Two medium-sized municipalities in the mid-Norway were chosen with the help of a stratified selection. In all, 18 of the total of 19 primary schools in these municipalities participated in the study (one of the smaller schools declined to participate). Pupils from the 18 primary schools were subsequently spread between 4 different secondary schools in the same municipalities in 2018 and 2019. At the start of the data collection in 2017, the sample consisted of 320 12–13-year-old pupils (7^th grade), who each had valid measurements of their activity from all three measurement periods, representing a 77% response rate. In all, 234 (124 girls and 110 boys) had valid measurements for all three time periods, concerning answering research questions 1 and 2, a response rate of 56%. Only 91 pupils (61 girls and 36 boys) had valid data for all three time periods regarding research question 3 – a 23% response rate. A sample size calculator show that we needed more than 30 participants to fulfil the criteria for observed power, and our study includes a much higher number of participants.

In this study, PA is defined as MVPA, which is the gold standard for the measurement of overall PA (30, 31), and in line with both national (2) and international (1) recommendations. All measurements were taken during the same week in April 2017, 2018, and 2019. MVPA was measured with the use of accelerometer ActiGraph GT1M, which is an approved and reliable instrument for the measurement of PA (32, 33). The pupils were instructed to carry the accelerometer on their right hip for 7 days, except only during water activities as the monitors do not tolerate water (30), and at night when sleeping. In keeping with large Norwegian population studies, the criterion for valid measurements was set at data over 8 h daily for at least 2 days, and MVPA was characterized as >2000 counts per min (12, 30). The storage interval for the raw data was set at 10 s. All data collection, from all measurements, was done by the same test leader. The same equipment was used throughout, and the procedures used were constant.

Data from the data collection was downloaded to Actilife v6.13.3 (ActiGraph, LLC, Pensacola, FL), where they were filtered and MVPA during schooltime separated from MVPA in leisure time. For the pupils' data to be valid, they must have a total of 480 min of counts (12). Periods of more than 20 min with no counts were not included, nor was activity between 00.00 and 06.00. The pupils must have a minimum of 180 min of counts for the school day to be considered valid.

The data material was analyzed using the program SPSS, version 27 (SPSS Inc., Chicago, IL, USA). As most of the dependent variables were not normally distributed according to the Kolmogorov Smirnov test (p < 0.05), non-parametric tests were used (34). Friedmann's test was used to study differences in MVPA during schooltime, and schooltime's contribution to the health recommendations over the three time periods, with the Wilcoxon test as a follow-up test to find pair-wise differences between the different years, using Bonferroni corrections. Finally, the Mann–Whitney U-test was used to examine gender differences. The significance level was set at p < 0.05, p < 0.01, and p < 0.001 respectively.

The results showed a significant reduction in weekly MVPA among the pupils between 7^th grade and 8^th grade, from the age of 13 to the age of 15. This applied to both girls and boys, representing a reduction in PA s during the transition from primary school to secondary school. There were, however, no significant differences in schooltime's contribution to weekly MVPA between 8^th grade and 9^th grade (13–14 years old). That schooltime's contribution to the pupils' weekly MVPA falls between primary school and secondary school is in agreement with the findings of Remmers et al. (25). Harding et al. (21) also found a significant reduction in pupils' MVPA during schooltime, although their study took only primary school. Compared to the findings in both Remmers' Dutch study and the findings in Hardings' English study, the decline in schooltime's contribution of MVPA, was significantly larger in this study. The smaller reduction in MVPA in Harding et al. (21), can be because the pupils had not moved schools between the two sets of measurements. The difference may also be due to cultural and school policy differences, as suggested by Andersen (19). In his cross-sectional study, Andersen found that schooltime accounted for less MVPA among fifteen-year-olds than among nine-year-olds, and his findings about the fifteen-year-olds MVPA were relatively similar to those in our study. Thus, both our study and the other studies presented here, point to a marked reduction in the contribution to pupils' MVPA made during schooltime after they start secondary school.

Research further suggests that physical education (PE) makes a similar, but quite small contribution to pupils' weekly MVPA in primary and in secondary school (19, 35), with 19 finding that PE lessons accounted for roughly 4%, in both primary and secondary school, of the recommended level of weekly MVPA. As teaching is organized in the same way in both primary and secondary schools (sitting at a desk for 45 min), our findings point in the direction of the changes taking place during recess.

There is evidence that the transition from primary to secondary school is a sensitive period (6). When pupils enter secondary school, there are many new things to deal with, and good activity habits at recess can be one of them. Being physically active may be more difficult because the other pupils are older, it is more difficult to stand out, and where physical play may be seen as childish. An interview study of some pupils from the same sample as the present study found that the transition to secondary school involved getting a new school environment, new teachers, and new classmates to get to know and relate to, as well as increased freedom and new and increased demands (24). The latter findings are in agreement with Ashton (27) which discovered that 7^th graders associated secondary school with increased responsibility, freedom, and choices. The increased experience of freedom in secondary school may be related to the possibility to choose where to spend their recess time. While before they were forced to be outside during recess, the pupils can choose to be inside at secondary school. This can allow for greater passivity during recess, as the outdoor area forms both a sanctuary for the pupils and a more conducive environment for physically active play (36). Results from Andersen (19), showed that while 9-year-olds spent 27% of their free time in MVPA, pupils in secondary school spent only 10.5% of their free time in MVPA. Further, the study found that breaks accounted for 18.2 and 6.8%, respectively, of the weekly attainment of the health recommendation for 9-year-old and 15-year-old pupils, respectively. This supports our contention that it is reduced activity during breaks at secondary school that occasions the substantial fall in weekly MVPA during schooltime between primary and secondary school. That our results show no significant difference in MVPA between 8^th grade and 9^th grade further strengthens this line of argument.

The results showed no significant differences, in any of the three sets of measurements, in school time's contribution of weekly MVPA in relation to the total samples' weekly MVPA. However, there were significant differences in schooltime's contribution to weekly MVPA for the girls between 7^th and 8^th grade, from the age of 13 to the age of 14. The results showed that the contribution of school time weekly MVPA to total weekly MVPA, was under 30% for both genders in 7^th, 8^th, and 9^th grade. This means that around 70% or more of the pupils' total MVPA occurs in their free time. Considering that MVPA is highest in spring and summer, compared with winter (37), it is worrying that the contribution of weekly MVPA during school time is so low in all three measurements, which were made in spring. It is also concerning that young people's PA behaviour seems to be significantly altered between the final year of primary school and the first of secondary school, especially given that habits formed at a young age are likely to persist into adult (14, 15). The numbers presented in the study by Remmers et al. (25) support this finding, with schooltime contributing 23.5 and 28.5% of, respectively, the girls' and the boys' during their final year in primary school and first year of secondary school. Other research also indicates that the time spent at school contributes a relatively low percentage of the pupils' total MVPA, both in school (38) and in pre-school (16).

Considering that pupils are at school for approximately 40% of their waking time (13), one could expect a larger contribution from schooltime. According to Dale et al. (39) pupils do not compensate for school days with little PA with a higher degree of PA in their free time. In fact, it seems that school days with a higher level of activity contribute to a higher level of activity outside school (39).

The results show a significant reduction in schooltime's contribution, toward the recommended level of PA (1, 2), between 7^th and 8^th grade (from the age of 13 to the age of 14), whilst there was no significant difference between 8^th grade and 9^th grade. The reduction in schooltime's contribution toward the recommended level of PA from 7^th grade to 8^th grade, was 41.3% for girls and 35.5% for boys. Results from earlier studies of Norwegian pupils also suggest that the contribution of schooltime to the pupils' attainment of the health recommendation is greater in primary school than in secondary school (19, 20). Longitudinal studies from other countries also support the findings of a reduction through time in school time's contribution to recommended levels of MVPA (21, 25). Thus, the results from this study show that, in all 3 sets of measurements (7^th, 8^th, and 9^th grade), the school stands for less than half of the number of minutes required to fulfill the recommended weekly amount. Earlier studies have pointed to similar findings (19–21, 25, 26).

The results showed that the boys had significantly more MVPA in schooltime than the girls at all three grades, and, further, that the contribution of schooltime to the fulfilment of the recommended level was significantly higher for the boys than for the girls in all three grades. That boys accumulate more MVPA during schooltime is supported by earlier research (19, 20, 25, 26).

Several studies have attempted to clarify the factors contributing to this gender difference in PA. Blatchford et al. (40) found in their study that boys were significantly more involved in ball games during breaks, while girls socialized more with friends in the form of conversations, and as verbal and sedentary play. Their study suggest that the genders use their break time for different activities, with “boys” activities' contributing to more MVPA (40). Girls' and boys' different maturational processes have also been discussed as a possible explanations for these gender differences. The psychological and biological changes occurring in connection with puberty can affect young people's self-perception, which in turn may affect their PA (41). It is possible that these changes may have a negative effect on girls' self-esteem, possibly leading to a reduction of PA. For boys, these changes may lead to an increase in self-esteem, showing itself in increased MVPA (41).

One can argue that such large gender differences in MVPA during schooltime are problematic, especially since the gender differences increase between primary school and secondary school. PE may also help to explain some of the gender differences in schooltime's contribution of MVPA. Several studies indicate that PE is better adapted to boys than to girls (42–44). As the subject is perceived by many as an arena where pupils can engage in the same activities as in organized sports, especially ball activities (45, 46), it is perhaps not unexpected that more boys than girls enjoy PE (46). Because wellbeing and enjoyment are factors in increased PA (47), wellbeing in PE may be thought of as having an impact on differences between boys' and girls' MVPA during school time. This finding gives reason for being particularly attentive to the need for adaptation and differentiation in PE.

The results also show that schooltime's contribution of MVPA in relation to the total weekly MVPA was significantly greater for the boys than for the girls in grades 7 and 8, but not in 9^th grade. The boys not only spent more time in MVPA in schooltime in all three measuring periods, but in relation to the pupils' total weekly MVPA, schooltime represented a significantly greater proportion among the boys than the girls. This means that the girls must accumulate a larger proportion of MVPA in their free time. In this way, the school contributes to gender differences.

A clear strength of the study is its use of a longitudinal design, in which reasonably many participants were measured in each of the three school years. All three sets of measurements used standardized and objective measuring instruments, and the test leader, equipment and procedures were the same throughout, and all three measurement periods were at the same time of the year. With longitudinal studies one can expect a certain level of drop out, and a response rate of 56% is therefore satisfactory (48). Statistical analysis (Mann–Whitney U-test) of the pupils returning valid results in all three time periods and those who dropped out, showed no significant differences between the two samples (Z = −1.57, p = 0.12). For this reason, the drop out is seen as being random, and those remaining as being representative of the entire sample. The use of accelerometers as a means of measurement of PA, is seen by many researchers as the best means of measuring PA among children and adolescents (30, 31).

There are, however, certain weaknesses in using accelerometers. They do not tolerate water, which could have significance for the measurement of the pupils' MVPA (30). Another weakness in the use of accelerometers is their limited capacity to register movements involving little vertical acceleration (30). This means that in the case of, for example, cycling or upper-body training, MVPA may be under-recorded. However, there was no possibility of cycling during schooltime in the periods when measurements were being recorded, nor swimming lessons at that time. For practical reasons, a stratified sample of municipalities was used, as there are difficulties with random sampling when conducting studies at this scale. At the same time, there is much that suggests that these two municipalities are not distinctly different from other Norwegian municipalities where gender balance, number of pupils from urban and rural areas, or spread of socioeconomic status are concerned.