The effectiveness of the HIPPA intervention in the sociocultural environment of ECEC physical activity

This is a self-archived version of an original article. This version may differ from the original in pagination and typographic details. Author(s): Title: Year: Version: Copyright: Rights: Rights url: Please cite the original version: CC BY-NC 4.0 https://creativecommons.org/licenses/by-nc/4.0/ The effectiveness of the HIPPA intervention in the sociocultural environment of ECEC physical activity © 2023 Anette Mehtälä, Arja Sääkslahti, Anne Soini, Marita Poskiparta, and Sami Kokko. Published version Mehtälä, Anette; Sääkslahti, Arja; Soini, Anne; Poskiparta, Marita; Kokko, Sami Mehtälä, A., Sääkslahti, A., Soini, A., Poskiparta, M., & Kokko, S. (2023). The effectiveness of the HIPPA intervention in the sociocultural environment of ECEC physical activity. Journal of Early Childhood Education Research, 12(1), 169-204. https://journal.fi/jecer/article/view/116848 2023 Journal of Early Childhood Education Research Volume 12, Issue 1, 2023, 169–204 © 2023 Anette Mehtälä, Arja Sääkslahti, Anne Soini, Marita Poskiparta, and Sami Kokko. Peer-review under responsibility of the editorial board of the journal. Publication of the article in accordance with the Creative Commons Non-Commercial license. ISSN 2323-7414; ISSN-L 2323-7414 online. Early Childhood Education Association Finland. The effectiveness of the HIPPA intervention in the sociocultural environment of ECEC physical activity Anette Mehtäläa, Arja Sääkslahtib, Anne Soinic, Marita Poskipartad & Sami Kokkoe a Jamk University of Applied Sciences, Jyväskylä, Finland & Faculty of Sport and Health Sciences, University of Jyväskylä, Finland, corresponding author, e-mail: anette.mehta[email protected], https://orcid.org/0000-0002-9299-473X b Faculty of Sport and Health Sciences, University of Jyväskylä, Finland, https://orcid.org/0000-0003-4354-0990 c Faculty of Education and Psychology, University of Jyväskylä, Finland, https://orcid.org/0000-0001-9168-9437 d Faculty of Sport and Health Sciences, University of Jyväskylä, Finland, https://orcid.org/0000-0002-2060-7892 e Faculty of Sport and Health Sciences, University of Jyväskylä, Finland, https://orcid.org/0000-0001-9436-5681 ABSTRACT: Studies have shown that, on average, children’s physical activity (PA) levels in early childhood education and care (ECEC) environments are low, thus opening up the possibility of interventions. We examined the effect of the teacherimplemented one-year long ‘home-and-childcare-based intervention to promote physical activity’ (HIPPA) study on children’s PA in ECEC settings in Finland. Participating four-year-old children (N = 128) were cluster-randomised into two groups, intervention (seven childcare centres) and control (seven childcare centres), in autumn 2011. The children were observed directly during their ECEC times. Multilevel linear analysis was used to test the PA differences between the intervention conditions. Post-intervention results showed that the HIPPA intervention increased the children’s PA. To enhance the real-life effectiveness of the present multicomponent intervention, we examined methods based on the intervention’s success and found areas of development for future studies. Overall, 170 Mehtälä, Sääkslahti, Soini, Poskiparta & Kokko. Journal of Early Childhood Education Research 12(1) 2023, 169–204. http://jecer.org promoting active play by ECEC personnel offers an effective way to enhance children’s PA. Furthermore, to ensure the sustainability of the intervention effects, specific changes in practice have been identified that should be transferred into the policies intended for ECEC settings. Keywords: children, physical activity, early childhood education and care, multicomponent intervention research Introduction Current studies suggest that young children’s physical activity (PA) behaviours have changed over the decades. The prevalence of early childhood overweight and obesity has increased rapidly worldwide (de Onis et al., 2010; Di Cesare et al., 2019; NCD-RisC, 2017). Environmental, behavioural, biological, and genetic factors all impact on weight gain (Di Cesare et al., 2019), and it seems very likely that changes in PA and sedentary behaviour have promoted this unfavourable development (de Onis et al., 2010; Di Cesare et al., 2019; NCD-RisC, 2017). Not only does living in a modern society require less physical effort than before (Hill et al., 2003; Wiklund, 2016), but it also offers a wide variety of sedentary activities (Gray et al., 2015). Opportunities for physical activities requiring a wide range of motor skills may have diminished. This can impair performance in tasks that require good coordinative skills, such as balancing and throwing (Roth et al., 2010). Therefore, attention should be paid to children’s PA, and time and space should be organised so that children have the opportunity to achieve sufficiently diverse PA on a daily basis. The World Health Organization (WHO) recommends PA for a total of 180 min per day, including at least one hour of moderate-to-vigorous PA (MVPA) for children threeto four-year-olds (WHO, 2019). These evidence-based guidelines aim to enhance the growth and development of children and are consistent in several countries across continents, such as Australia, Canada, the United Kingdom, and the United States of America. In Finland, corresponding recommendations apply for children under eight years old (Ministry of Education and Culture, 2016). Studies have shown an increasing interest in children’s PA behaviour in early childhood education and care (ECEC) settings, with results pointing out that children’s PA in various ECEC environments is generally low, thus opening up the possibility of an intervention (Jones et al., 2019). ECEC environments can potentially impact children’s PA because most children aged 3–5 years engage in ECEC (OECD, 2022). Review articles addressing the efficiency of health behaviour (i.e., PA and diet) interventions in childcare-age children have been published in recent years (Stacey et al., 2017). It is also worth noting that successfully implemented PA and fundamental motor skills interventions in ECEC may enhance children’s overall well-being and support their cognitive and learning skills 171 Mehtälä, Sääkslahti, Soini, Poskiparta & Kokko. Journal of Early Childhood Education Research 12(1) 2023, 169–204. http://jecer.org (Jylänki et al., 2022). The most recent meta-analysis of 21 published ECEC-setting intervention studies found a small but significant impact on children’s MVPA but no significant effect on light PA (LPA). A more substantial overall impact on MVPA than LPA may reflect the contents of the intervention programmes, which focused mainly on enhancing MVPA by promoting, for example, active play outdoors or structured gross movement sessions in ECEC settings (Hnatiuk et al., 2018.) Reducing playground density and providing portable play equipment have proven to be effective intervention methods for increasing children’s outdoor PA (Stacey et al., 2017). Although there is significant variation in the frequency and intensity of outdoor PA, children spend more than half of their outdoor time engaged in sedentary activities (Truelove et al., 2018). This suggests that free play alone may not be sufficient to promote PA (Alhassan et al., 2007; Cardon et al., 2009) or to develop gross motor skills in all children, which is vital for lifelong PA (Brian et al., 2017; Stodden et al., 2008). ECEC personnel involvement (e.g., prompts to PA, trips to a forest to play or learn through play) or structured PA (e.g., obstacle courses, rule games) during recess periods is needed alongside free play. Nevertheless, outdoor play is an important form of children’s PA. Outdoors, children have space to move and play, giving them opportunities to develop their creativity and social skills (Aras et al., 2016; Truelove et al., 2017, 2018). Children have been observed in childcare settings to be up to two times more physically active outdoors as indoors (Tandon et al., 2018). The provision of structured PA (e.g., integrated into the preschool curriculum) was one of the potentially effective methods that can facilitate children’s PA, as observed in two recent reviews and meta-analyses listing several childcare centres’ policies and practices (Hnatiuk et al., 2018; Stacey et al., 2017). Studies have reported that the most promising effects result from methods that target the improved knowledge and qualifications of ECEC personnel through training, and physical environments, such as modifications of playground space (e.g., by reducing its density) and providing portable equipment (Stacey et al., 2017). ECEC personnel should have knowledge of children’s motor skills development and the ability to support development in an age-appropriate way. Another approach is to tailor an intervention to the distinct cultural and societal needs of the target group. Tailoring can mean providing ongoing support to address context-specific barriers or modifying materials to suit the local community (Hnatiuk et al., 2018). This is an important factor to consider in today’s standards because ECEC personnel already face challenges, such as time-consuming administrational duties, when performing their basic routines (Aras, 2016). It is essential to ensure that ECEC personnel can change their practices to help promote children’s PA under their supervision (Hnatiuk et al., 2018). From the perspective of long-term health behaviour effects, the following factors should be considered: children enjoy the offered activities, interventions are easy to implement 172 Mehtälä, Sääkslahti, Soini, Poskiparta & Kokko. Journal of Early Childhood Education Research 12(1) 2023, 169–204. http://jecer.org and are embedded into ECEC environments without demanding extra time, there are enough resources (money or personnel), and parents and children are consistently informed (Finch et al., 2014; Zhou et al., 2014). Despite the growing interest in promoting PA in ECEC settings, little is known about the characteristics of the interventions facilitating children’s PA. Furthermore, it has been suggested that a more comprehensive approach should be used, rather than simply focusing on a single determinant when considering the effectiveness of an intervention (Gubbels et al., 2014). Thus, the first purpose of the current study is to examine the effects of the teacher-implemented, one-year-long ‘home-and-childcare-based intervention to promote physical activity’ (HIPPA) intervention on children’s directly observed PA in ECEC settings. Second, this study explores the moderating factors in the observed physical (outdoors and indoors) and social environments (adult initiation of PA or adult involvement or presence in PA), wherein children’s PA occurs at childcare centres using the observational system for recording physical activity in preschool children (OSRAC-P). Observed social-environmental factors were the specifically primary location of the child in the ECEC centre, time of day, group composition (e.g., child group, adult group), and activity initiator (child or adult). Observed outdoor and indoor factors included activity contexts, for example playing in open space, with fixed equipment, or in the sandbox; participation in group time; and teacher-arranged gross motor activity or sociodramatic play (see Appendix 1). Differences in the cross-sectional and longitudinal associations of children’s PA between the intervention (HIPPA) and control groups were assessed. Finally, the sensitivity of the intervention was examined by separately assessing the associations amongst and between boys and girls. Methods This cluster-randomised HIPPA intervention study is part of a more extensive study on Dutch and Finnish twoto six-year-old children’s PA in childcare and home settings (see Mehtälä et al., 2017; Soini, 2015). A follow-up study (2010–2013) was developed to estimate children’s PA behaviours and those of their parents using several different methods. Direct observation (OSRAC-P; Brown et al., 2006), accelerometers (Actigraph GT3X) and proxy reports were used to assess the children’s PA. In the present study, only data collected by the OSRAC-P were used. In Finland, the study consisted of two phases. The first phase of the study (Phase I, autumn 2010 to winter 2011) identified the level and characteristics of children’s PA at different times of the year (Figure 1). The purpose of the second phase (Phase II, spring 2011 to winter 2013) was to plan and implement an intervention appropriate to the environment 173 Mehtälä, Sääkslahti, Soini, Poskiparta & Kokko. Journal of Early Childhood Education Research 12(1) 2023, 169–204. http://jecer.org to increase the PA of childcare-age children based on the current best knowledge and the evaluation of the study results of the first phase (Soini, 2015). Ethical approval for the study was obtained from the local ethical committee and the social affairs and health officer of the city where the study was conducted. Study design and participants In spring 2010, 60 public childcare centres located throughout a city in central Finland were invited to participate in the study (Figure 1). Fourteen childcare centres accepted the invitation after a regional administrative meeting and distributed the information letters and consent forms to eligible families. The sole inclusion criterion was that the children enrolled in the participating childcare centres must be born in 2007. The 14 participating childcare centres were representative of the typical Finnish ECEC system. Most childcare centres in Finland are municipal (Finnish National Agency for Education, 2022). In this study, 13 were municipal childcare centres, and 1 was private. In the present study, we examined the intervention data (Phase II) from autumn 2011 to winter 2013. The time included four data collection points: initiation, midway, postintervention, and follow-up of the HIPPA intervention (Mehtälä et al., 2017). There were 220 children born in 2007 at the participating childcare centres in autumn 2011. The families of 128 eligible children provided informed consent. Eleven children had missing intervention initiation data because of sickness or other reasons for their absence during the measurement days, thus leading to a final sample of 117 children (53%). The participating children were cluster-randomised into two groups: intervention (seven childcare centres) and control (seven childcare centres). Paired and random allocations were used, and the childcare centres were matched based on their locations. The participants in the intervention centres were exposed to the HIPPA, whereas those in the control centres received their usual daily programmes (Mehtälä et al., 2017). 174 Mehtälä, Sääkslahti, Soini, Poskiparta & Kokko. Journal of Early Childhood Education Research 12(1) 2023, 169–204. http://jecer.org FIGURE 1 Flowchart of study participants. The present study analysed only the data collected during Phase II. 60 childcare centres Baseline Autumn 2010 (14 centres) Baseline Winter 2011 (14 centres) Randomised by centre (n = 220 children) Intervention initiation (n = 70) Autumn 2011 Absent during measurement days (n = 6) Allocated to HIPPA intervention, 7 centers (n = 76) Allocated to control group, 7 centers (n = 52) Did not wish to participate (n = 92 children) Usual ECEC (n = 47) Autumn 2011 Absent during measurement days (n = 5) Intervention mid-way (n = 59) Winter 2012 Dropped out of the study (n = 1) Absent during measurement days (n = 10) Usual ECEC (n = 43) Winter 2012 Dropped out of the study (n = 1) Absent during measurement days (n = 3) Post-intervention (n = 52) Autumn 2012 Dropped out of the study (n = 7) Absent during measurement days (n = 10) Usual ECEC (n = 37) Autumn 2012 Dropped out of the study (n = 7) Absent during measurement days (n = 2) Follow-up (n = 51) Winter 2013 Dropped out of the study (n = 3) Absent during measurement days (n = 8) Usual ECEC (n = 34) Winter 2013 Dropped out of the study (n = 2) Absent during measurement days (n = 3) Did not wish to participate (n = 46 centres) Analysed (n = 70) Analysed (n = 47) Phase I Spring 2010– winter 2011 Phase II Spring 2011– winter 2013 175 Mehtälä, Sääkslahti, Soini, Poskiparta & Kokko. Journal of Early Childhood Education Research 12(1) 2023, 169–204. http://jecer.org In each data collection period, each child’s height was measured by two researchers to the nearest 0.1 cm and their weight to the nearest 0.1 kg using a portable stadiometer (Charder HM 200P) and a digital scale (Seca 877), respectively. The body mass index (BMI) was calculated for each child as weight (kg) divided by the squared height (m2). The descriptive data of the participating children at the start of the intervention are presented in Table 1. In general, the children in the HIPPA group were younger, shorter, and weighed less than those in the control group. Notably, age differences appeared between girls, while a difference in mean weight was recorded between boys. Amongst the HIPPA group, girls had lower observed mean PA than boys. TABLE 1 Participants’ characteristics at the start of the HIPPA intervention CHARACTE RISTIC CONTROL GROUP HIPPA INTERVENTION GROUP Boys Girls Total Boys Girls Total n Mean (SD) n Mean (SD) N Mean (SD) n Mean (SD) n Mean (SD) N Mean (SD) Age, year 23 4.24 (0.24) 24 4.31 (0.26)* 47 4.28 (0.25)* 34 4.13 (0.28) 36 4.09 (0.29)* 70 4.11 (0.28)* Height, cm 23 108.1 (4.9)* 23 106.3 (3.9)* 46 107.2 (4.4)* 31 104.9 (6.2)* 36 103.3 (3.6)* 67 103.6 (5.0)* Weight, kg 23 18.7 (2.6)* 23 17.8 (2.1) 46 18.2 (2.4)* 31 17.3 (2.6)* 36 17.3 (2.1) 67 17.3 (2.3)* BMI, kg/m2 23 15.9 (1.5) 23 15.7 (1.2) 46 15.8 (1.3) 31 15.9 (1.1) 36 16.2 (1.4) 67 16.1 (1.3) Mean PA 23 2.57 (0.40) 24 2.44 (0.29) 47 2.51 (0.35) 34 2.50 (0.15)** 36 2.37 (0.20)** 70 2.43 (0.19) Note. *Statistically significant difference (p < .05) between intervention conditions and **genders. Intervention The socio-ecological model was applied as a framework in the HIPPA intervention. This approach is commonly used to guide the selection of strategies and methods for intervention studies because it considers the dynamic and complex interplay between an individual and the environmental factors that can affect one’s health behaviours (McLeroy et al., 1988). One of the leading ideas was to plan the intervention collaboratively with ECEC personnel. The tailored intervention based on each centre’s own needs would be primarily implemented by the ECEC personnel and must be able to overcome real-life challenges, such as lack of resources. The intervention was planned with the ECEC personnel based on the data collected in autumn 2010 and winter 2011 and it utilised the former Finnish early childhood PA recommendations (Ministry of Social Affairs and Health, 2005). The main targets to be modified were outdoor playtime, indoor facilities of 176 Mehtälä, Sääkslahti, Soini, Poskiparta & Kokko. Journal of Early Childhood Education Research 12(1) 2023, 169–204. http://jecer.org the centres, the PA knowledge and skills of the principals, ECEC personnel’s competence and support for PA, and the children’s motivation and self-efficacy for PA. In the intervention planning meetings held in the spring of 2011 between researchers and ECEC personnel, the targets were considered, and various methods were presented to promote children’s PA in childcare centres’ typical daily life situations. Children’s guardians had the opportunity to record their ideas to anonymously promote children’s PA. Several methods were identified: modifying outdoor time and space, that is by an implementation of adult-led or adult-initiated physical activities or by staggering outdoor time of different child groups; parental activation, for example at a parents’ evening; inservice education of ECEC personnel; making indoor spaces more stimulating for PA; more effective use of existing PA programmes; and colleagues’ support. The discussions also raised issues that were potential barriers to the implementation of these methods in the intervention. They were related to children’s safety, for example increased risk of injuries, size of the indoor facilities and child groups, and concerns about personnel as well as equipment adequacy. The attitudes of the personnel and guardians towards PA were highlighted as both a barrier and an enabler to increase children’s opportunities for PA. Various intervention materials were used to increase children’s PA and to promote children’s as well as ECEC personnel’s and guardians’ motivation, knowledge, and selfreliance: the ‘moving bead box’ to increase children’s MVPA, especially; family PA tip cards and PA monitoring cards to enhance joint activity in the family; the ‘best practices’ poster with 20 tips on how to promote PA among children in ECEC; and intervention folders to families where to collect their monthly letters considering issues related to overall well-being. As much as possible, the intervention utilised existing materials to promote children’s PA of in ECEC settings (i.e., materials of the Young Finland association Varpaat Vauhtiin! and Pihaseikkailu). The personnel were encouraged to discuss PA with parents during daily encounters and upon finalization of the individual ECEC plan. The emphasis of the plan is on the objectives set for the pedagogical activities. The ECEC partnership combines the knowledge and experience of parents and personnel, whose relationship plays an essential role in ensuring a child’s well-being (Heikkilä et al., 2004). The researchers participated in parents’ evenings, where they discussed the baseline results (Phase I; Soini et al., 2012) on the importance of PA for health and well-being and how guardians can enhance it through their own actions. Monthly letters related to health and well-being were also distributed to families through the ECEC. In addition, families, as important role models for children, were offered the opportunity to borrow and use pedometers during the measurement period. 183 Mehtälä, Sääkslahti, Soini, Poskiparta & Kokko. Journal of Early Childhood Education Research 12(1) 2023, 169–204. http://jecer.org TABLE 4 The moderating effects of the general socio-environmental context covariates on the relationship between the intervention condition and the children’s observed PA from the start of the intervention up to post-intervention EFFECT ESTIMATE SE 95% CI p* LL UL Gender –0.013 0.007 –0.028 0.001 .066 Primary location 0.025 0.007 0.012 0.038 <.001 Time of day –0.037 0.006 –0.049 –0.024 <.001 Group composition 0.005 0.010 –0.015 0.025 .621 Initiator –0.056 0.011 –0.078 –0.033 <.001 Note. Children’s PA = mean of the highest observed PA level at each observation interval (1–8). The variable was cube root transformed. Level 1 Observed intervals Control N = 5,360, HIPPA N = 6,935; Level 2 Children control N = 47, HIPPA N = 70, Level 3 Childcare centre N = 14; CI = confidence interval; LL = lower limit; UL = upper limit. *p-values from the cross-level interaction test. Gender: girl = 0, boy = 1; Primary location: indoors = 0, outdoors = 1; Time of day: morning = 0, afternoon = 1; Group composition: adult = 0, child = 1; Initiator: adult = 0, child = 1. The stratified subgroup analysis showed that the OR values of MVPA and LMVPA increased in favour of the HIPPA group indoors (OR = 4.65, 95% CI [2.45, 8.82]; OR = 2.95, 95% CI [1.46, 5.97], respectively) and in the afternoons (OR = 3.18, 95% CI [1.81, 5.58]; OR = 3.08, 95% CI [1.61, 5.90], respectively). The odds of MVPA, but not LMVPA were higher in the HIPPA group compared to the control group outdoors (OR = 2.33, 95% CI [1.41, 3.85]; OR = 1.51, 95% CI [0.93, 2.46], respectively) and in the mornings OR = 2.25, 95% CI [1.39, 3.63]; OR = 1.46, 95% CI [0.80, 2.66], respectively). The OR values of MVPA and LMVPA also increased in child-initiated (OR = 2.58, 95% CI [1.77, 3.77]; OR = 2.33, 95% CI [1.30, 4.18], respectively) but not in adult-initiated activities (OR = 1.51, 95% CI [0.58, 3.94]; OR = 0.61, 95% CI [0.17, 2.21], respectively). Figure 2 demonstrates the interaction effect between time and intervention condition based on the observed primary locations. 184 Mehtälä, Sääkslahti, Soini, Poskiparta & Kokko. Journal of Early Childhood Education Research 12(1) 2023, 169–204. http://jecer.org FIGURE 2 Interaction effects between time and intervention condition on children’s MVPA indoors and outdoors. Multilevel logistic linear regression models were adjusted with a grand mean centred children’s age. Indoor activities Significant interaction effects were observed in all indoor activities except transition (Table 5). Differences in the children’s average interval PA levels between intervention conditions were reduced due to the effects of these indoor factors. TABLE 5 The moderating effects of indoor activities on the relationship between the intervention condition and children’s observed PA from the start of the intervention to post-intervention EFFECT ESTIMATE SE 95% CI p* LL UL Toys –0.047 0.011 –0.068 –0.026 <.001 Other –0.051 0.011 –0.073 –0.030 <.001 Art –0.167 0.016 –0.199 –0.136 <.001 Group time –0.046 0.014 –0.073 –0.020 .001 Sociodramatic –0.036 0.016 –0.067 –0.004 .026 Teacher arranged –0.050 0.018 –0.085 –0.015 .005 Transition –0.005 0.017 –0.039 0.029 .776 Note. Children’s PA = mean of the highest observed PA level at each observation interval (1–8). The variable was cube root transformed. Control Level 1 Observed intervals N = 2,488, HIPPA N = 3,234, Level 2 Children control N = 47, HIPPA N = 70, Level 3 Childcare centre N = 14; CI = confidence interval; LL = lower limit; UL = 185 Mehtälä, Sääkslahti, Soini, Poskiparta & Kokko. Journal of Early Childhood Education Research 12(1) 2023, 169–204. http://jecer.org upper limit. *p-values from the cross-level interaction test. Toys = playing with toys, dolls, dollhouses, Lego, puzzles etc.; Other = being in some other indoor context or engaging in some activity other than the options listed in the OSRAC-P; Art = engaging in art activities or being in an art centre or activity area; Group time = participating in a group activity that is teacher organised or led; Sociodramatic = engaging in sociodramatic or pretend to play activities or being in a sociodramatic play centre; Teacher arranged = engaging in teacher planned, arranged, and led gross motor physical activities with or without equipment; Transition = moving from one classroom activity context to another area without engaging in materials. However, the stratified subgroup analysis showed that the odds of MVPA and LMVPA were higher when children played with toys (OR = 24.36, 95% CI [4.95, 119.78]; OR = 4.53, 95% CI [1.63, 12.56], respectively) in the HIPPA group compared to the control group. The odds of MVPA were higher in sociodramatic play and in other activities than listed in the OSRAC-P in the HIPPA group compared to the control group (OR = 44.50, 95% CI [8.42, 235.27]; OR = 4.50, 95% CI [1.57, 12.94], respectively]. Outdoor activities Significant interaction effects were observed in almost all outdoor activities (Table 6). The effects of these outdoor factors increased the differences in the children’s average interval PA levels between intervention conditions. TABLE 6 The moderating effects of outdoor activities on the relationship between the intervention condition and children’s observed PA from the start of the intervention up to the post-intervention EFFECT ESTIMATE SE 95% CI p* LL UL Fixed 0.325 0.065 0.198 0.451 <.001 Open space 0.418 0.064 0.294 0.543 <.001 Portable 0.393 0.109 0.179 0.608 <.001 Sandbox 0.395 0.067 0.264 0.526 <.001 Sociodramatic 0.441 0.094 0.256 0.626 <.001 Wheel 0.454 0.087 0.282 0.625 <.001 Other 0.154 0.100 0.041 0.349 .123 Note. Children’s PA = mean of the highest observed PA level at each observation interval (1–8). Control Level 1 Observed intervals N = 2,304, HIPPA N = 2,829, Level 2 Children control N = 47, HIPPA N = 70, Level 3 Childcare centre N = 14; CI = confidence interval; LL = lower limit; UL = upper limit. *p-values from the cross-level interaction test. Fixed = Engaging in activity on fixed playground equipment or being on the fixed playground equipment; Open space = Being in an open outdoor area that is not one of the other outdoor activity contexts; Portable = Engaging in activity with equipment brought to the playground other than balls or wheel toys; Sandbox = Engaging in activities using sandbox materials or being in a sandbox; Sociodramatic = Engaging in activity with sociodramatic play props or similar materials; Wheel = Touching, riding, or pushing wheel toys that are not fixed equipment (e.g., tricycles, scooters, wagons); Other = Outdoor activity context other than the options listed in the OSRAC-P. 186 Mehtälä, Sääkslahti, Soini, Poskiparta & Kokko. Journal of Early Childhood Education Research 12(1) 2023, 169–204. http://jecer.org The stratified subgroup analysis showed that the odds of MVPA were higher in the HIPPA group compared to the control group while playing in open spaces (OR = 2.49, 95% CI [1.30, 4.77]), playing with portable equipment (OR = 5.93, 95% CI [2.18, 16.12]), and playing in sandboxes (OR = 9.02, 95% CI [3.75, 21.74]). The odds of LMVPA were higher in riding or pulling/pushing toys with wheels in the HIPPA group compared to the control group (OR = 17.39, 95% CI [1.99, 151.84]). Girls The odds of MVPA and LMVPA indoors (OR = 3.22, 95% CI [1.36, 7.64]; OR = 2.21, 95% CI [1.07, 4.59], respectively) and in the afternoons (OR = 5.12, 95% CI [2.38, 11.02]; OR = 2.18, 95% CI [1.09, 4.34], respectively) were higher in the HIPPA girls compared to the girls in the control group. The odds of MVPA, but not LMVPA outdoors (OR = 2.73, 95% CI [1.49, 4.99]; OR = 0.92, 95% CI [0.50, 1.69], respectively), in the mornings (OR = 1.92, 95% CI [1.04, 3.57]; OR = 0.93, 95% CI [0.49, 1.78], respectively); and in child-initiated activities (OR = 2.67, 95% CI [1.81, 3.93]; OR = 1.45, 95% CI [0.88, 2.45], respectively), were higher in the HIPPA girls compared to those in the control group. Furthermore, the odds of MVPA outdoors were higher in HIPPA girls when playing in open spaces (OR = 2.46, 95% CI [1.07, 5.64]) compared to the control girls. Boys The odds of MVPA and LMVPA indoors (OR = 7.21, 95% CI [3.88, 13.39]; OR = 2.72, 95% CI [1.16, 6.36]), in the afternoons (OR = 2.43, 95% CI [1.04, 5.67]; OR = 4.91, 95% CI [1.76, 13.73]), and in child-initiated activities (OR = 2.53, 95% CI [1.18, 5.42]; OR = 3.74, 95% CI [1.36, 10.76]) were higher in the HIPPA boys compared to the control boys. Furthermore, the odds of MVPA indoors were higher in the HIPPA boys in playing with toys (OR = 8.91, 95% CI [1.15, 69.30]), and in other activities apart from those listed in OSRAC-P (OR = 15.31, 95% CI [4.10, 57.24]), compared to the control boys. The odds of LMVPA indoors were higher in playing with toys (OR = 5.41, 95% CI [1.56, 18.77]), and in sociodramatic play (OR = 23.47, 95% CI [2.01, 274.77]), but lower in transition (OR = 2.48 x 10-7, 95% CI [6,97 x 10-8, 1.74 x 10-6]), compared to the control boys. The odds of LMVPA, but not MVPA outdoors, were higher in the HIPPA boys compared to those in the control group (OR = 3.43, 95% CI [1.06, 11.08]; OR = 2.10, 95% CI [0.83, 5.30], respectively), whilst the odds of MVPA, but not LMVPA in the mornings, were higher in the HIPPA boys compared to those in the control group (OR = 2.56, 95% CI [1.18, 5.56]; 187 Mehtälä, Sääkslahti, Soini, Poskiparta & Kokko. Journal of Early Childhood Education Research 12(1) 2023, 169–204. http://jecer.org OR = 1.89, 95% CI [0.87, 4.13], respectively). The OR values of MVPA increased in favour of the HIPPA group in sandbox activities (OR = 21.90, 95% CI [9.75, 49.17]), and decreased in activities apart from those listed in the OSRAC-P (OR = 0.19, 95% CI [0.04, 0.84]). Furthermore, the odds of LMVPA outdoors were higher in the HIPPA boys in sandbox activities (OR = 12.01, 95% CI [4.98, 28.98]), in sociodramatic play (OR = 3.66, 95% CI [1.58, 8.5]), and in riding or pulling or pushing with wheel (OR = 3.29 x 108, 95% CI [1.68 x 107, 6.43 x 109]), compared to the control boys. Discussion This research focused on examining the effects of the HIPPA intervention on children in ECEC settings. A real-life HIPPA intervention was implemented by ECEC personnel to promote the PA of children aged four to five years. The intervention proved to be successful, based on the increased PA post-intervention amongst HIPPA children in ECEC settings. This study adds information to the scarce research on the long-term promotion of childcare-aged children’s PA (Jones et al., 2019). In the present study, children’s PA increased in the afternoons and during child-initiated activities, indicating that the intervention especially affected the more sedentary time of day (Soini et al., 2016) and unstructured time of ECEC. The increase in PA indoors suggests that the HIPPA intervention affected the practices of the childcare centres. Discussions with ECEC personnel highlighted the presence of restrictions and rules that were more likely to inhibit children’s PA indoors than outdoors. The safety of the children justified the existence of the rules, but some of the rules could be removed by thinking about the practices in a new way. Interestingly, gross motor activities were observed more frequently in the HIPPA centres post-intervention than at the start of the intervention (82% of the observed intervals were at post-intervention) and more frequently than in the control childcare centres (Table S3). In other words, HIPPA children were more likely than usual to engage in gross motor activities without adult initiation or immediate presence. The results also indicate that sociodramatic play and transition indoors were more often at physically higher intensive levels than at the start of the intervention. The findings support the notion that ECEC practices have changed to a more permissive and supportive direction in terms of PA indoors. Findings regarding gross motor activities and sociodramatic play are promising, although they do not seem to enhance the intervention effect on the observed mean PA. These activities can be considered as active play—an unstructured form of PA with or without equipment where, more importantly, children have fun (Truelove et al., 2017). Overall, the results support the hypothesis of Truelove and colleagues (2017) that active play may be easier to 188 Mehtälä, Sääkslahti, Soini, Poskiparta & Kokko. Journal of Early Childhood Education Research 12(1) 2023, 169–204. http://jecer.org promote amongst young children than PA from the perspective of ECEC personnel and that educators’ perceptions of PA for childcare-age children being only structured activities or simply involving running or jumping (Jones et al., 2019) may have changed. The intervention was supported by researchers; however, the ECEC personnel implemented the methods in accordance with the context of their respective centres. At the follow-up, the participating ECEC personnel were asked to fill in a questionnaire regarding the extent and quality of the implementation of the HIPPA intervention in their centres, and almost 60% answered the questionnaire. The results showed that the most frequently implemented intervention method was the modification of indoor facilities to promote PA. They used floor tapes and figures and built obstacle courses to encourage the children to move in various ways (balancing, jumping) but to do so safely in specific areas of the centre. Four out of five (81%) intended to keep PA equipment available to children during free play in the future (Mehtälä et al., 2017). Considering that children in ECEC are more sedentary in afternoons than in mornings (Soini et al., 2016) and indoors than outdoors (Tandon et al., 2018), it is logical that the HIPPA intervention affected more children’s PA in those contexts. Due to the lower PA indoors, there was more room to increase children’s PA compared to the outdoor setting, along with a variety of targets that the intervention could influence. The ECEC programme is structured indoors, especially in the mornings, but children’s spontaneous, free play is a highly cherished part of the ECEC setting (Arash, 2016, see also Table 2). The adult-led group sessions (i.e., group times) are also held in the mornings. Group time was the third most common indoor activity and proved to be very sedentary in both intervention conditions in the present study (Table S3). The level of group-time activity in the HIPPA intervention group even decreased slightly during the intervention. Regarding the relative contribution of group time to the total ECEC time in the mornings, teaching ECEC personnel how to activate those learning sessions physically could increase children’s PA whilst also providing opportunities to enhance their academic skills and fundamental motor skills (Jylänki et al., 2022; Trost et al., 2008; Van der Fels et al., 2015). Good motor skills enable children to enjoy various physical activities and may help them maintain a life-long active lifestyle (Stodden et al., 2008). The HIPPA intervention included methods to facilitate outdoor play time (i.e., stressing the importance of outdoor time and encouraging ECECs to modify the environment to make it more inspiring to PA with equipment or playground markings, with organised/adult-led or adult-initiated PA, or by using existing PA campaign materials) because of the unquestionable evidence of the importance of outdoor time in providing PA opportunities for children (Sääkslahti & Niemistö, 2021; Truelove et al., 2018). The level of intervention implementation was lower outdoors than indoors (59% vs. 86%), 189 Mehtälä, Sääkslahti, Soini, Poskiparta & Kokko. Journal of Early Childhood Education Research 12(1) 2023, 169–204. http://jecer.org but the outdoor time provided increased by 45 minutes per day in the HIPPA childcare centres (Mehtälä et al., 2017). In contrast to indoor conditions, outdoor facilities are usually suitable and safe for every kind of PA, which is why children are also allowed to be more physically active outdoors. Outdoor time is considered free time for children (Iivonen et al., 2021), which is also evident in the results of the present study and may be reflected in the implementation levels of the intervention. Adult-initiated activities and prompts for PA occurred very rarely. Teacher-arranged activities were not observed outdoors in the HIPPA centres and only in 1% of all outdoor intervals in the control centres. After half a year of intervention, children’s PA was slightly higher in the control group compared to the HIPPA group. In Finland, pre-primary school education is provided for six-year-old children. Thus, at the follow-up time, the participating children were already enrolled in pre-primary school education, so the adults around them, and even the group and the facility where they were during ECEC time, may have already changed. Previous studies have reported that increasing children’s outdoor play opportunities and the availability of PA equipment promote children’s PA, but these methods alone are insufficient to maintain such an increase over a more extended period (Alhassan et al., 2007; Cardon et al., 2009). This may be why children’s PA in the HIPPA group decreased after half a year of intervention. Specifically, changes in practice may not have gone to a policy level. Anticipating staff turnover in general, the entire ECEC unit must be committed to changing its patterns to promote children’s PA (Repo et al., 2020). The primary strength of the present study is that it is a long-term PA intervention with a follow-up, which is still scarce in this research area (Jones et al., 2019). The HIPPA intervention was part of the usual ECEC time, so the whole child group, and not only the children participating in the research, were influenced. This work also has limitations that should be recognised. First, PA was measured by observing randomly selected children amongst the participants. As a method, observation is valuable when searching for detailed information about the social, physical, and pedagogical environments associated with PA (Brown et al., 2006; Loprinzi & Cardinal, 2011). However, it should be noted that activities during the observation recordings were not observed and, therefore, not recorded. The intervals were also not recorded if the next child participating in the study was not immediately found for observation. Thus, only parts of their ECEC times were recorded. Second, the highest PA level achieved by an observed child during the observation interval was recorded. At the same time, all the activities in which they participated during this time frame were recorded. This might have slightly increased the mean PA of typically 190 Mehtälä, Sääkslahti, Soini, Poskiparta & Kokko. Journal of Early Childhood Education Research 12(1) 2023, 169–204. http://jecer.org very sedentary activities. However, the amount of the increase is expected to be the same for both the intervention and control groups. Third, the number of participating children was relatively low, which may have affected the precision of our results. Thus, the effects of a small sample size should be kept in mind when interpreting the results. Finally, the PA knowledge and skills of the ECEC managers, the personnel’s competence in supporting children’s PA, and the children’s own motivations and self-efficacies were not assessed; however, they were objects in the HIPPA intervention. Hence, the impacts of these factors on the children’s PA remain unresolved. Conclusions The results of the present study show that the real-life PA intervention, implemented over one year by ECEC personnel with the support of researchers, increased PA in four-yearold children. To the best of our knowledge, this study is the first attempt to determine from the child’s level where the intervention effects lie and what the methods are like amidst the background of the impact of a multi-component intervention in ECEC settings. Our results indicate that children’s PA could be promoted by allowing them to be physically active indoors during unstructured time. Creating small indoor areas suitable for active play is a cost-effective and easy-to-implement strategy (Hendersson et al., 2015). Keeping PA equipment, such as balls, jumping ropes, and trampolines, available for children, along with indoor facilities, can inspire children to engage in active play more effectively. Implementing these methods may require changes in the practices of the ECEC settings and personnel’s attitudes towards children’s physically active play, especially indoors. Furthermore, influencing personnel’s perceptions of their competence in supporting children’s active play through ECE teacher education could also be a long-term strategy that can be translated into a policy-level strategy (Soini et al., 2021). Current evidence suggests that increasing childcare-aged children’s PA requires multicomponent interventions (Jones et al., 2019; Mehtälä et al., 2014) and that the ECEC setting is ideal for implementing an effective intervention; this is also supported by the present study. However, the ECEC setting per se is a complex environment, with various interactions among personnel, children, and their environment (Jones et al., 2019; Mehtälä et al., 2014). To improve the effectiveness of an intervention on a larger scale in real-life scenarios, it is essential to evaluate the effective methods in multi-component interventions and why they are effective. 191 Mehtälä, Sääkslahti, Soini, Poskiparta & Kokko. Journal of Early Childhood Education Research 12(1) 2023, 169–204. http://jecer.org Acknowledgements This work was supported by the Ministry of Social Affairs and Health, the Ministry of Culture and Education, and by a research grant awarded by the Juha Vainio Foundation in Finland. We wish to thank the Jamk University of Applied Sciences for granting the educational leave that facilitated the finalization of the article. We are also grateful to the MSc students who helped us with the data collection. Finally, special thanks are offered to the participating ECEC personnel and the children’s families. References Alhassan, S., Sirard, J. R., & Robinson, T. N. (2007). The effects of increasing outdoor play time on physical activity in Latino preschool children. International Journal of Pediatric Obesity, 2, 153–158. https://doi.org/10.1080/17477160701520108 Aras, S. (2016). Free play in early childhood education: a phenomenological study. Early Child Development and Care, 186(7), 1173–1184. https://doi.org/10.1080/03004430.2015.1083558 Brian, A., Goodway, J. D., Logan, J. A., & Sutherland, S. (2017). SKIPing with Head Start teachers: Influence of T-SKIP on object-control skills. Research Quarterly for Exercise and Sport, 88(4), 479–491. https://doi.org/10.1080/02701367.2017.1375077 Brown, W. H., Pfeiffer, K. A., McIver, K. L., Dowda, M., Almeida, M. J. C. A., & Pate, R. R. (2006). Assessing preschool children´s physical activity: the observational system for recording physical activity in children-preschool version. Research Quarterly for Exercise and Sport 77(2), 167–176. https://doi.org/10.1080/02701367.2006.10599351 Brown, W. H., Pfeiffer, K. A., McIver, K. L., Dowda, M., Addy, C. L., & Pate, R. R. (2009). Social and environmental factors associated with preschoolers’ nonsedentary physical activity. Child Development, 80(1), 45–58. https://doi.org/10.1111/j.1467-8624.2008.01245.x Cardon, G., Labarque, V., Smits, D., & De Bourdeaudhuij, I. (2009). Promoting physical activity at the pre-school playground: the effects of providing markings and play equipment. Preventive Medicine, 48(4), 335–340. https://doi.org/10.1016/j.ypmed.2009.02.013 Finch, M., Wolfenden, L., Morgan, P. J., Freund, M., Jones, J., & Wiggers, J. (2014). A cluster randomized trial of a multi-level intervention, delivered by service staff, to increase physical activity of children attending center-based childcare. Preventive Medicine, 58, 9– 16. https://doi.org/10.1016/j.ypmed.2013.10.004 Finnish National Agency for Education (EDUFI). (2022). National core curriculum for ECEC in a nutshell. Publications of the Ministry of Education and Culture. https://www.oph.fi/en/education-and-qualifications/national-core-curriculum-ececnutshell Gray, C., Gibbons, R., Larouche, R., Sandseter, E. B. H., Bienenstock, A., Brussoni, M., Chabot, G., Herrington, S., Janssen, I., Pickett, W., Power, M., Stanger, N., Sampson, M., & Tremblay, M. S. (2015). What is the relationship between outdoor time and physical activity, sedentary 192 Mehtälä, Sääkslahti, Soini, Poskiparta & Kokko. Journal of Early Childhood Education Research 12(1) 2023, 169–204. http://jecer.org behaviour, and physical fitness in children? A systematic review. International Journal of Environmental Research and Public Health, 12(6), 6455–6474. https://doi.org/10.3390/ijerph120606455 Gubbels, J. S., Kremers, S. P. J., van Kann, D. H. H., Staufleu, A., Dagnelie, P. C., Thijs, C., & de Vries, N. K. (2011). Interaction between physical environment, social environment, and child characteristics in determining physical activity at child care. Health Psychology, 30(1), 84–90. https://doi.org/10.1037/a0021586 Gubbels, J. S., Van Kann, D. H., de Vries, N. K., Thijs, C., & Kremers, S. P. (2014). The next step in health behavior research: The need for ecological moderation analyses—An application to diet and physical activity at childcare. International Journal of Behavioral Nutrition and Physical Activity, 11, 52. https://doi.org/10.1186/1479-5868-11-52 Hallgren, K. A. (2012). Computing inter-rater reliability for observational data: An overview and tutorial. Tutorials in Quantitative Methods for Psychology, 8, 23–34. https://doi.org/10.20982/tqmp.08.1.p023 Heikkilä, M., Ihalainen, S., & Välimäki, A. (2004). National curriculum guidelines on early childhood education and care in Finland. Stakes. https://www.julkari.fi/bitstream/handle/10024/75535/267671cb-0ec0-4039-b97b7ac6ce6b9c10.pdf?sequence=1 Henderson, K. E., Grode, G. M., O’Connell, M. L., & Schwartz, M. B. (2015). Environmental factors associated with physical activity in childcare centers. International Journal of Behavioral Nutrition and Physical Activity, 12, 43. https://doi.org/10.1186/s12966-015-0198-0 Hill, J. O., Wyatt, H. R., Reed, G. W., & Peters, J. C. (2003). Obesity and the environment: Where do we go from here? Science, 299(5608), 853–855. https://doi.org/10.1126/science.1079857 Hnatiuk, J. A., Brown, H. E., Downing, K. L., Hinkley, T., Salmon, J., & Hesketh, K. D. (2018). Interventions to increase physical activity in children 0-5 years old: a systematic review, meta-analysis and realist synthesis. Obesity Reviews, 20(1), 75–87. https://doi.org/10.1111/obr.1276 Iivonen, S., Niemistö, D., Sääkslahti, A., & Kettukangas, T. (2021). What makes John move? Outdoor play physical environmental factors changing a child’s activity from sedentary to physically active: longitudinal mixed-method case study. Journal of Early Childhood Education Research, 10(3), 21–53. Jones, R. A., Sousa-Sá, E., Peden, M., & Okely, A. D. (2019). Childcare physical activity interventions: A discussion of similarities and differences and trends, issues, and recommendations. International Journal of Environmental Research and Public Health, 16(23), 4836. https://doi.org/10.3390/ijerph16234836 Jylänki, P., Mbay, T., Hakkarainen, A., Sääkslahti, A., & Aunio, P. (2022). The effects of motor skill and physical activity interventions on preschoolers' cognitive and academic skills: A systematic review. Preventive Medicine, 155, 106948. https://doi.org/10.1016/j.ypmed.2021.106948 Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical data. Biometrics, 33(1), 159–174. https://doi.org/10.2307/2529310 199 Mehtälä, Sääkslahti, Soini, Poskiparta & Kokko. Journal of Early Childhood Education Research 12(1) 2023, 169–204. http://jecer.org APPENDIX 2 Kappa coefficients and interval-by-interval agreement for observed categories in Observational System for Recording Physical Activity in Children – Preschool (OSRAC-P; Brown et al., 2006) (n = 76,800 observations). OBSERVED CATEGORY M SD 95%CI Physical activity levels, all Kappa coefficient Interval-by-interval agreement (%) .497 87.6 .128 8.3 .407 – .587 80.3 – 94.9 Sedentary Kappa coefficient Interval-by-interval agreement (%) .695 85.5 .065 N/A .653 – .738 N/A Light Kappa coefficient Interval-by-interval agreement (%) .483 78.4 .098 N/A .425 – .537 N/A MVPA Kappa coefficient Interval-by-interval agreement (%) .611 90.8 .122 N/A .530 – .691 N/A Physical activity types Kappa coefficient Interval-by-interval agreement (%) .667 97.4 .200 4.2 .539 – .795 95.6 – 99.2 Group compositions, all Kappa coefficient Interval-by-interval agreement (%) .732 92.9 .122 3.4 .663 – .802 90.0 – 95.9 Adult involved or present Kappa coefficient Interval-by-interval agreement (%) .803 93.7 .170 N/A .752 – .853 N/A Child/ren only involved or present Kappa coefficient Interval-by-interval agreement (%) .793 93.6 .199 N/A .746 – .840 N/A Primary locations Kappa coefficient Interval-by-interval agreement (%) .836 99.7 .105 0.1 .761 – 911 99.6 – 99.8 Indoor activities Kappa coefficient Interval-by-interval agreement (%) .779 99.1 .181 1.5 .698 – .859 98.4 – 99.7 Outdoor activities Kappa coefficient Interval-by-interval agreement (%) .721 98.7 .181 1.7 .630 – .812 97.9 – 99.6 Initiator of activities Kappa coefficient Interval-by-interval agreement (%) .874 94.8 .077 0.0 .846 – .902 94.8 – 94.8 Prompts for PA, all Kappa coefficient Interval-by-interval agreement (%) .499 99.7 .218 0.4 .296 – .701 99.3 – 100.0 No prompts Kappa coefficient Interval-by-interval agreement (%) 0.491 99.2 .231 N/A .282 – .700 N/A All categories Kappa coefficient Interval-by-interval agreement (%) .705 95.8 .155 2.0 .617 – .793 94.3 – 97.3 200 Mehtälä, Sääkslahti, Soini, Poskiparta & Kokko. Journal of Early Childhood Education Research 12(1) 2023, 169–204. http://jecer.org APPENDIX 3 Percentage of intervals at the post-intervention (T2) and percentage changes in observed categories between the start of the intervention (T1) and the post-intervention (T2) by PA levels and by intervention condition. CONTROL (n = 36)* INTERVENTION (n = 52)* OBSERVED CATEGORIES OBSERVED CODES OBSERVED INTERVALS (T2 % OF TOTAL) PERCENTAGE OF INTERVALS AT T2 BY ACTIVITY LEVELS (T2% - T1%) OBSERVED INTERVALS (T2 % OF TOTAL) PERCENTAGE OF INTERVALS AT T2 BY ACTIVITY LEVELS (T2% - T1%) Sedentary (Levels 1-2) Light (Level 3) MVPA (Levels 4-5) Sedentary (Levels 1-2) Light (Level 3) MVPA (Levels 4-5) Primary locations Inside 2,468 (51) 64 (-1) 28 (2) 9 (-1) 3,182 (51) 51 (-19) 34 (8) 16 (11) Outside 2,304 (51) 26 (-9) 45 (10) 29 (-1) 2,829 (51) 19 (-19) 47 (4) 34 (15) Transition 20 (80) 50 (0) 44 (19) 6 (-19) 44 (52) 35 (-18) 48 (0) 17 (17) Total observed intervals 4,792 (51) 6,055 (53) Physical activity types Sit or squat 2,568 (50) 64 (-1) 32 (5) 4 (-4) 3,286 (56) 51 (-28) 38 (19) 11 (9) Stand 1,805 (49) 39 (-5) 46 (6) 15 (-1) 2,267 (50) 24 (-21) 56 (9) 20 (12) Walk 1,022 (49) 5 (-7) 66 (5) 29 (3) 1,827 (47) 1 (-10) 56 (-18) 42 (28) Run 414 (53) 1 (-3) 15 (-7) 84 (10) 532 (61) 1 (-5) 11 (-14) 88 (19) Climb 326 (56) 5 (-5) 37 (-10) 59 (16) 287 (62) 0 (-7) 42 (-19) 58 (27) Pull or push 339 (50) 6 (-2) 68 (22) 27 (-20) 280 (46) 2 (-2) 50 (9) 48 (-7) Swing 222 (38) 16 (-10) 32 (8) 52 (2) 152 (36) 17 (-17) 56 (-35) 27 (53) Jump or skip 206 (47) 1 (-4) 44 (14) 55 (-11) 134 (69) 0 (-2) 27 (-27) 73 (29) Ride 132 (44) 10 (4) 31 (-3) 59 (-1) 125 (52) 0 (-28) 29 (-11) 71 (39) Crawl 107 (42) 2 (-8) 71 (11) 27 (-4) 75 (35) 0 (-22) 50 (-11) 50 (34) Lie down 97 (40) 41 (-6) 51 (13) 8 (-8) 94 (43) 37 (-50) 40 (30) 24 (21) Balance 83 (25) 10 (0) 52 (4) 38 (-4) 60 (58) 3 (-1) 60 (-20) 37 (21) Throw 46 (52) 8 (4) 50 (0) 42 (-4) 56 (71) 0 (-6) 40 (-41) 60 (48) Rough and tumble 50 (44) 5 (-6) 55 (22) 41 (-16) 35 (71) 0 (0) 36 (-34) 64 (34) Slide 29 (55) 6 (-1) 69 (15) 25 (-14) 37 (54) 0 (0) 25 (-52) 75 (52) Swim 6 (50) 0 (0) 100 (67) 0 (-67) 46 (85) 0 (0) 44 (15) 56 (-15) Rock 9 (0) 0 (NA) 0 (NA) 0 (NA) 27 (52) 7 (-15) 63 (-12) 30 (28) Roll 18 (67) 8 (-8) 42 (8) 50 (0) 11 (55) 0 (0) 83 (43) 17 (-43) Dance 21 (86) 0 (0) 61 (-39) 39 (39) 2 (100) 0 (NA) 50 (NA) 50 (NA) Other 11 (64) 29 (29) 29 (-21) 43 (-7) 12 (8) 0 (-55) 0 (-18) 100 (73) Total observed intervals 7,511 (49) 9,345 (53) Appendix 3 Continues 201 Mehtälä, Sääkslahti, Soini, Poskiparta & Kokko. Journal of Early Childhood Education Research 12(1) 2023, 169–204. http://jecer.org CONTROL (n = 36)* INTERVENTION (n = 52)* OBSERVED CATEGORIES OBSERVED CODES OBSERVED INTERVALS (T2 % OF TOTAL) PERCENTAGE OF INTERVALS AT T2 BY ACTIVITY LEVELS (T2% - T1%) OBSERVED INTERVALS (T2 % OF TOTAL) PERCENTAGE OF INTERVALS AT T2 BY ACTIVITY LEVELS (T2% - T1%) Sedentary (Levels 1-2) Light (Level 3) MVPA (Levels 4-5) Sedentary (Levels 1-2) Light (Level 3) MVPA (Levels 4-5) Indoor activities Toys 706 (50) 64 (-4) 35 (9) 2 (-5) 1,141 (48) 50 (-31) 42 (24) 8 (7) Other 494 (57) 51 (-12) 38 (12) 11 (0) 771 (57) 45 (-9) 45 (1) 11 (9) Group time 539 (48) 87 (6) 11 (-4) 2 (-2) 418 (59) 90 (2) 9 (1) 1 (-3) Art 245 (81) 84 (-11) 14 (10) 2 (2) 310 (49) 78 (-16) 17 (10) 6 (6) Sociodramatic 255 (25) 32 (-25) 62 (27) 6 (-2) 194 (61) 19 (-15) 50 (-12) 31 (27) Teacher arranged 197 (71) 26 (-7) 31 (-8) 43 (15) 153 (38) 24 (-12) 35 (-10) 41 (21) Transition 143 (36) 14 (-18) 60 (7) 27 (12) 151 (63) 6 (3) 33 (-48) 61 (45) Music 107 (69) 81 (11) 14 (-17) 5 (15) 70 (51) 97 (15) 0 (-18) 3 (3) Selfcare 56 (38) 38 (-28) 33 (5) 29 (23) 89 (49) 59 (-5) 36 (5) 5 (0) Gross motor 14 (43) 0 (-13) 17 (-8) 83 (21) 105 (82) 1 (1) 43 (6) 56 (-7) Housework 37 (11) 0 (-42) 75 (36) 25 (7) 48 (90) 63 (63) 28 (-72) 9 (9) Manipulative 13 (62) 88 (-13) 13 (13) 0 (0) 53 (28) 94 (-7) 7 (7) 0 (0) Videos 48 (58) 96 (1) 4 (-1) 0 (0) 7 (14) 0 (-100) 100 (100) 0 (0) Pool activity 8 (0) 0 (NA) 0 (NA) 0 (NA) 47 (83) 0 (-13) 44 (19) 56 (-6) Large block 37 (0) 0 (NA) 0 (NA) 0 (NA) 12 (67) 0 (-75) 25 (0) 75 (75) Snack 13 (46) 83 (-2) 17 (2) 0 (0) 21 (100) 38 (NA) 57 (NA) 5 (NA) Preacademic 16 (100) 75 (NA) 25 (NA) 0 (NA) 3 (100) 0 (0) 100 (0) 0 (0) Tantrum 2 (0) 0 (NA) 0 (NA) 0 (NA) 5 (80) 0 (0) 0 (0) 0 (0) Time out 2 (0) 0 (NA) 0 (NA) 0 (NA) 3 (0) 0 (NA) 0 (NA) 0 (NA) Total observed intervals 2,902 (52) 3,601 (54) Appendix 3 Continues 202 Mehtälä, Sääkslahti, Soini, Poskiparta & Kokko. Journal of Early Childhood Education Research 12(1) 2023, 169–204. http://jecer.org CONTROL (n = 36)* INTERVENTION (n = 52)* OBSERVED CATEGORIES OBSERVED CODES OBSERVED INTERVALS (T2 % OF TOTAL) PERCENTAGE OF INTERVALS AT T2 BY ACTIVITY LEVELS (T2% - T1%) OBSERVED INTERVALS (T2 % OF TOTAL) PERCENTAGE OF INTERVALS AT T2 BY ACTIVITY LEVELS (T2% - T1%) Sedentary (Levels 1-2) Light (Level 3) MVPA (Levels 4-5) Sedentary (Levels 1-2) Light (Level 3) MVPA (Levels 4-5) Outdoor activities Fixed equipment 688 (50) 19 (-10) 47 (8) 37 (2) 673 (65) 19 (-11) 46 (3) 35 (8) Open space 376 (56) 27 (1) 45 (-2) 28 (1) 958 (58) 16 (-13) 41 (-9) 44 (21) Sandbox 526 (54) 32 (-13) 58 (19) 10 (-6) 724 (52) 30 (-27) 53 (16) 17 (11) Wheel 264 (48) 24 (8) 35 (2) 40 (-10) 310 (40) 4 (-23) 38 (6) 58 (17) Sociodramatic 178 (34) 47 (-15) 31 (20) 22 (-4) 299 (41) 34 (-30) 43 (8) 23 (22) Other 265 (66) 40 (-17) 41 (6) 19 (11) 125 (38) 29 (-21) 58 (21) 13 (1) Portable equipment 131 (31) 15 (-10) 65 (23) 20 (-13) 162 (55) 20 (-30) 53 (8) 37 (22) Games 199 (71) 18 (-14) 29 (10) 53 (4) 65 (51) 6 (-6) 46 (11) 49 (-5) Forest 126 (18) 17 (-19) 70 (37) 13 (-18) 100 (38) 0 (-19) 79 (14) 21 (5) Transition 114 (32) 27 (17) 49 (-2) 24 (-15) 67 (7) 0 (-26) 60 (-6) 40 (32) Ball and object 35 (89) 19 (-56) 45 (45) 36 (-11) 6 (100) 0 (NA) 50 (NA) 50 (NA) Teacher arranged 31 (74) 26 (-24) 35 (-15) 39 (39) 0 (0) 0 (NA) 0 (NA) 0 (NA) Tantrum 5 (80) 75 (-25) 25 (25) 0 (0) 19 (47) 11 (-89) 67 (67) 22 (22) Snacks 0 (0) NA NA NA 11 (0) 0 (NA) 0 (NA) 0 (NA) Pool 0 (0) NA NA NA 7 (100) 0 (NA) 43 (NA) 57 (NA) Time out 0 (0) NA NA NA 1 (0) 0 (NA) 0 (NA) 100 (NA) Total observed intervals 2,938 (51) 3,527 (53) Appendix 3 Continues 203 Mehtälä, Sääkslahti, Soini, Poskiparta & Kokko. Journal of Early Childhood Education Research 12(1) 2023, 169–204. http://jecer.org CONTROL (n = 36)* INTERVENTION (n = 52)* OBSERVED CATEGORIES OBSERVED CODES OBSERVED INTERVALS (T2 % OF TOTAL) PERCENTAGE OF INTERVALS AT T2 BY ACTIVITY LEVELS (T2% - T1%) OBSERVED INTERVALS (T2 % OF TOTAL) PERCENTAGE OF INTERVALS AT T2 BY ACTIVITY LEVELS (T2% - T1%) Sedentary (Levels 1-2) Light (Level 3) MVPA (Levels 4-5) Sedentary (Levels 1-2) Light (Level 3) MVPA (Levels 4-5) Activity initiators Child Initiated 3,617 (50) 39 (-6) 41 (8) 20 (-2) 4,956 (53) 30 (-23) 44 (10) 26 (13) Adult Initiated 1,175 (53) 62 (-3) 24 (0) 14 (3) 1,099 (53) 66 (7) 20 (-14) 14 (7) Total observed intervals 4,792 (51) 6,055 (53) Group composition Group child only 1556 (56) 42 (-3) 40 (7) 18 (-4) 2082 (54) 25 (-23) 46 (9) 30 (15) One-to-one peer 1,577 (49) 36 (-9) 42 (8) 22 (1) 1,861 (50) 29 (-24) 44 (9) 28 (16) Group with adult 1417 (56) 59 (-6) 26 (3) 15 (3) 1523 (50) 62 (1) 24 (-8) 14 (7) Solitary 894 (42) 33 (-15) 45 (13) 22 (2) 1,430 (55) 31 (-17) 45 (6) 24 (11) One-to-one adult 187 (48) 47 (-10) 39 (8) 14 (2) 223 (46) 29 (-40) 52 (26) 19 (15) Total observed intervals 5,631 (52) 7,119 (52) Prompts None 4,736 (51) 48 (-5) 33 (6) 19 (0) 6,038 (53) 45 (-18) 37 (6) 18 (12) Teacher increase 48 (19) 11 (-7) 78 (34) 11 (-27) 18 (0) 0(NA) 0(NA) 0(NA) Teacher decrease 5 (60) 33 (33) 33 (-17) 33 (-17) 2 (50) 100 (100) 0 (0) 50 (-100) Peer increase 4 (75) 0 (0) 0 (-100) 100 (100) 2 (50) 0 (0) 0 (0) 100 (0) Peer decrease 0 (0) 0 (NA) 0 (NA) 0 (NA) 1 (0) 0 (NA) 0 (NA) 0 (NA) Total observed intervals 4796 (51) 6061 (53) Note. *The observed children are the same at T1 and T2 (N = 88); Green color = percentual increase in the observed PA level between T1 and T2, Yellow color = no change in the observed PA level between T1 and T2, Red color = percentual decrease in the observed PA level between T1 and T2. 204 Mehtälä, Sääkslahti, Soini, Poskiparta & Kokko. Journal of Early Childhood Education Research 12(1) 2023, 169–204. http://jecer.org APPENDIX 4 Results of multilevel modeling Null model Random Intercept and Fixed Slope Two-way interactions Random Intercept and Random Slope Cross-Level Interactions Effect Estimate 95% CI (LL, UL) p Estimate 95% CI (LL, UL) p Estimate 95% CI (LL, UL) p Estimate 95% CI (LL, UL) p Estimate 95% CI (LL, UL) p Intercept 1.357 1.344, 1.370 <.001 1.396 1.379, 1.414 <.001 1.421 1.402, 1.439 <.001 1.424 1.403, 1.445 <.001 1.430 1.409, 1.452 <.001 Time 0.038 0.035, 0.042 <.001 0.012 0.005, 0.020 0.002 0.008 -0.007, 0.023 0.285 -0.007 -0.024, 0.009 .375 Condition 0.008 -0.015, 0.031 .444 -0.032 -0.057, -0.006 0.018 -0.037 -0.066, -0.009 0.011 -0.049 -0.079, -0.020 .002 Age 0.034 0.014, 0.055 .001 0.029 0.009, 0.049 0.005 0.019 0.001, 0.039 0.062 0.020 8.4X10-5, 0.040 .049 BMI 0.008 0.005, 0.011 <.001 0.008 0.005, 0.011 <.001 0.006 0.003, 0.010 <.001 0.006 0.003, 0.009 <.001 Gender -0.023 -0.036, -0.011 <.001 -0.024 -0.043, -0.005 0.013 -0.027 -0.045, -0.009 0.004 -0.030 -0.048, -0.011 .002 Primary location -0.083 -0.086, -0.080 <.001 -0.101 -0.106, -0.095 <.001 -0.101 -0.107, -0.095 <.001 -0.095 -0.101, -0.088 <.001 Time of day 0.006 0.003, 0.009 <.001 -0.004 -0.010, 0.001 0.117 -0.004 -0.009, 0.002 0.159 -0.014 -0.020, -0.007 <.001 Group composition -0.013 -0.018, -0.008 <.001 0.001 -0.010, 0.009 0.884 -0.002 -0.012, 0.007 0.653 0.001 -0.010, 0.013 .835 Initiator -0.014 -0.020, -0.009 <.001 -0.025 -0.035, -0.014 <.001 -0.024 -0.035, -0.014 <.001 -0.040 -0.053, -0.028 <.001 Time Condition 0.038 0.032, 0.045 <.001 0.043 0.024, 0.063 <.001 0.071 0.048, 0.094 <.001 Time Gender 0.005 -0.001, 0.012 0.115 0.011 0.004, 0.018 0.002 0.018 0.008, 0.029 .001 Time Primary location 0.005 -0.001, 0.012 0.112 0.005 -0.001, 0.012 0.108 -0.010 -0.019, 3.1X10-4 .058 Time Time of day 0.005 -0.002, 0.011 0.145 0.004 -0.002, 0.010 0.218 0.025 0.016, 0.034 <.001 Time Croup composition -0.011 -0.021, -0.001 0.026 -0.009 -0.019, 0.001 0.064 -0.013 -0.028, 0.003 .011 Time Initiator -0.006 -0.017, 0.006 0.326 -0.007 -0.019, 0.004 0.213 0.024 0.007, 0.041 .006 Condition Gender 0.001 -0.026, 0.024 0.932 0.001 -0.023, 0.025 0.938 0.006 -0.019, 0.030 .638 Condition Primary location 0.026 0.020, 0.033 <.001 0.026 0.020, 0.033 <.001 0.017 0.008, 0.025 <.001 Condition Time of day 0.012 0.006, 0.019 <.001 0.013 0.006, 0.019 <.001 0.029 0.021, 0.038 <.001 Condition Croup composition -0.009 -0.020, 0.001 0.07 -0.009 -0.019, 0.002 0.101 -0.012 -0.026, 0.002 .095 Condition Initiator 0.026 0.014, 0.037 <.001 0.026 0.015, 0.038 <.001 0.052 0.036, 0.068 <.001 Time Condition Gender -0.013 -0.028, 0.001 .066 Time Condition Primary location 0.025 0.012, 0.038 <.001 Time Condition Time of day -0.037 -0.049, -0.024 <.001 Time Condition Croup composition 0.005 -0.015, 0.025 .621 Time Condition Initiator -0.056 -0.078, -0.033 <.001 Variance components Within child variance 9.3x10-3 9.1x10-3, 9.5x10-3 <.001 6.9x10-3 6.7x10-3, 7.0x10-3 <.001 6.7x10-3 6.5x10-3, 6.9x10-3 <.001 6.7x10-3 6.5x10-3, 6.8x10-3 <.001 6.6x10-3 6.4x10-3, 6.8x10-3 <.001 Variance of intercepts across children 1.2x10-3 8.5x10-4, 1.6x10-3 <.001 9.7x10-3 7.1x10-4, 1.3x10-3 <.001 9.0x10-4 6.5x10-4, 1.2x10-3 <.001 8.2x10-4 6.0x10-4, 1.1x10-3 <.001 8.1x10-4 5.9x10-4, 1.1x10-3 <.001 Variance of intercepts and slopes across centres 3.2x10-4 9.0x10-5, 1.1x10-3 .119 2.2x10-4 5.0x10-5, 9.7x10-4 .187 2.1x10-4 5.0x10-5, 9.0x10-4 .175 1.3x10-4 6.2x10-5, 2.6x10-4 .007 1.4x10-4 6.9x10-5, 2.9x10-4 .006 Additional information ICC 0.029 AIC -22310 -25645 -25858 -25923 -25980 BIC -22288 -25623 -25836 -25901 -25958 Number of estimated parameters 4 13 24 24 29