Ten-year old children participated in a wide range of activities which often conformed to gender stereotypes. For example, a greater proportion of boys reported playing football compared to girls. This indicates that within physical activity promotion interventions and public health strategies it may be necessary to allow for personal and gender specific activity preference.
The most commonly reported activity was running, which contrasts with previous studies which have reported walking, basketball, swimming and football as the most common activities in young people [7, 9–11, 35]. Differences in activity participation from previous research may be due to heterogeneous cultures, age ranges, activity questionnaires and season in which data were collected. The current data were collected in spring and summer (April–July) and may reflect seasonal fluctuation in activity participation. For example, Rugby, which in Britain is typically played in winter, had lower participation than Cricket, which is traditionally played in summer. Furthermore, 9% of children reported participation in snow-based activities. Although there was no snow in Norfolk at the time of data collection there is a large ‘dry ski-slope’ in the county, which could explain this finding.
Differences in Variety and Frequency of activities between population subgroups (sex, activity status and parent/guardian education level) suggest that the characteristics of children’s physical activity behaviour are not homogenous. Tailoring interventions and public health strategies to the physical activity characteristics within the target population may therefore be appropriate. For example, an intervention targeting children who do not meet government physical activity guidelines may focus on increasing the Variety and Frequency of activities performed, as both are significantly lower than in children meeting the guidelines.
Results from separate models suggest that Variety and Frequency reflect objectively measured MVPA to a similar extent. However, the associations between Variety and MVPA were not independent of Frequency. This appears to suggest that if the total number of activity sessions performed is high, it may not matter whether children perform many activities few times per week or few activities many times per week. If both approaches result in high total Frequency of activity, they may both result in more MVPA overall. Therefore, both approaches could be relevant to health promotion depending on the population or personal preference. In addition, the lack of significant interactions indicates that the associations of Variety and Frequency with physical activity are similar across population subgroups. This suggests that Variety and Frequency may be relevant components of physical activity to promote for all children. The lack of independence between Variety and Frequency in associations with MVPA could be explained if Frequency is on the causal pathway between Variety and MVPA. It is possible that increasing Variety could increase Frequency and in turn MVPA. Therefore, offering a wide range of activities in interventions promoting physical activity or including a minimum recommended number of different activities in future physical activity guidelines may be relevant. Longitudinal data analysis or a randomized controlled trial is necessary to confirm or refute this suggestion as this cross-sectional study leaves potential for reverse causality.
Strengths and limitations
It was possible to assess associations between objectively measured MVPA and two potentially important and mostly unexplored components of physical activity (Variety and Frequency). Moreover, data were obtained from a large population based sample, aiding the external validity of these results. However, the population of Norfolk is not representative of the rest of the UK , and results should be generalised to other populations with caution.
Accelerometers cannot accurately record activities without vertical hip movement e.g. cycling, and they must be removed for water based activities, therefore physical activity may have been underestimated in children who regularly cycle or swim. The focus of this paper was MVPA so sedentary time and light activity were not incorporated into the main models. The analyses could be repeated with objectively measured sedentary time and light activity as the main outcomes of interest as there may be different associations with Variety and Frequency of specific sedentary and light activities.
Self-report questionnaires are limited in their accuracy for estimating physical activity intensity and duration  but they can provide information about activity type that is not yet easily available from accelerometers. By examining self-reported Variety and Frequency, but not duration of activity the recall bias resulting from self-reported duration of activity should be minimised. We assumed that children’s ability to recall whether or not they participated in an activity may be more accurate than their recollection of activity duration. This is in line with the YPAQ being able to adequately rank children’s physical activity . The questionnaire presented a list of pre-specified activity categories with multiple activities in some items such as ‘Basketball/volleyball’. Therefore, it is unavoidable that for some activities we lose detail about the specific activities that children participated in. Furthermore, the frequency of participation was reported in days per week with a maximum of ‘4 or more times’ in the last seven days. This may result in under-reporting of activities which could be participated in on multiple days and also multiple times per day. We assumed that the 202 children who had less than 1/3rd of YPAQ data missing did not participate in activities with no response. Although this could have led to some error, the sensitivity analysis conducted excluding children with any missing YPAQ data, indicated that this assumption did not affect the results.
Only a small amount of the variance in objectively measured physical activity could be accounted for by self-reported Variety and Frequency. This could be partly because self-reported and objective physical activity was not assessed simultaneously. In addition, the YPAQ only asked about non-school activities, whereas objectively measured physical activity was determined over the whole day. This approach was chosen to inform the development of interventions aiming to facilitate children to increase their activity level over the whole day.
Variety and Frequency were calculated such that when children reported an extra activity (i.e. increased their Variety by one) they also reported at least one extra session (i.e. Increased their Frequency by at least one). Data were tested for multicollinearity and results indicated that it was acceptable to include Variety and Frequency in the same model and therefore the observed associations are unlikely to be artefacts due to the study design .
In future it may be possible to use novel technologies to objectively measure the types of activity being performed, allowing further exploration of the patterns and characteristics of physical activity mode to better inform the design and targeting of physical activity promotion strategies. Longitudinal research with objectively measured physical activity is also necessary to examine whether Variety and Frequency of activities in childhood are associated with activity maintenance, uptake and drop out throughout adolescence and adulthood .
Conclusion and public health implications
Given the wide range of activities reported and the apparent sex differences in participation it may be beneficial for future interventions and public health strategies promoting physical activity in children to allow scope for personal and sex specific activity preference. Furthermore, tailoring interventions to the physical activity characteristics within a target population may increase the chance of developing intervention programs that are acceptable, thus facilitating uptake and adoption. There was no evidence in 9–10 year old British children that activity Variety was associated with objectively measured MVPA independent of activity Frequency. Increasing the Frequency of existing activity or introducing new types of activity may therefore both be potential physical activity promotion strategies. However, longitudinal research is needed to examine whether both aspects of activity are similarly associated with activity maintenance.