of our knowledge, this is the first time that complementary methodologies (accelerometry and self-report) have been used to describe sedentary behaviour in a sample of children this large. In adults, it is recommended that population-based monitoring of sedentary time include both self-reported and device-based measurement . However, several limitations warrant discussion. Firstly, this was a cross-sectional study, which does not allow for inferences of causality, or capture change across time. Secondly, there was a 45% non-participation rate and it is possible that non-participants differed from participants, though the study population was nationally representative by ethnicity, age and geography. During the original survey parents were approached to recruit their child, thus selection bias was based on parent rather than the young person, which might dilute this effect.
Data from a subset of participants aged 10-18 years were analysed rather than the total sample. Exclusion of those aged 19-24 years was considered the most appropriate approach to provide meaningful comparisons of sedentary behaviour amongst young people at a similar stage of development (pre-adolescent or adolescent), as this older age group are often at a transitional stage of life in which they move from school into sedentary, office-based occupations. Exclusion of those aged 5-9 years was for pragmatic reasons because of differences in the way data was collected. For this younger age group, a parent gave a proxy report of activities the child performed whilst being directly supervised by the parent. Therefore, all of the time spent at school was coded as "school", and no further information was provided about sedentary behaviour or physical activity during this time. This was not comparable with older participations who recalled a full 24 hours of activity. Finally, the use of a 10-18 year old age group allows for comparison with other international data sets.
The tools used to capture sedentary behaviour are each associated with their own limitations. As discussed, the MARCA data may be affected by the biases associated with self-report, though the psychometric properties have been shown to be sound, and there was a very high proportion of valid data. For accelerometry, it is important to note that no current consensus exists on the most appropriate way to define sedentary behaviour, with published definitions ranging from 100-1100 activity counts per minute (cpm) . Furthermore, estimates relied on participant compliance with wearing the accelerometer during waking hours. When compared to the MARCA, a smaller proportion of participants provided valid data (99.6% vs. 72.4%).
Further discussion of issues associated with the measurement of sedentary behaviour via self-report and accelerometry is warranted. Sedentary behaviour is an emerging field of research in public health, but there remains little consensus regarding the optimal way of defining and measuring it. In this survey, the MARCA and accelerometry captured distinct aspects of sedentary behaviour. A seated or lying body posture was the key criterion for defining sedentary behaviour according to the MARCA, with a secondary MET threshold applied. For example, "watching TV sitting - 1.2 METs" was classified as a sedentary behaviour, whereas "playing the drums - 4.0 METs" was not, even though this was a seated behaviour. Conversely, accelerometry defines sedentary behaviour as low movement and does not provide any information on body posture. Standing still can elicit low accelerometer counts and therefore accelerometer-derived sedentary time will likely include a mixture of time spent both sitting and standing.
For total sedentary behaviour, estimates from the MARCA (521 [SE 5.29] minutes per day) were higher than those from accelerometry (420 [SE 4.26] minutes per day). The 100 minute discrepancy between methods raises the question as to which estimate best represents the "true" value. Differences in the operationalisation of sedentary behaviour, as well as differences in what the tools actually measure, most likely accounts for the discrepancy in outputs. Because of their objectivity, accelerometers have inherent appeal, but are not a criterion method for assessing sedentary behaviour. A new class of devices based on inclinometry (e.g., the activPAL) show promise for the objective assessment of body posture . Though it relies on self-report, the MARCA has potential to provide additional information on what individuals do when they are sedentary. It is likely that the "true" value lies somewhere between the two estimates. Though the absolute value of the estimates differed between the two methods, the patterns of sedentary behaviour among groups (e.g., younger vs. older age groups, ethnic groups) were strikingly similar.