The aim of this study was to investigate the association between a range of enabling and reinforcing recess variables and children's sedentary, moderate and vigorous physical activity in this context. The results revealed that girls engaged in 13.8% more sedentary activity and 8.2% less vigorous activity than boys during recess. This supports previous research highlighting how boys are more physically active than girls when measured using objective [8, 10, 25] and self-report measures .
Play space was significantly negatively associated with sedentary activity and positively associated with vigorous activity. Previous studies have noted that more space available per child is associated with recess physical activity  and that children are more active in spacious compared to restricted environments . Conversely, Sallis et al.  found that area size was not significantly associated with physical activity, though their study assessed available space in different school ground areas rather than the space available per child. The current study's findings may partly be explained by the types of recess activities the children engaged in. Armitage  stated that soccer often dominates over half the elementary school playground yet it is played by approximately one quarter of the school population, leading to the remaining children situating themselves around the playground perimeter and engaging in inactive behaviours. As such, recess strategies to increase physical activity could consider reducing the dominance of soccer by allocating specific areas for this activity [7, 10], which could enable more children to be active as they have more space on the playground.
Interventions conducted in recess have used games equipment to enable greater engagement in active behaviours, with both positive effects [7, 10, 29] and no effects  found. This present study found that equipment availability was a positive predictor of moderate and negative predictor of sedentary activity. Interestingly, this contrasts Haug et al.  and Zask et al.  who found that, in elementary school children, equipment availability was not significantly associated with moderate or vigorous physical activity. Zask et al.  did find, however, that the ball-to-child ratio was a significant predictor of vigorous physical activity, whilst Willenberg et al.  found that more children engaged in vigorous physical activity when loose equipment (e.g. bats, balls, jump ropes) was provided. It may be that equipment enables children to be active and certain equipment simulates greater activity compared to others. For example, soccer balls may stimulate continuous physical activity engagement, while jump rope could facilitate intermittent bouts of activity. Overall, these findings suggest that the use of equipment may be a simple way to enable children to be active during recess, though further research is needed to determine how the availability of equipment, the type of equipment available and the amount of equipment provided influence physical activity and sedentary behaviours. Such information could inform future recess interventions.
In the present study, temperature was negatively associated with vigorous activity during recess. This supports Ridgers et al.  and Farley et al. , who collected data in June-July in North-West England and over a 2-year period in New Orleans, USA, respectively. However, these findings contrast Zask et al. , who collected data in February-April in Australia. A recent review by Carson and Spence  noted that meteorological variables have not generally been specifically researched in physical activity studies, though seasonal variations suggest that higher activity levels occur during summer months. Duncan et al.  found that boy's activity increased as temperature increased, whilst in adolescents increasing temperatures are associated with more physical activity sessions . Notably, these latter studies have focused on habitual physical activity, where children choose to be outside and are likely influenced by weather conditions. This differs from recess in the UK where children are required to be on the playground during recess with the exception of when they eat their lunch or when it is raining. It is possible that whilst warmer weathers encourage children to spend time outdoors, in a context where children already access the outdoor environment temperature may be an important consideration when assessing physical activity levels and implementing interventions. As such, it is recommended that recess studies should report temperature and weather to aid cross-study comparisons.
Interestingly, physical playground features were not associated with physical activity at any intensity. This supports the findings of Cardon et al.  but contrasts previous research that have changed the physical environment to enable physical activity engagement [9, 10]. It is possible that the measure used in this study was not sensitive enough to detect associations with physical activity. Alternatively, it may be that the age of the playground markings and features that are associated with physical activity during recess, as recent data have suggested that physical playground features have a positive effect on children's activity, but the effects begin to decrease after 6 months . More research in this area is needed.
The number of supervisors on the playground was also not associated with physical activity during recess. Mixed findings have been reported concerning this issue to date [11, 20, 27, 31]. In the present study, supervising adults generally allowed the children to engage in free play, and were not involved in the organisation of children's games. The presence of adults could be one potential strategy to decrease children's sedentary activity and reinforce physical activity , though since recess is an opportunity for activity engagement which is largely free from adult control, more research is needed to identify factors in this context which could enable and reinforce children's activity levels.
There are several limitations that warrant attention. Firstly, the sample size at the school level is relatively small, though multilevel modelling analyses have been used previously on a similar number of schools . In addition, the playgrounds were not broken down in to specific areas (e.g. fixed equipment, soccer pitches), which has been the approach adopted in European and American studies and effects comparability across studies. In this study, children only had access to one playground which contained different fixed structures, which made it difficult to break the available outdoor space in to specific areas. Thirdly, collected data during short recess periods may not be representative of activity levels and behaviour during longer recess periods, though research suggests that physical activity data collected within discrete periods of the day may be more consistent in the school environment . Notably, no differences were found between the physical activity variables and the recess periods in this study. Lastly, it should also be noted that the three outcome variables of interest are correlated, therefore engagement in one physical activity intensity will influence the proportion of time that children are active at another intensity.