BACKGROUND: Pupils in secondary schools do not meet the targets for physical activity levels during physical education (PE) sessions, and there is a lack of data on the vigorous physical activity domain (VPA) in PE known to be positively associated with cardio metabolic health While PE session intensity depends on a variety of factors, the large majority of studies investigating these factors have not taken into account the nested structure of this type of data set. Therefore, the aim of this study was to investigate the relationship between various factors (gender, activity type, class location and class composition) and various activity levels during PE classes in secondary schools, using a multi-level statistical approach. METHODS: Year eight (12-13 years old) adolescents (201 boys and 106 girls) from six schools were fitted with accelerometers during one PE session each, to determine the percentage (%) of the PE session time spent in sedentary (SPA), light (LPA), moderate (MPA), vigorous (VPA) and moderate-to-vigorous (MVPA) intensity levels. Two- and three-level (pupils, n = 307; classes, n = 13, schools, n = 6) mixed-effect models were used to assess the relationship between accelerometer-measured physical activity levels (% of class time spent in various activity levels) and gender, activity type, class location and composition. RESULTS: Participants engaged in MVPA and VPA for 30.7 ± 1.2% and 11.5 ± 0.8% of PE classes, respectively. Overall, no significant association between gender or class composition and PA was shown. A significant relationship between activity type and PA was observed, with Artistic classes significantly less active than Fitness classes for VPA (5.4 ± 4.5 vs. 12.5 ± 7.1%, p = 0.043, d:1.19). We also found a significant association between class location and PA, with significantly less time spent in SPA (24.8 ± 4.8% vs. 30.0 ± 3.4%, p = 0.042, d:0.77) and significantly more time spent in VPA (12.4 ± 3.7% vs. 7.6 ± 2.0%, p = 0.022, d:1.93) and MVPA (32.3 ± 6.7% vs.24.8 ± 3.8%, p = 0.024, d:1.33) in outdoors vs. indoors classes. CONCLUSIONS: The results suggest that class location and activity type could be associated with the intensity of PA in PE. It is essential to take into account the clustered nature of this type of data in similar studies if the sample size allows it.